INDICE

1. 2. 3. 4. 5. 6. 7. 8. 9.

INTRODUCTION TO CDMA ....................................................................................................... 01 FEATURES AND ADVANTAGES OF CDMA .......................................................................... 19 FORWARD CDMA CHANNEL ................................................................................................... 32 REVERSE CDMA CHANNEL ..................................................................................................... 37 FREQUENCY PLANS ................................................................................................................... 41 WALSH FUNCTIONS ................................................................................................................... 42 LONG CODE MASK ..................................................................................................................... 43 SHORT CODE ................................................................................................................................ 44 REVERSE SPREADING ............................................................................................................... 45

10. LONG CODE MASK ..................................................................................................................... 47 11. SPREAD SPECTRUM, INTRODUCTION .................................................................................. 50 12. SPREAD SPECTRUM, APPLICATIONS .................................................................................... 81

http://www.qualcomm.com/ProdTech/cdma/training/cdma25/m5/m5p01.html Introduction CDMA signal generation The types of codes used in CDMA Forward and reverse link code channels CDMA call processing to CD M A

The words "code" and "division" are important parts of how CDMA works. CDMA uses codes to convert between analog voice signals and digital signals. CDMA also uses codes to separate (or divide) voice and control data into data streams called "channels." These digital data stream channels should not be confused with frequency channels There are five steps in generating a CDMA signal. 1. 2. 3. 4. 5. analog to digital conversion vocoding encoding and interleaving channelizing the signals conversion of the digital signal to a Radio Frequency (RF) signal

The use of codes is a key part of this process.

Analog to digital conversion The first step of CDMA signal generation is analog to digital conversion, sometimes called A/D conversion. CDMA uses a technique called Pulse Code Modulation (PCM) to accomplish A/D conversion.

Voice Compression The second step of CDMA signal generation is voice compression. CDMA uses a device called a vocoder to accomplish voice compression. The term "vocoder" is a contraction of the words "voice" and "code." Vocoders are located at the BSC and in the phone.

How compression works People pause between syllables and words when they talk. CDMA takes advantage of these pauses in speech activity by using a variable rate vocoder.

Variable Rate Vocoder A CDMA vocoder varies compression of the voice signal into one of four data rates based on the rate of the user's speech activity. The four rates are: Full, 1/2, 1/4 and 1/8.

The vocoder uses its full rate when a person is talking very fast. It uses the 1/8 rate when the person is silent or nearly so.

Vocoder types CDMA systems can use either an 8 kbps (kilobytes per second) or a 13 kbps vocoder. The earliest CDMA systems used the 8kbps vocoder to maximize capacity. The 13 kbps vocoder was later developed to provide a more land-line quality voice signal. The great improvement in quality was worth the slight reduction in capacity. Recently the CDMA community adopted a new 8 kbps vocoder. This new vocoder is usually referred to as the EVRC (Extended Variable Rate Coding). It combines the quality of 13 kbps vocoding with the capacity of the 8kbps data rate.

Encoding and interleaving Encoders and interleavers are built into the BTS and the phones. The purpose of the encoding and interleaving is to build redundancy into the signal so that information lost in transmission can be recovered.

How encoding works The type of encoding done at this stage is called "convolutional encoding." A simplified encoding scheme is shown here. A digital message consists of four bits (A, B, C, D) of vocoded data. Each bit is repeated three times. These encoded bits are called symbols. The decoder at the receiver uses a majority logic rule. Thus, if an error occurs, the redundancy can help recover the lost information.

Burst errors A burst error is a type of error in received digital telephone signals. Burst errors occur in clumps of adjacent symbols. These errors are caused by fading and interference. Encoding and interleaving reduce the effects of burst errors.

How interleaving works Interleaving is a simple but powerful method of reducing the effects of burst errors and recovering lost bits. In the example shown here the symbols from each group are interleaved (or scrambled) in a pattern that the receiver knows. De-interleaving at the receiver unscrambles the bits, spreading any burst errors that occur during transmission.

Channelizing The encoded voice data is further encoded to separate it from other encoded voice data. The encoded symbols are then spread over the entire bandwidth of the CDMA channel. This process is called channelization. The receiver knows the code and uses it to recover the voice data.

Two kinds of codes CDMA uses two important types of codes to channelize users. Walsh codes channelize users on the forward link (BTS to mobile). Pseudorandom Noise (PN) codes channelize users on the reverse link (mobile to BTS).

Walsh codes Walsh codes provide a means to uniquely identify each user on the forward link. Walsh codes have a unique mathematical property--they are "orthogonal." In other words, Walsh codes are unique enough that the voice data can only be recovered by a receiver applying the same Walsh code. All other signals are discarded as background noise.

PN codes Pseudorandom Noise (PN) codes uniquely identify users on the reverse link. A PN code is one that appears to be random, but isn't. The PN codes used in CDMA yield about 4.4 trillion combinations of code. This is a key reason why CDMA is so secure.

Digital to Radio Frequency (RF) conversion The BTS combines channelized data from all calls into one signal. It then converts the digital signal to a Radio Frequency (RF) signal for transmission.

10

Digital to analog conversion After the CDMA signal is transmitted, the receiver must reverse the signal generation process to recover the voice, as follows: 1. 2. 3. 4. 5. Conversion of RF signal to digital signal Despreading the signal Deinterleaving and decoding Voice decompression Digital to analog voice recovery

Code channels used in CDMA A code channel is a stream of data designated for a specific use or person. This channel may be voice data or overhead control data. Channels are separated by codes.

11

The forward and reverse links use different types of channels.

Forward link channels The forward link uses four types of channels to transmit voice and control data to the mobile. The types of forward link channels are: Pilot Sync Paging Traffic

Pilot channel The BTS constantly transmits the pilot channel. The mobile uses the pilot signal to acquire the system. It then uses the pilot signal to monitor and adjust the power needed in order to transmit back to the BTS.

12

Sync channel The BTS constantly transmits over the sync channel so the mobile can synchronize with the BTS. It provides the mobile with the system time and the identification number of the cell site. The mobile ignores the sync channel after it is synchronized.

Paging channel CDMA uses up to seven paging channels. The paging channel transmits overhead information such as commands and pages to the mobile. The paging channel also sends commands and traffic channel assignment during call set-up. The mobile ignores the paging channel after a traffic channel is established.

13

Forward link traffic channel CDMA uses between fifty-five and sixty-one forward traffic channels to send both voice and overhead control data during a call. Once the call is completed, the mobile tunes back in to the paging channel for commands and pages.

Reverse link channels

14

The reverse link uses two types of channels to transmit voice and control data to the BTS. The types of reverse link channels are: Access Traffic

Access channel The mobile uses the access channel when not assigned to a traffic channel. The mobile uses the access channel to: Register with the network Originate calls Respond to pages and commands from the base station Transmit overhead messages to the base station

Reverse link traffic channel

15

The reverse traffic channel is only used when there is a call. The reverse traffic channel transmits voice data to the BTS. It also transmits the overhead control information during the call.

Call processing stages There are four stages or modes in CDMA call processing: Initialization mode Idle mode Access mode Traffic mode.

Initialization mode

16

During initialization, the mobile: acquires the system via the Pilot code channel synchronizes with the system via the Sync code channel

Idle mode The mobile is not involved in a call during idle mode, but it must stay in communication with the base station: The mobile and the base station communicate over the access and paging code channels The mobile obtains overhead information via the paging code channel.

Access mode The mobile accesses the network via the Access code channel during call origination. The Access channel and Paging channel carry the required call set-up communication between the mobile phone and the BTS until a traffic channel is established. Traffic mode During a land to mobile (LTM) call: The mobile receives a page on the paging channel. The mobile responds on the access channel. The traffic channel is established and maintained throughout the call.

During a mobile to land call (MTL): The call is placed using the Access channel. The base station responds on the paging channel. The traffic channel is established and maintained throughout the call.

Call processing (messages) During the call overhead messaging continues on the traffic channel in a limited fashion. This messaging uses "Dim and Burst" or "Blank and Burst" signaling, which replaces part of the voice traffic with system messages. The user does not detect this signaling, however, due to the strong data recovery schemes inherent to CDMA.

17

18

Features

and

Advantages

of

CDMA

CDMA has several unique features that make it a cost-effective, high quality wireless solution. In this module you will learn what those features are and how they provide advantages. Features of CDMA The following features are unique to CDMA technology: Universal frequency reuse Fast and accurate power control Rake receiver Different types of handoff

Frequency reuse The frequency spectrum is a limited resource. Therefore, wireless telephony, like radio, must reuse frequency assignments. For example, two radio stations might transmit at 91.3 FM. There is no interference as long as the radio stations are far enough apart.

Cell interference Cell A and B of a conventional, analog system are using the same frequency. The area of overlap, area C, has a frequency conflict and interference. This is similar to what you experience when you are driving between the broadcast zones of two radio stations transmitting at the same frequency.

19

FDMA and TDMA frequency reuse planning A frequency (channel) can be used again within an FDMA or TDMA network, but cells using the same frequency must be separated by an appropriate distance. Adjacent cells must be assigned a different set of frequencies. For example, a cell using frequency A must not be adjacent to another cell using frequency A. As a result, each cell site in the site is able to use only 1/7 of the possible frequencies.

20

CDMA frequency reuse planning Each BTS in a CDMA network can use all available frequencies. Adjacent cells can transmit at the same frequency because users are separated by code channels, not frequency channels. This feature of CDMA, called "frequency reuse of one," eliminates the need for frequency planning.

Power control Power control is a CDMA feature that enables mobiles to adjust the power at which they transmit. This ensures that the base station receives all signals at the appropriate power. The CDMA network independently controls the power at which each mobile transmits. Both forward and reverse links use power control techniques.

21

Why power control is needed If all mobiles transmitted at the same power level, the base station would receive unnecessarily strong signals from mobiles nearby and extremely weak signals from mobiles that are far away. This would reduce the capacity of the system. This problem is called the near-far problem.

Reverse link power control Reverse link power control consists of two processes: Open loop Closed loop

Open loop is an initial estimate of the power the mobile needs to transmit to the BTS. Closed loop is a refinement of the open loop estimate. Open loop power control Open loop is the mobile's estimate of the power at which it should transmit. The open loop estimate is based on the strength of the pilot signal the mobile receives. As the pilot signal gets weaker or stronger, the mobile adjusts its transmission strength upwards or downwards. Open loop is used any time the mobile transmits.

22

Closed loop power control In closed loop, the BTS sends a command to the mobile to increase or decrease the strength at which it is transmitting. The BTS determines this command based on the quality of the signal it receives from the mobile. Closed loop is only used during a call. Closed loop commands are sent on the forward traffic channel.

23

Forward link power control The BTS independently adjusts the power for each forward traffic channel based on the information it receives from the mobile.

Rake Receiver The rake receiver is a CDMA feature that turns what is a problem in other technologies into an advantage for CDMA.

24

The multi-path problem Signals sent over the air can take a direct path to the receiver, or they can bounce off objects and then travel to the receiver. These different paths, called multi-paths, can result in the receiver getting several versions of the same signal but at slightly different times. Multi-paths can cause a loss of signal through cancellation in other technologies.

How the rake receiver works CDMA's rake receiver is multiple receivers in one. The rake receiver identifies the three strongest multi-path signals and combines them to produce one very strong signal. The rake receiver therefore uses multipath to reduce the power the transmitter must send. Both the mobile and the BTS use rake receivers. Handoff in CDMA Handoff is the process of transferring a call from one cell to another. This is necessary to continue the call as the phone travels. CDMA is unique in how it handles handoff.

25

Types of CDMA handoff CDMA has three primary types of handoff: hard soft idle

The type of handoff depends on the handoff situation. Soft handoff A soft handoff establishes a connection with the new BTS prior to breaking the connection with the old one. This is possible because CDMA cells use the same frequency and because the mobile uses a rake receiver. The CDMA mobile assists the network in the handoff. The mobile detects a new pilot as it travels to the next coverage area. The new base station then establishes a connection with the mobile. This new communication link is established while the mobile maintains the link with the old BTS. Soft handoffs are also called "make-before-break."

Variations of the soft handoff There are two variations of soft handoffs involving handoffs between sectors within a BTS: Softer Soft-softer

26

The softer handoff occurs between two sectors of the same BTS. The BTS decodes and combines the voice signal from each sector and forwards the combined voice frame to the BSC. The soft-softer handoff is combination handoff involving multiple cells and multiple sectors within one of the cells.

CDMA hard handoff A hard handoff requires the mobile to break the connection with the old BTS prior to making the connection with the new one. CDMA phones use a hard handoff when moving from a CDMA system to an analog system because soft handoffs are not possible in analog systems. A Pilot Beacon Unit (PBU) at the analog cell site alerts the phone that it is reaching the edge of CDMA coverage. The phone switches from digital to analog mode as during the hard handoff. Hard handoffs are also called "break-before -make."

27

When does CDMA use a hard handoff? The CDMA hard handoff may be used when moving from a CDMA network to an analog one. It may also be used when moving to a different: RF channel MTSO Carrier Market

Analog to CDMA handoff is not available due to the limitations of analog technology.

CDMA idle handoff An idle handoff occurs when the phone is in idle mode. The mobile will detect a pilot signal that is stronger than the current pilot. The mobile is always searching for the pilots from any neighboring BTS. When it finds a stronger signal, the mobile simply begins attending to the new pilot. An idle handoff occurs without any assistance from the BTS.

28

TDMA and FDMA handoff TDMA and FDMA systems use a hard handoff when the mobile is moving from one cell site to another. These technologies do not allow for any type of make-before -break handoff. A hard handoff can increase the likelihood of a dropped call.

Advantages of CDMA CDMA technology has numerous advantages including: Coverage Capacity Clarity Cost Compatibility Customer satisfaction

Coverage CDMA's features result in coverage that is between 1.7 and 3 times that of TDMA: Power control helps the network dynamically expand the coverage area. Coding and interleaving provide the ability to cover a larger area for the same amount of available power used in other systems.

29

Capacity CDMA capacity is ten to twenty times that of analog systems, and it's up to four times that of TDMA. Reasons for this include: CDMA's universal frequency reuse CDMA users are separated by codes, not frequencies Power control minimizes interference, resulting in maximized capacity. CDMA's soft handoff also helps increase capacity. This is because a soft handoff requires less power.

Clarity Often CDMA systems can achieve "wireline" clarity because of CDMA's strong digital processing. Specifically: The rake receiver reduces errors The variable rate vocoder reduces the amount of data transmitted per person, reducing interference. The soft handoff also reduces power requirements and interference. Power control reduces errors by keeping power at an optimal level. CDMA's wide band signal reduces fading. Encoding and interleaving reduce errors that result from fading.

30

Cost CDMA's better coverage and capacity result in cost benefits: Increased coverage per BTS means fewer are needed to cover a given area. This reduces infrastructure costs for the providers. Increased capacity increases the service provider's revenue potential. CDMA costs per subscriber has steadily declined since 1995 for both cellular and PCS applications.

Compatibility CDMA phones are usually dual mode. This means they can work in both CDMA systems and analog cellular systems. Some CDMA phones are dual band as well as dual mode. They can work in CDMA mode in the PCS band, CDMA mode in the cellular band, or analog mode in an analog cellular network.

Customer satisfaction CDMA results in greater customer satisfaction because CDMA provides better: Voice quality Longer battery life due to reduced power requirements No cross-talk because of CDMA's unique coding Privacy--again, because of coding.

31

http://www.cdg.org/tech/a_ross/Forward.asp Forward CDMA Channel

The FORWARD CDMA CHANNEL is the cell-to-mobile direction of communication. It carries traffic, a pilot signal, and overhead information. The pilot is a spread, but otherwise unmodulated DSSS signal. The pilot and overhead channels establish the system timing and station identity. The pilot channel also is used in the mobile assisted handoff (MAHO) process as a signal strength reference. Frequency Plan The base station transmit frequency is 45 MHz above the mobile station transmit frequency in the cellular service (IS-95A), and 80 MHz above in the PCS service (ANSI J-STD-008). Permissible frequency assignments are on 30 kHz increments in cellular and 50 kHz in PCS. See Frequency Plans for further details. Transmission Parameters The IS-95A forward link currently supports a 9600 bps rate family in the three data-bearing channel types, as shown in the table. In all cases the FEC code rate is 1/2 and the PN rate is 1.2288 MHz. Note that 1.2288 MHz = 128*9600 bps. Forward Link Channel Parameters, Rate Set 1 Channel Sync Paging Traffic Data rate Code repetition Modulation symbol rate 1200 4800 9600 1200 2400 4800 9600 2 2 1 8 4 2 1 4800 19,200 19,200 19,200 19,200 19,200 19,200 64 256 64 128 64 1024 64 512 64 256 64 128 bps sps

PN chips/modulation symbol 256 PN chips/bit 1024

J-STD-008 supports, in addition to the above rates, a second traffic channel rate family with a maximum rate of 14,400 bps. This is termed Rate Set 2, the original 9600 bps family being Rate Set 1. Rate Set 2 uses an FEC code rate of 3/4, created by puncturing the code used in Rate Set 1.

Forward Link Channel Parameters, Rate Set 2 Channel Traffic Data rate 1800 3600 7200 14400 Code repetition 8 4 2 1 Modulation symbol rate 19,200 19,200 19,200 19,200 PN chips/modulation symbol 64 64 64 64 PN chips/bit 682.67 341.33 170.67 85.33

bps sps

32

Signal Structure Channelization The forward link consists of up to 64 logical channels (code channels). The channels are independent in that they carry different data streams, possibly at different rates, and are independently adjustable in amplitude (Cf. Reverse Channelization). Coding and Interleaving The figure shows the core processing that generates one forward code channel, rate set 1. Rate set 2 is identical except the coding rate is 3/4 rather than 1/2, yielding the same code symbol rate with 3/2 times the data rate.

Walsh Codes The code channels, as transmitted, are mathematically orthogonal. The orthogonality is established by covering the FEC code symb ols with one of a set of 64 so-called Walsh functions. "Mutually orthogonal" means that their cross correlations are small (ideally zero). Because only whole periods of the Walsh functions occur in each code symbol, the effect of the Walsh cover is to make the channels completely separable in the receiver, at least in the absence of multipath. The orthogonality not only means that there is no co-mingling of channels, it means there is no interference between users in the same cell, again in the absence of multipath. This has a substantial beneficial effect on the forward link capacity. Multipath delay spread that exceeds a chip does introduce mutual interference between users in one cell. In any particular Rake finger the uncorrelated channels contribute an effective interference level. This level varies from zero, when there is only one multipath component, up to (N-1)/N of the total signal power if there are N discrete, equal-amplitude multipath comp onents. Note that one of the Walsh functions is always a constant, code number zero, by the numbering convention. This channel is always reserved to serve as the pilot.

33

Spreading Each forward code channel is spread by the Short Code, which has I- and Q-components. The spreading is thus quadrature. That is, from a single binary-valued, covered, symbol stream, two binary sequences are generated by mod 2 addition of the short code PN sequences , as shown in the figure.

Modulation The two coded, covered, and spread streams are vector-modulated on the RF carrier. The spreading modulation is thus QPSK, superimposed on a BPSK code symbol stream. The spectrum shaping of the forward link is carefully prescribed in the IS-95A air interface and the IS-97 performance specification. The latter is in terms of the so-called Rho meter, a measurement of the correlation between the actual transmitter output with the ideal transmitter output. The air interface also specifies a slightly nonlinear phase characteristic the purpose of which is partial pre -equalization of the mobile receiver. In-band ripple is specified as less than 1.5 dB. Stopband rejection is 40 dB beginning 740 kHz from band center. An equi-ripple, 48 tap FIR baseband filter is suggested, although not required. Overhead channels There are three types of overhead channel in the forward link: pilot, sync, and paging. The pilot is required in every station. Pilot Channel The pilot channel is always code channel zero. It is both a demodulation reference for the mobile receivers, and for handoff level measurements, and thus must be present in every station. It carries no information. It is pure short code, with no additional cover or information content. The amplitude of the pilot and its spatial distribution must be carefully controlled because their relative amplitudes control handoff boundaries between stations. The PNI and PNQ modified linear feedback shift register sequences that comprise the short code have period 215 chips, which is 80/3 = 26.667 ms at the 1.2288 MHz chip rate. All stations use the same short code, and thus have the same pilot waveform. They are distinguished from one another only by the phase of the pilot. The short period of the short code, 215 , facilitates rapid pilot searches by the mobiles. The air interfaces stipulate that pilot phases always be assigned to stations in multiples of 64 chips, giving a total of 215-6 = 512 possible assignments. The 9-bit number that identifies the pilot phase assignment is called the Pilot Offset.

34

Sync Channel The sync channel carries a repeating message that identifies the station, and the absolute phase of the pilot sequence. The data rate is always 1200 bps. The interleaver period is 80/3 = 26.667 ms, equal to the period of the short code. This simplifies finding frame boundaries, once the mobile has located the pilot. The Sync Channel carries a single, repeating message that conveys the timing and system configuration information to the mobile station. The mobile station can derive accurate system time, to within a fraction of a spreading chip (833 ns) by synchronizing to the short code. The short code synchronization and the pilot offset, which is part of the sync message, fix system time modulo 26.667 ms. The code period ambiguity is then resolved by the long code state and system time fields that are also part of the sync message. Paging Channel The paging channel is the vehicle for communicating with mobile stations when they are not assigned to a traffic channel. As the name implies, its primary purpose is to convey pages, that is, notifications of incoming calls, to the mobile stations. It carries the responses to mobile station accesses, both page responses and unsolicited originations. Successful accesses are normally followed by an assignment to a dedicated traffic channel. Once on a traffic channel, signaling traffic between base and mobile can continued interspersed with the user traffic. The paging channel may run at either 4800 or 9600 bps. Each base station must have at least one paging channel per sector, on at least one of the frequencies in use. All paging can be done on one frequency, or it can be distributed over multiple frequencies. Traffic Channel Traffic channels are assigned dynamically, in response to mobile station accesses, to specific mobile stations. The mobile station is informed, via a paging channel message, which code channel it is to receive (it is tempting, but inappropriate to use the word "tune"!). The traffic channel always carries data in 20 ms frames. Frames at the higher rates of Rate Set 1, and in all frames of Rate Set 2, include CRC codes to help assess the frame quality in the receiver. Soft Handoff During soft handoffs each base station participating in the handoff transmits the same traffic over its assigned code channel. The code channel assignments are independent, and in general will be different in each cell. Whatever code channels are not in use for overhead channels are available, up to either a total of 64 or the available equipment limit, whichever is smaller. Rate Traffic channels carry variable rate traffic frames, either 1, 1/2, 1/4, or 1/8 of the maximum rate. In IS-95A only a 9600 bps rate family is currently available in the standard. In J-STD-008 a second rate set, based on a maximum rate of 14,400 bps is available. The Rate Set 2 will be added in a future revision of IS-95. The rate variation is accomplished by 1, 2, 4, or 8-way repetition of code symbols. Transmission is continuous, with the amplitude reduced at the lower rates so as to keep the energy per bit approximately constant, regardless of rate. The rate is independently variable in each 20 ms frame.

35

Power Control Subchannel The 800 bps reverse link power control subchannel is carried on the traffic channel by puncturing 2 from every 24 symbols transmitted. The punctured symbols both carry the same power control bit, so they can be coherently combined by the receiver. Each base station participating in a soft handoff makes its own power control decision, independent of the others, unless they are different sectors of the same cell, in which case they all transmit a common decision. This special circumstance is made known to the mobile when the handoff is set up. Timing All base stations must be synchronized within a few microseconds for the station identification mechanisms to work reliably and without ambiguity. Any convenient mechanism can be used for this purpose, but the system was designed under the assumption that the Global Positioning System (GPS) would be used. This is a family of lowearth-orbit satellites that broadcast a spread-spectrum signal and ephemeris information from which a sophisticated Kalman filter algorithm in a receiver can derive both a very accurate position and a very accurate time.

36

http://www.cdg.org/tech/a_ross/Reverse.asp Reverse CDMA Channel

The REVERSE CDMA CHANNEL is the mobile-to-cell direction of communication. It carries traffic and signaling. Any particular reverse channel is active only during calls to the associated mobile station, or when access channel signaling is taking place to the associated base station. Frequency Plan The mobile station transmit frequency is 45 MHz below the base station transmit frequency in the cellular service (IS-95A), and 80 MHz below in the PCS service (ANSI J-STD-008). Permissible frequency assignments are on 30 kHz increments in cellular and 50 kHz in PCS. See Frequency Plans for further details. Transmission Parameters The IS-95A Reverse CDMA Channel currently supports a 9,600 bps rate family in the Access Channel and Traffic Channels, as shown in the table. The transmission duty cycle varies with data rate. In all cases the FEC code rate is 1/3, the code symbol rate is always 28,800 symbols per second after there are 6 code symbols per modulation symbol, and the PN rate is 1.2288 MHz. The modulation is 64-ary orthogonal, using the same Walsh functions that are used in the forward link for channelization. Each period of the Walsh function is repeated for four chips of the PN code. The Walsh symbol rate is thus 1.2288 MHz/(4 chips per Walsh chip)/(64 Walsh chips per Walsh symbol) = 4,800 modulation symbols per second. Note that 1.2288 MHz = 128*9,600 bps. Reverse CDMA Channel Parameters, Rate Set 1 Channel Access Traffic Data rate Code Rate Symbol Rate before Repetition Symbol Repetition Symbol Rate after Repetition Transmit Duty Cycle Code Symbols/Modulation Symbol PN Chips/Modulation Symbol PN chips transmitted/bit 4,800 1,200 2,400 4,800 9,600 1/3 1/3 1/3 1/3 1/3 14,400 3,600 7,200 14,400 28,800 2 8 4 2 1 28,800 28,800 28,800 28,800 28,800 1 1/8 1/4 1/2 1 6 256 256 6 256 128 6 256 128 6 256 128 6 256 128 bps sps sps

37

J-STD-008 supports, in addition to the above rates, a second traffic channel rate family with a maximum rate of 14,400 bps. This is termed Rate Set 2, the original 9600 bps family being Rate Set 1. Rate Set 2 uses a rate 1/2 code in place of the rate 1/3 code of Rate Set 1. Reverse CDMA Traffic Channel Parameters, Rate Set 2 Channel Traffic Data rate 1,800 3,600 7,200 14,400 bps Code Rate 1/2 1/2 1/2 1/2 Symbol Rate before Repetition Symbol Repetition Symbol Rate after Repetition Transmit Duty Cycle Code Symbols/Modulation Symbol PN Chips/Modulation Symbol PN chips transmitted/bit Signal Structure Channelization The Reverse CDMA Channel consists of 242 -1 logical channels. One of these logical channel is permanently and uniquely associated with each mobile station. That logical channel is used by the mobile whenever it passes traffic. The channel does not change upon handoff (Cf. Forward Channelization). Other logical channels are associated with base stations for system access. This reverse link addressing is accomplished through manipulation of period 242 -1 Long Code, which is part of the spreading process. Coding and Interleaving Figure 1 shows the core processing that generates one Reverse CDMA Channel. 3,600 7,200 14,400 28,800 8 4 2 1 28,800 28,800 28,800 28,800 1/8 6 256 1/4 6 256 1/2 6 256 1 6 256 sps sps

256/3 256/3 256/3 256/3

Figure 1. Reverse CDMA Channel signal generation. Separation of Users The reverse CDMA Channel, in contrast to the Forward CDMA Channel, does not use strict orthogonality in any sense to separate logical channels. Rather, it uses a very long period spreading code, in distinct phases. The correlations between stations are not zero, but they are acceptably small.

38

Orthogonal Modulation Reverse link data modulation is 64-ary orthogonal, and is applied prior to the spreading. Groups of six code symbols select one of 64 orthogonal sequences. The 64-ary orthogonal sequences are the same Walsh functions that are used in the Forward CDMA Channel, here are used for a totally different purpose. Each period of the Walsh sequence (a Walsh Chip) is four PN chips in duration. The modulation symbol rate is thus always 4,800 sps. Spreading Each Reverse CDMA Channel is spread by both the channel-unique Long Code and the Short Code, which has Iand Q-components. The spreading is thus quadrature. That is, from a single binary-valued symbol stream, two binary sequences are generated by mod 2 addition of the short code PN sequences (Figures 1 and 2). The effect of adding long and short codes is to produce a supersequence that has an extraordinarily long period, about 257 , or 3700 years at the 1.2288 MHz spreading rate (see Reverse Spreading).

Figure 2. Reverse CDMA Channel modulation. RF Modulation The two coded, covered, and spread streams are vector-modulated on the RF carrier. The Q-axis modulation is delayed by 1/2 chip. The spreading modulation is thus offset QPSK. Offset modulation was chosen in an effort to reduce the envelop modulation of the RF signal and reducing performance requirements on the power amplifiers in the subscriber station. The spectrum shaping of the reverse link is carefully prescribed in the IS-95A air interface and the IS-98 performance specification. The latter is in terms of the so-called Rho meter, a measurement of the correlation between the actual transmitter output with the ideal transmitter output. The air interface also specifies a slightly nonlinear phase characteristic the purpose of which is partial pre -equalization of the mobile receiver. In-band ripple is specified as less than 1.5 dB. Stopband rejection is 40 dB beginning 740 kHz from band center. An equi-ripple, 48 tap FIR baseband filter is suggested, although not required.Access channel There is only one type of overhead channel in the Reverse CDMA Channel: the Access Channel. The Access Channel is the vehicle for communicating with mobile stations when they are not assigned to a traffic channel. As the name implies, its primary purpose is to service originations and page responses by the mobile stations. Successful accesses are normally followed by an assignment to a traffic channel. Once on a traffic channel, signaling traffic between base and mobile can continued interspersed with the user traffic. The access channel always runs at 4800 bps. Each base station must service at least one Access Channel, on at least one of the frequencies in use. The Long Code Mask for the Access Channel is derived from the station identity, the paging channel number with which the access channel is associated, and Access Channel number within that base station.

39

Traffic Channel Traffic channels are in the Reverse CDMA Channel are mobile-unique. That is, each station has a unique Long Code Mask, based on its electronic serial number. Whenever the mobile is assigned to traffic, it uses its specific long code mask. The traffic channel always carries data in 20 ms frames. Frames at the higher rates of Rate Set 1, and in all frames of Rate Set 2, include CRC codes to help assess the frame quality in the receiver. Soft Handoff During soft handoffs a mobile station transmits the same information that it would in the absence of the handoff, that is, there is no change in the content of the mobile transmissions. There is a change in the way that reverse link power control is applied to the reverse link. Power control "down" commands from all participants in the handoff are logically "or'ed" together - if any of the handoff participant base stations says "down," then the mobile station is required to reduce its reverse link power. Rate Traffic channels carry variable rate traffic frames, either 1, 1/2, 1/4, or 1/8 of the maximum rate. In IS-95A only a 9600 bps rate family is currently available in the standard. In J-STD-008 a second rate set, based on a maximum rate of 14,400 bps is available. The Rate Set 2 will be added in a future revision of IS-95. The data rate variation is accomplished by varying the duty cycle of transmission in accordance with a 1, 1/2, 1/4, 1/8 plan, according to the rate requested by the data source. Transmission always occurs in 1.25 ms segments. The segments that are actually transmitted are pseudo-randomly selected, using a decimated long code sequence. Timing Mobile stations are required to adjust their transmission time according to the timing that they are able to derive from the pilot and sync channels, and adjusted for the know base station pilot offset. That is, the pilot offset index read from the sync channel must be used to correct transmitter timing so that the Reverse CDMA Channel signal arrives at the base station aligned in time with system time, to within a round-trip of the air interface propagation delay.

40

Frequency Cellular (IS -95A)

Plans

CDMA cellular service is intended to share the existing AMPS spectral allocation, shown below.

Consecutive AMPS channels are spaced by 30 kHz. CDMA stations are permitted to operate on any AMPS channel, except for guard bands at the edges of the allocations. CDMA stations, of course, would normally be assigned channel at least 1.25 MHz apart (about 42 channels). The mobile station transmit frequency is always 45 MHz lower than the base station transmit frequency. Both A and B operators have 12.5 MHz of spectrum in each direction. Each allocation, however, is split, and the splits are not the same for the two operators, as shown in the figure. Note that the A' and B' allocations present problems, both for the RF hardware design, and for the allocation of CDMA channels. The B' band, in particular, accommodates two CDMA channels only if they are overlapped slightly, at some small loss of capacity. PCS (J-STD-008) PCS is allocated 60 MHz total in each direction, as three 15 MHz bands plus three 5 MHz bands, shown below.

Consecutive frequency assignments are spaced by 50 kHz. Assignments near band edges are conditional, depending on whether the neighboring bands are held by the same operator. Operation near the edges of the service is forbidden in 1.2 MHz guard bands. In contrast to the cellular service, the standard suggests particular channel numbers as preferred CDMA frequency assignments as follows. CDMA Preferred Frequency Assignments Band Preferred Channels A 25, 50, 75, 100, 125, 150, 175, 200, 225, 250, 275 D B E F C 325, 350, 375 425, 450, 475, 500, 525, 550, 575, 600, 625, 650, 675 725, 750, 775 825, 850, 875 925, 950, 975, 1000, 1025, 1050, 1075, 1100, 1125, 1150, 1175

41

Walsh

Functions

Hadamard-Walsh functions are binary orthogonal sequences, with power-of-two lengths. They can be generated by the recursion

where

and M is a power of two. The rows of any instance form a mutually orthogonal set over the inner product. The sequences of order 8, for example, are:

. The Walsh functions of order 64 are used as orthogonal cover on the forward link channels. They are also used as orthogonal modulation symbols in the reverse link. Although the functions are the same, they are used for entirely different purposes in the forward and reverse links.

42

Long Code The long code is a period 242 -1 LFSR sequence that is used for spreading the reverse link. There is only one long code sequence. Different stations are distinguished not by the sequence itself but by its relative phase. The fact that the long code is added to each of the two (I and Q) short code sequences ensures that cross correlations between the signals from distinct stations are always small. The long code LFSR tap polynomial is:

The different phases of the long code are generated by use of one of the well-known properties of LFSR sequences. Any modulo-2 sum of different phases of a LFSR sequence gives a third phase of that same sequence. A corollary of this property is the fact that all internal nodes of any LFSR generator also run through the same sequence as the generator output, but with different phases. This addition property of LFSR sequences is exploited in the long code generation process for the reverse link spreading. A 42-bit number, the Long Code Mask, is used to select particular bits of the 42-bit long code generator register. The selected nodes are summed, modulo 2. The resultant of the sum, that is, the modulo-2 inner product of the generator state with the mask, is the generator output corresponding to that mask.

Figure 1. Long Code Mask logic.

It can be shown that each distinct mask results in a distinct phase. The Long Code Mask thus serves as a reverse link address.

43

Short

Code

The short code is a pair of period 215 sequences that are used for spreading the CDMA Forward Channel. They are also used in conjunction with the Long Code for spreading the CDMA Reverse Channel. They are both derived from period 215 -1 LFSR sequences, augmented with an extra zero bit to bring the length to an even power of two. The short code LFSR tap polynomials are, for the I-sequence (1) and for the Q-sequence . (2)

The extra zero bit is inserted in each sequence immediately after the occurrence of 14 consecutive zeros from the generator register. This occurs once per period. At the spreading rate of 1.2288 MHz the period of the short code is 80/3 = 26.666... ms.

44

Reverse

Spreading

There are two different criteria that apply to the reverse link spreading. When a mobile is engaged in user traffic, i.e., in a conversation, it is desirable that that mobile use a unique code that is distinct from all others. A mobile unique code, rather than a base-station-associated code, facilitates handoff. With a mobile -unique code, nothing needs to change about the mobile's modulation or coding when handoff occurs. The second situation occurs when a mobile is attempting to gain the attention of a base station. Initially the base station has no knowledge that any particular mobile is in its service area. It is wildly impractical for each base station to search simultaneously for millions of potential subscriber codes. For these initial accesses, or any other non-traffic uses of the air interface, it is desirable to have some Reverse Spreading codes that are base stationassociated. If there are only a few associated codes for each base station, then it is practical for the base station to search for them continuously and simultaneously, awaiting the arrival of any user who wants service. Traffic Channel Spreading Spreading is based on a single, universal 42-bit Long Code LFSR sequence. In a sense, the reverse spreading code borrows an idea from the Forward Spreading. Rather than completely unique codes, a single, universal 42-bit LFSR sequence is used. This sequence is known as the Long Code. Again, phase of the long code is used to distinguish stations. All 242 -1 possible phases are available as logical addresses. The distinct phases of the Long Code are generated by a simple means that takes advantage of one of the curious properties of maximal-length LFSR sequences. A 42-bit Long Code Mask (LCM) selects particular cells of the generator register. The output is the modulo-2 sum of the contents of the selected cells. It turns out, oddly, that any such linear combination of any cells of the generator runs through the same sequence, but with a different phase. All 242 -1 phases that can be generated by a non-zero 42-bit mask are distinct (the all-zero mask gives a constant zero output). The LCM for each mobile station is based on the mobile manufacturer's identity and the serial number of that unit. The LCM is, effectively, a logical address of the Reverse CDMA Channel. The period of the long code, for what it's worth, is 3.6 million seconds, or about 41 days! Short code is added to resolve ambiguities. Even though there are a large number of Long Code phases, without some modification, there is potential ambiguity between units that have similar LCMs. Timing errors in real systems will be of the order of a round trip delay to the farthest cell, in practice perhaps in excess of 100 microseconds. This would make groups of LCMs ambiguous if the Long Code were the only component to the spreading. A delay of X microseconds would be indistinguishable from a long code that was offset by X microseconds. For this reason the Long Code is modified by adding the Short Code. Each mobile, when it acquires a candidate base station, synchronizes its Short Code and its Long Code generators to System Time. The mobile applies its unique LCM to the long code generator, and modulo-2 adds the output, that is the unique-phase long code, to the universal short code. As in the Forward CDMA Channel, the spreading modulation is quadrature, so as to homogenize the phase of the interference. Again, both short code sequences are used. See Figure 1.

Figure 1. Reverse CDMA channel spreading logic.

45

Astute readers might note that the addition of the Short Code to the Long Code is not without residual ambiguity. The combined "supersequence" code is still periodic, but with a larger period. The periods of the short and long codes are relatively prime, so the period of the supersequence is the product of the original periods: (242 -1)*215 = approximately 257 , or about 3700 years. Any offset long code sequence, added to the short code sequence, is some different phase of the supersequence. Why is this not a problem? Although the supersequence has period 257 , we are only ever using, at most, 242 -1 of the possible phases of it. We have constructed the spreading sequence generator in such a way that the smallest code phase separation between users is one short code period. One is more than enough. The mobile will have system time in error by at most a fraction of one millisecond due to propagation delays. Although everyone is using the same supersequence, their phases are so far apart that they will never be confused by the base stations. This is the contribution of the short code to the process: it changes the time ambiguity from one chip displacement to 215 chips displacement. The number of Long Code phases, 242 -1, is about 4.4 * 1012 , so the number of available addresses should suffice for the near future! Actually, because of the way they are assigned, which includes space for the access channels, the available addresses are a bit less than this, but there is still plenty of space. See Long Code Mask for further details. Access Channel Spreading Access channel spreading is generated by the same logic as the mobile -unique traffic channel spreading. The difference is that the Long Code Mask is associated with the base station, not the mobile that is using it. The LCM for an access channel is derived from a prefix that basically says "this is an access channel", the base station identity, and an integer index. The index selects one of the several possible access channels that are supported by this station. The LCMs that are assigned to each base station are shared by all mobiles. That is, any mobile station desiring to use that base station must use one of the base station's assigned LCMs. Each base station must support an access channel on at least one RF carrier of each sector. Each station is permitted, but not required, to support more than one access channel per sector. If more than one access channel is available then the one that a mobile uses is chosen through a Resource Hashing algorithm. The hashing ensures that the load on the available channels is statistically uniform. A mobile attempting to gain access to a base station first identifies the base station by decoding the Sync Channel message and later acquires the information it needs to do system accesses by reading the Access Parameters Message on a paging channel. These message contains all information needed to construct the appropriate long code mask, including the identity of the base station and the number of access channels that it supports. See System Access for further details.

46

Long

Code

Mask

The long code mask is a 42-bit number that serves as a logical address for Reverse CDMA Channel spreading codes. It is used to select specific bits from the long code linear feedback shift register to be added, modulo-two, in order to produce the actual long code, at the proper phase. This method is taking advantage of one of the peculiar properties of maximal length shift register sequences (see the LFSR page). The logic is shown in Figure 1.

Figure 1. Long Code Mask logic. When transmitting on an access channel, a mobile station constructs the mask from the base station identity, the paging channel number, and the access channel number. When transmitting traffic, a mobile station uses a mask constructed from its Electronic Serial Number.

Figure 2. Long code mask contents. Linear Feedback Shift Registers Ideally the spreading codes used in direct sequence spread spectrum systems would be truly random binary sequences, as might be produced by consecutive tosses of an unbiased, memoryless coin. This is not practical. Both transmitter and receiver must generate the same sequence, time-aligned, in order to communicate with one another. Receivers thus must perform synchronization searches by changing their time offset hypothesis until the transmitter timing is located. Achieving high capacity in the CDMA environment also requires that the spreading rate be high: 1.2288 MHz in IS-95A CDMA. If truly random sequences were to be used then they would have to be pre-generated and pre-stored in all transmitters, with a matching copy in all receivers. Deterministic methods of generating the pseudo-random sequences are preferable. The CDMA air interfaces use linear feedback shift register (LFSR) generators for this purpose. The maximal-length binary sequences produced by linear feedback shift registers are widely used for direct sequence spectrum spreading. There are several reasons for this: LFSR sequences are easily generated by very simple binary logic circuits. Very high speed generators are possible because of the simple logic. Maximal length sequence generators are easily designed using finite (Galois) field mathematics. The full period autocorrelation functions of maximal-length LFSR sequences are binary-valued, facilitating synchronization searching.

47

It is easily shown, by the way, that any linear feedback binary state machine that generates a maximum length output sequence must be equivalent to some maximal length LFSR. See, for example, Peterson, et al. Linear feedback shift registers can be implemented in two ways (Figure 1). The so-called Fibonacci implementation consists of a simple shift register in which a binary weighted modulo 2 sum of the taps is fed back to the input.

The Galois implementation consists of a shift register, the contents of which are modified at every step by a binary weighted value of the output stage.

If the tap weights are identical, and configured as shown in the figures, then the two implementations will produce exactly the same sequence (this can be verified by simple arguments). Initial conditions required to produce the same phase of the sequence are obviously not identical, however. There are actually two sequences produced by each of these generators. One is the trivial one, of length one, that occurs in both cases when the initial state of the generator is all zeros. The other, the useful one, has length 2m-1. Together these two sequences account for all 2m states of the m-bit state register. The mathematics of these generators is equivalent to the operation of ordinary algebra applied to abstract polynomials over an indeterminate X, with binary-valued coefficients. This is a finite (Galois) field of order 2m . Each sequence is based on a generator polynomial

whose coefficients are binary, and are the weights shown in the figures. The polynomial is said to be primitive if it does not factor and it divides Xr+1, where r=2m-1. A primitive polynomial of degree m necessarily has g m = g 0 = 1. If the generator polynomial in a LFSR is primitive then the sequence produced by that generator has maximum length, which is 2 m-1. Maximal length sequences are sometimes called m-sequences. For a discussion of the mathematics of finite fields see, for example, Golomb . Properties Some properties of m-sequences: 1. An m-bit register produces a sequence of period 2 m-1. 2. An m-sequence contains exactly 2 m-1 ones and 2m-1 -1 zeros. 3. The sum, modulo 2 of an m-sequence and another phase of the same m-sequence yields a third phase of the

48

sequence. 3a. (A corollary of 3) Each node of the generator of an m-sequence runs through some phase of the sequence. 4. A sliding window of length m, passed along an m-sequence for 2 m-1 positions, will span every possible m-bit number except all zeros once and only once. 5. Define a run of length r to be a sequence of r consecutive identical symbols, bracketed by non-equal symbols. Then in any m-sequence there are: 1 run of ones of length m 1 run of zeros of length m-1 1 run of ones and 1 run of zeros, each of length m-2 2 runs of ones and 2 runs of zeros, each of length m-3 4 runs of ones and 4 runs of zeros, each of length m-4 ... 2m-3 runs of ones and 2m-3 runs of zeros, each of length 1

If the sequence is mapped to a binary valued waveform by mapping a binary zero to -1 and binary 1 to +1, then the autocorrelation is unity for zero delay, and -1/N at all other times. Other curiosities: If g(X) is a primitive polynomial of degree m, then X mg(X-1 )=g'(X) is also a primitive polynomial. This is easily shown using the definition of a primitive polynomial, and the properties of modulo-2 arithmetic. A little thought shows that the register taps of g'(X) are the mirror image of those of g(X). And the sequence produced by the tapreversed generator is the time reversal of the original. This seems like it ought to be almost trivially obvious: if time is reversed in a generator, then the arrows in the block diagram are reversed, and in the taps the flow reversal doesn't change the values on any of the three connections! Example Consider as an example, polynomials of degree 3, so that maximal length sequences are 7 bits long. The primitive polynomials of degree 3 are found as the non-trivial factors of Xr +1, where r=2 m-1=7:

The two degree 3 polynomial factors are, as noted above, mirror images of one another. Choosing the first as g(X), the generator is:

The length 7 sequence produced by this generator, starting from the state [001] is {1011100...}. The other, mirror image, generator

produces {1110100} starting from the state [001].

49

Offsets and Time Shifts of m-Sequences The linearity of the m-sequence generators and their properties as a representation of a finite field make it rather simple to offset a state by some prescribed number of states, or to create a transformation matrix that will produce a delayed version of the sequence from an undelayed state register. The Galois generator, in particular, can be regarded as a counter in a Galois field. Counting is equivalent to multiplication of the generator state by a primitive element of the field. The multiplication is equivalent to ordinary matrix multiplication of the state, regarded as a column matrix, by a transformation matrix T.

General offsets, say by k states, can be accomplished by calculating Tk and then multiplying the initial state by Tk . In particular, power-of-two powers of T can be pre-calculated, and an arbitrary offset accomplished my multiplying by the powers of T corresponding to the one bits in the binary representation of k.

50

Spread Spectrum (SS) introduction ir. J. Meel

jme@denayer.wenk.be Studiedag Spread Spectrum - 6 okt. 99

51

CONTENTS 1. DEFINITION OF SPREAD SPECTRUM (SS)............................................................................................... 2. BASIC PRINCIPLE OF SPREAD SPECTRUM SYSTEMS: DSSS AND FHSS ....................................... 3. BASIC PRINCIPLE OF DIRECT SEQUENCE SPREAD SPECTRUM (DSSS ) ...................................... 3.1 3.2 M ODULATION .......................................................................................................................................... DEMODULATION...................................................................................................................................... 3.2.1 pnr = pnt................................................................................................................................................. 3.2.2 pnr pnt ................................................................................................................................................. 4. PERFORMANCE IN THE PRESENCE OF INTERFERENCE .................................................................. 4.1 NARROWBAND INTERFERENCE............................................................................................................. 4.2 WIDEBAND INTERFERENCE .................................................................................................................... 5. PSEUDO-NOISE SEQUENCES PN................................................................................................................. RANDOM WHITE GAUSSIAN NOISE ............................................................................................................. PSEUDO-RANDOM NOISE............................................................................................................................... 5.3 PROPERTIES OF PN SEQUENCES.............................................................................................................. TYPES.................................................................................................................................................................... 5.4.1 m-sequence .............................................................................................................................................. 5.4.2 Barker Code............................................................................................................................................. 5.4.3 Gold Codes.............................................................................................................................................. Hadamard-Walsh Codes................................................................................................................................... 6. TRANSMITTER ARCHITECTURE.............................................................................................................. 7. RECEIVER ARCHITECTURE....................................................................................................................... 8. PN DECORRELATORS.................................................................................................................................... 8.1 PN MATCHED FILTER ............................................................................................................................... 8.2 PN A CTIVE CORRELATOR (INTEGRATE AND DUMP)........................................................................ 9. PN SYNCHRONIZATION .............................................................................................................................. 9.1 A CQUISITION PHASE (COARSE SYNCHRONIZATION) ...................................................................... 9.1.1 Serial Synchronization (Sliding Correlator) ................................................................................................ 9.1.2 Serial/Parallel Synchronization................................................................................................................... 9.2 TRACKING PHASE (FINE SYNCHRONIZATION)................................................................................... 10. MULTIPLE ACCESS ..................................................................................................................................... 11. MULTIPATH CHANNELS ........................................................................................................................... 12. JAMMING........................................................................................................................................................ 13. ISM BANDS..................................................................................................................................................... 14. EVALUATION OF SS ................................................................................................................................... 15. REFERENCES..................................................................................................................................................

52

1. Definition of Spread Spectrum (SS)

A transmission technique in which a pseudo-noise code, independant of the information data, is employed as a modulation waveform to spread the signal energy over a bandwidth much greater than the signal information bandwidth. At the receiver the signal is despreading using a synchronized replica of the pseudo-noise code.

2. Basic principle of Spread Spectrum Systems: DSSS and FHSS

Direct Sequence Spread Spectrum (DSSS)

A pseudo-noise sequence pnt generated at the modulator, is used in conjunction with an M-ary PSK modulation to shift the phase of the PSK signal pseudorandomly, at the chipping rate Rc (=1/Tc) a rate that is an integer multiple of the symbol rate Rs (=1/Ts). The transmitted bandwidth is determined by the chip rate and by the baseband filtering. The implementation limits the maximum chiprate Rc (clock rate) and thus the maximum spreading. The PSK modulation scheme requires a coherent demodulation. A short-code system uses a PN code length equal to a data symbol. A long-code system uses a PN code length that is much longer than a data symbol, so that a different chip pattern is associated with each symbol.

53

Frequency Hopping Spread Spectrum

A pseudo-noise sequence pnt generated at the modulator is used in conjunction with an M-ary FSK modulation to shift the carrier frequency of the FSK signal pseudorandomly, at the hopping rate R h. The transmitted signal occupies a number of frequencies in time, each for a period of time Th (=1/Rh), referred to as dwell time. FHSS divides the available bandwidth into N channels and hops between these channels according to the PN sequence. At each frequency hop time the PN generator feeds the frequency synthesizer a frequency word FW (a sequence of n chips) which dictates one of 2n frequency positions fhi. Transmitter and receiver follow the same frequency hop pattern. The transmitted bandwidth is determined by the lowest and highest hop positions and by the bandwidth per hop position (Dfch). For a given hop, the instantaneous occupied bandwidth is identical to bandwidth of the conventional M-FSK, which is typically much smaller than Wss. So the FSSS signal is a narrowband signal, all transmission power is concentrated on one channel. Averaged over many hops, the FH/M-FSK spectrum occupies the entire spread spectrum bandwidth. Because the bandwidth of an FHSS system only depends on the tuning range, it can be hopped over a much wider bandwith than an DSSS system. Since the hops generally result in phasediscontinuity (depending on the particular implementation) a noncoherent demodulation is done at the receiver. With slow hopping there are multiple data symbols per hop and with fast hopping there are multiple hops per data symbol.

54

3. Basic principle of Direct Sequence Spread Spectrum (DSSS )

For BPSK modulation the building blocks of a DSSS system are:

Input:
Binary data dt with symbol rate Rs = 1/Ts (= bitrate Rb for BPSK) Pseudo-noise code pnt with chip rate Rc = 1/Tc (an integer of Rs)

Spreading:
In the transmitter, the binary data dt (for BPSK, I and Q for QPSK) is directly multiplied with the PN sequence pnt, which is independant of the binary data, to produce the transmitted baseband signal tx b: txb = dt . pnt The effect of multiplication of dt with a PN sequence is to spread the baseband bandwidth Rs of dt to a baseband bandwidth of Rc.

Despreading:
The spread spectrum signal cannot be detected by a conventional narrowband receiver. In the receiver, the received baseband signal rxb is multiplied with the PN sequence pnr. If pnr = pnt and synchronized to the PN sequence in the received data, than the recovered binary data is produced on dr. The effect of multiplication of the spread spectrum signal rxb with the PN sequence pnt used in the transmitter is to despread the bandwidth of rxb to Rs. If pnr pnt , than there is no despreading action. The signal dr has a spread spectrum. A receiver not knowing the PN sequence of the transmitter cannot reproduce the transmitted data. To simplify the description of modulation and demodulation, the spread spectrum system is considered for baseband BPSK communication (without filtering) over an ideal channel.

55

3.1 Modulation

Spread spectrum systems are spreading the information signal dt which has a BW info, over a much larger bandwidth BWSS:

The SS-signal spectrum is white noise-like. The amplitude and thus the power in the SS-signal txb is the same as in the original information signal dt. Due to the increased bandwidth of the SS- signal the power spectral density must be lower. The bandwidth expansion factor, being the ratio of the chip rate Rc and the data symbol rate Rs, is usually selected to be an integer in practical SS systems:

56

3.2 Demodulation

To demodulate, the received signal is multiplied by pnr, this is the same PN sequence as pnt (the pseudo-noise code used in the transmitter), synchronized to the PN sequence in the received signal rxb. This operation is called (spectrum) despreading, since the effect is to undo the spreading operation at the transmitter. The multiplier output in the receiver is then (since pnr = synchronized pnt) : dr = rxb . pnr = (dt . pnt). pnt The PN sequence pnt alternates between the levels -1 and +1, in the example: pnt = +1 +1 +1 -1 +1 -1 1 The alternation is destroyed when the PN sequence pnt is multiplied with itself (perfectly synchronized), because: pnt . pnt = +1 for all t Thus: autocorrelation Ra (t=0) = average (pnt . pnt) = +1 The data signal is reproduced at the multiplier output: dr = dt

57

If the PN sequence at the receiver is not synchronized properly to the received signal, the data cannot be recovered.

3.2.2 pnr pnt

If the received signal is multiplied by a PN sequence pnr, different from the one used in the modulator, the multiplier output becomes: dr = rxb . pnr = (dt . pnt ). pnr In the receiver, detection of the desired signal is achieved by correlation against a localreference PN sequence. For secure communications in a multi-user environment, the transmitteddata dt may not be recovered by a user that doesnt know the PN sequence pnt used at thetransmitter. Therefore: crosscorrelation Rc ( t) = average (pnt . pnr) << 1 for all t is required. This orthogonal property of the allocated spreading codes, means that the output of the correlator used in the receiver is approximately zero for all except the desired transmission.

4. Performance in the presence of interference

To simplify the influence of interference, the spread spectrum system is considered for baseband BPSK communication (without filtering).

The received signal rxb consists of the transmitted signal tx b plus an additive interference i (noise, other users, jammer, ... ): rxb = txb + i = dt . pnt + i To recover the original data dt , the received signal rxb is multiplied with a locally generated PN sequence pnr that is an exact replica of that used in the transmitter (that is pnr = pnt and synchronized). The multiplier output is therefore given by: dr = rxb . pnt = dt . pnt . pnt + i . pnt The data signal dt is multiplied twice by the PN sequence pnt , whereas the unwanted interference i is multiplied only once. Due to the property of the PN sequnence: pnt . pnt = +1 for all t The multiplier output becomes: dr = dt + i . pnt The data signal dt is reproduced at the multiplier output in the receiver, except for the interference represented by the additive term i . pnt . Multiplication of the interference i by the locally generated PN sequence, means that the

58

spreading code will affect the interference just as it did with the information bearing signal at the transmitter. Noise and interference, being uncorrelated with the PN sequence, become noise-like, increase in bandwidth and decrease in power density after the multiplier. After despreading, the data component dt is narrow band (Rs) whereas the interference component is wideband (Rc). By applying the dr signal to a baseband (low-pass) filter with a bandwidth just large enough to accommodate the recovery of the data signal, most of the interference component i is filtered out. The effect of the interference is reduced by the processing gain (Gp).

4.1 Narrowband interference

The narrowband noise is spread by the multiplication with the PN sequence pnr of the receiver.The power density of the noise is reduced with respect to the despread data signal. Only 1/G p of the original noise power is left in the information baseband (Rs). Spreading and despreading enables a bandwith trade for processing gain against narrow band interfering signals. Narrowband interference would disable conventional narrowband receivers. The essence behind the interference rejection capability of a spread spectrum system: the usefull signal (data) gets multiplied twice by the PN sequence, but the interference signal gets multiplied only once.

4.2 Wideband interference

59

Multiplication of the received signal with the PN sequence of the receiver gives a selective despread of the data signal (smaller bandwidth, higher power density). The interference signal is uncorrelated with the PN sequence and is spread. Origin of wideband noise: Multiple Spread Spectrum users: multiple access mechanism. Gaussian Noise: There is no increase in SNR with spread spectrum. The larger channel bandwidth (Rc instead of Rs) increases the received noise power with Gp :

The spread spectrum signal has a lower power density than the directly transmitted signal.

5. Pseudo-Noise Sequences PN
5.1 Random White Gaussian Noise
Zero-mean White Gaussian Noise (WGN) has the same power spectral density G WGN (f) for all frequencies. The adjective white is used in the sense that white light contains equal amounts of all frequencies within the visible band of electromagnetic radiation. The autocorrelation function of WGN is given by the inverse Fourier transform of the noise power spectral density GWGN (f):

60

The autocorrelation function RaWGN( ) is zero for 0. This means that any two different samples of WGN, no matter how close together in time they are taken, are uncorrelated. The noise signal WGN(t) is totally decorrelated from its time-shifted version for any 0. The amplitude of integrated bandlimited) WGN has a Gaussian probability density distribution p(WGNi):

5.2-Pseudo -Random Noise

A Pseudo-Noise (PN) code sequence acts as a noiselike (but deterministic) carrier used for bandwidth spreading of the signal energy. The selection of a good code is important, because type and length of the code sets bounds on the system capability. The PN code sequence is a Pseudo-Noise or Pseudo-Random sequence of 1s and 0s, but not a real random sequence (because periodic). Random signals cannot be predicted. The autocorrelation of a PN code has properties simular to those of white noise.

Pseudo-Random:
Not random, but it looks randomly for the user who doesnt know the code. Deterministic, periodical signal that is known to both the transmitter and the receiver. The longer the period of the PN spreading code, the closer will the transmitted signal be a truly random binary wave, and the harder it is to detect. Statistical properties of sampled white-noise.

Length:
Short code: The same PN sequence for each data symbol (Nc.Tc = Ts). Long code: The PN sequence period is much longer than the data symbol, so that a different chip pattern is associated with each symbol (Nc.Tc >> Ts).

61

5.3 Properties of PN Sequences Balance Property

In each period of the sequence the number of binary ones differs from the number of binaryzeros by at most one digit (for Nc odd).

pn = +1 +1 +1 -1 +1 -1 1 = + 1 When modulating a carrier with a PN coding sequence, one-zero balance (DC component) can limit the degree of carrier suppression obtainable, because carrier suppression is dependant on the symmetry of the modulating signal.
When modulating a carrier with a PN coding sequence, one-zero balance (DC component) can limit the degree of carrier suppression obtainable, because carrier suppression is dependant on the symmetry of the modulating signal.

Run-length Distribution
A run is a sequence of a single type of binary digits. Among the runs of ones and zeros in each period it is desirable that about one-half the runs of each type are of length 1, about one-fourth are of length 2, one-eigth are of length 3, and so on.

Autocorrelation
The origin of the name pseudo-noise is that the digital signal has an autocorrelation function which is very similar to that of a white noise signal: impuls like. The autocorrelation function for the periodic sequence pn is defined as the number of agreements less the number of disagreements in a term by term comparison of one full period of the sequence with a cyclic shift (position t) of the sequence itself:

It is best if Ra(t) is not larger than one count if not synchronized (=0).

62

For PN sequences the autocorrelation has a large peaked maximum (only) for perfect synchronization of two identical sequences (like white noise). The synchronization of the receiver is based on this property.

Frequency Spectrum
Due to the periodic nature of the PN seqeunce the frequency spectrum has spectral lines which become closer to each other with increasing sequence length Nc. Each line is further smeared by data scrambling, which spreads each spectral line and further fills in between the lines to make the spectrum more nearly continuous. The DC component is determined by the zero-one balance of the PN sequence.

Cross-correlation:
Cross-correlation describes the interference between codes pni and pnj :

Cross-correlation is the measure of agreement between two different codes pni and pnj. When the cross-correlation Rc(t) is zero for all t, the codes are called orthogonal. In CDMA multiple users occupy the same RF bandwidth and transmit simultaneous. When the user codes are orthogonal, there is no interference between the users after despreading and the privacy of the communication of each user is protected. In practice, the codes are not perfectly orthogonal; hence the cross-correlation between user codes introduces performance degradation (increased noise power after despreading), which limits the maximum number of simultaneous users.

5.4 Types 5.4.1 m-sequence

A Simple Shift Register Generator (SSRG) has all the feedback signals returned to a single input of a shift register (delay line). The SSRG is linear if the feedback function can be expressed as a modulo-2 sum (xor).

The feedback function f(x1,x2, ,xn) is a modulo-2 sum of the contents xi of the shift register cells with ci being the feedback connection coefficients (ci=0=open, ci=1=connect). An SSRG with L flip- flops produces sequences that depend upon register length L, feedback tap connections and initial conditions. When the period (length) of the sequence is exactly N c = 2L -1, the PN sequence is called a maximum-length sequence or simply an m-sequence. An m-sequence generated from a linear SSRG has an even number of taps. If an L-stage SSRG has feedback taps on stages L, k, m and has sequence ... , ai, ai+1, ai+2, ... than the reverse SSRG has feedback taps on L, L-k, L-m and sequence ... , ai+2, ai+1, ai, ... .

63

In the following table the feedback connections (even number) are tabulated for m-sequences generated with a linear SSRG (without image set).

For every set [L, k, ... , p] feedback taps listed in the table, there exists an image set (reverse set) of feedback taps [L, L-k, ... , L-p] that generates an identical sequence reversed in time.

64

Properties Balance
For an m-sequence there is one more one than zero in a full period of the sequence. Since all states but the allzero state are reached in an m-sequence, there must be 2L-1 ones and 2L-1-1 zeros.

Run-length distribution
For every m-sequence period, half the runs (of all 1s or all 0s) have length 1, one-fourth have length 2, one-eight have length 3, etc. For each of the runs there are equally many runs of 1s and 0s.

Autocorrelation
The autocorrelation function of the m-sequence is -1 for all values of the chip phase shift t, except for the [-1, +1] chip phase shift area, in which correlation varies linearly from the -1 value to 2L-1 = Nc (the sequence length). The autocorrelation peak increases with increasing length Nc of the m-sequence and approximates the autocorrelation function of white noise. Other codes can do no better than equal this performance of m-sequences!

Crosscorrelation
Cross-correlation is the measure of agreement between two different codes. Unfortunatly, crosscorrelation is not so well behaved as autocorrelation. When large numbers of transmitters, using different codes, are to share a common frequency band (multi-user environment), the code sequences must be carefully chosen to avoid interference between users.

65

Security
The m-sequence codes are linear, and thus not usable to secure a transmission system. The linear codes are easily decipherable once a short sequential set of chips (2L+1) from the sequence is known. (The overall system could still be secure if the information itself where encoded by a cryptographically secure technique).

5.4.2 Barker Code

The number of stages L in the SSRG als o determines the length (period) Nc =2L 1 of the msequence codes. The Barker code gives codes with different lengths and simular autocorrelation properties as the m-sequences.

The autocorrelation function of the balanced 11 chip Barker code is shown in the next figure.

66

5.4.3 Gold Codes

The autocorrelation properties of the m-sequences cannot be bettered. But a multi-user environment (Code Devision Multiple Access) needs a set of codes with the same length and with good cross-correlation properties. Gold code sequences are usefull because a large number of codes (with the same length and with controlled crosscorrelation) can be generated, although they require only one pair of feedback tap sets. Gold codes are product codes achieved b y the exclusive or-ing (modulo-2 adding) of two maximum-length sequences with the same length (factor codes). The code sequences are added chip by chip by synchronous clocking. Because the m-sequences are of the same length, the two code generators maintain the same phase relationship, and the codes generated are of the same length as the two base codes which are added together, but are nonmaximal (so the autocorrelation function will be worse than that of m -sequences). Every change in phase position between the two generated m-sequences causes a new sequence to be generated.

Any 2-register Gold code generator of length L can generate 2L - 1 sequences (length 2L - 1) plus the two base msequences, giving a total of 2L + 1 sequences. In addition to their advantage in generating large numbers of codes, the Gold codes may be chosen so that over a set of codes available from a given generator the autocorrelation and the crosscorrelation between the codes is uniform and bounded. When specially selected msequences, called preferred m -sequences, are used the generated Gold codes have a three valued crosscorrelation.

This important subset of Gold codes are the Preferred Pair Gold codes.

67

Predictable cross-correlation properties are necessary in an environment where one code must be picked from several codes which exist in the spectrum. Only part of the generated Gold codes are balanced.

68

5.4.4 Hadamard-Walsh Codes

The Hadamard -Walsh codes are generated in a set of N = 2n codes with length N = 2n. The generating algorithm is simple:

The rows (or columns) of the matrix Hn are the Hadamard-Walsh codes .

In each case the first row (row 0) of the matrix consist entirely of 1s and each of the other rows contains N/2 0s and N/2 1s. Row N/2 starts with N/2 1s and ends with N/2 0s. The distance (number of different elements) between any pair of rows is exactly N/2. For H8 the distance between any two rows is 4, so the Hamming distance of the Hadamard code is 4. The Hadamard-Walsh code can be used as a block code in a channel encoder: each sequence of n bits identifies one row of the matrix (there are N =2n possible rows). All rows are mutually orthogonal:

for all rows i and j. The cross-correlation between any two Hadamard-Walsh codes of the same matrix is zero, when perfectly synchronized. In a synchronous CDMA system this ensures that there is no interference among signals transmitted by the same station. Only when synchronized, these codes have good orthogonal properties. The codes are periodic, which results in less spreading efficiency and problems with synchronization based on autocorrelation.

69

6. Transmitter Architecture
A typical architecture of a Direct Sequence Spread Spectrum (DS-SS) transmitter:

7. Receiver Architecture
A typical architecture of a Direct Sequence Spread Spectrum (DS-SS) receiver:

The basic building blocks of a DS-SS (digital) receiver are: .coherent IQ vector-demodulator with waveform synthesizer (Direct Digital Synthesis) at the IF-carrier frequency (fIF) and chip matched filters (usually Square Root Raised Cosine) .despreading (correlation of the received symbols with the locally generated PN-sequence(s) pnI and pnQ) .decorrelated IQ to data demodulator mapping synchronization loops for the IF-carrier (fIF, phase error DjIF measured after despreading to reduce the influence of noise) and chip frequency (fchip)

70

8. PN Decorrelators
Two PN decorrelator architectures can be used for despreading spread spectrum signals: the matched filter and the active correlator. They are optimum from a SNR point of view.

8.1 PN Matched Filter

A typical matched filter implements convolution using a finite impuls response filter (FIR) whose coefficients are the time reverse of the expected PN sequence, to decode the transmitted data.

The output of the FIR filter is the convolution of the received IQ-demodulated and filtered signal iqc (Ichip or Qchip on the receiver architecture blockdiagram) with the FIR impulsreponse h=[ h0 h1 ... hNc-1]. Due to the time reversion, the output of the filter is the correlation of rxb with the local PN sequence.

In the shown example, 1 sample of the received signal per chip is taken (fsample = fchip). To increase the accuracy of synchronisation oversampling with a factor s can be used. In this case there are s samples per chip (fsample = s.fchip). The dimensions of the matched filter are also increased with a factor s (each filter coefficient hi is used s time). If the receiver is not synchronized, then the received signal will propagate through the matched filter, which outputs the complete correlation function. The large peak confirms that the correct code is indeed being received and provides accurate timing information for the synchronisation of the received signal. The output Rc of the FIR PN matched filter is immediately the decorrelated data: the polarity of the large correlation peaks indicates the data value.

8.2 PN Active Correlator (Integrate and Dump)

When timing information is already available, then the simpler active correlator receiver can be used. This receiver only operates correctly when the local PN sequence pnr is accurately matched and correctly timed, with respect to the spreading code within the received signal rxb.

71

Synchronization can be obtained by sliding the reference signal through the received signal. This can be an extremely slow process, however, for large spreading waveforms (long codes).

9. PN Synchronization
For its proper operation, a SS communication system requires that the locally generated PN sequence (pnr used in the receiver Rx to despread the received signal) is synchronized to the PN sequence of the transmitter generator (pnt used to spread the transmitted signal in the transmitter Tx) in both its rate and its position. Due to the sharp peak in the autocorrelation function, a misalignment in the PN sequence of Tc/2 gives a loss of a factor 2 in processing gain.

Sources of Synchronization Uncertainty Time uncertainty:

Uncertainty in distance between Tx -Rx (propagation delay) Relative clock shifts Different phase between Tx-Rx (carrier, PN sequence)

Frequency uncertainty:
Relative velocity vr between Tx-Rx (Doppler frequency shift) affects the carrier frequency fcarrier (with c the speed of light in the propagation medium):

For a carrier frequency of 2.4 GHz this gives a frequency shift Dfcarrier = 2.2Hz/km/hr. For a relative velocity vr = 100 km/hr this gives Dfcarrier = 220Hz. The process of synchronizing the locally generated PN sequence with the received PN sequence is usually accomplished in two steps. The first step, called acquisition, consists of bringing the two spreading signals into coarse alignment with one a nother. Once the received PN sequence has been acquired, the second step, called tracking, takes over and continuously maintains the best possible waveform fine alignment by means of a feedback loop. This is essential to achieve the highest correlation power and thus highest processing gain (SNR) at the receiver.

72

9.1 Acquisition Phase (Coarse Synchronization)

The acquisition problem is one of searching throughout a region of time and frequency (chip, carrier) in order to synchronize the received spread-spectrum signal with the locally generated PN sequence. Since the despreading process typically takes place before carrier synchronization, and therefore the carrier is unknown at this point, most acquisition schemes utilize noncoherent detection. A common feature of all acquisition methods is that the received signal and the locally generated PN sequence are first correlated with a coarse time step (mostly Tc/2) to produce a measure of simularity between the two. This measure is then compared to a threshold to decide if the two signals are in synchronism. If they are, a verification algorithm is started. To prevent false locking, it is necessary to dwell for some time to test synchronism. Than the tracking loop takes over. For proper synchronization, a peaked autocorrelation is required from the PN sequence.

Matched Filter (parallel)

A matched filter calculates the correlation function at each sample timestep T sample. This gives the shortest acquisition time but the fully parallel implementation requires a lot of hardware. The hardware increases with the PN codelength and and oversampling factor s. Therefore it is mostly used for short codes.

Active Correlator (serial)

An active correlator needs an integration over a total period Nc.Tc of the PN sequence to calculate one point of the correlation function. Less hardware is needed, but a larger acquisition time is required. This can be reduced by using parallelism as explained below.

9.1.1 Serial Synchronization (Sliding Correlator)

The sliding correlator is based on the correlation result of one active correlator. The correlator cycles through the time uncertainty, usually in discrete time intervals of T c/2 seconds or less. The correlation is performed over the period of the PN sequence T s = Nc.Tc. After each integration interval the correlator output is compared with a threshold to determine if the known PN sequence is present. If the threshold is not exceeded, the known PN sequence of the receiver (pnr) is advanced by Tc/2 seconds and the correlation process is repeated. These operations are performed until a signal is detected or until the search has been performed over the time uncertainty interval Tu. For a coarse time step of Tc/2 the worst case acquisition time is (Tu = NcTc):

This becomes unaccepatable long for long codes (large Nc).

73

9.1.2 Serial/Parallel Synchronization

More active correlators are placed in parallel (3 in this example) with PN sequences spaced one half chip (Tc/2) apart. After the integration period Nc.Tc the results of the correlator outputs are compared. The correlation function is thus calculated in 3 successive points (spaced one half chip apart). When no comparator output exceeds the threshold the sequences are advanced over 3Tc/2 seconds. When the threshold is exceeded, the correlator output with the largest output is chosen. For a search with 3 parallel correlators over the time uncertainty Tu interval in time steps of Tc/2 the worst case acquisition time is(Tu = NcTc):

The search time is reduced at the expense of a more complex and costly implementation.

74

9.2 Tracking Phase (Fine Synchronization)

The tracking maintains the PN code generator at the receiver in synchronism with the received signal. This is needed to achieve maximum processing gain. For a PN sequence phase error of Tc/2 the processing gain is reduced with a factor 2.

Delay-Locked Loop (DLL)

The locally generated code pnr(t) of the tracking loop is offset in phase from the incoming pn(t) by a time t, with t < Tc/2. In the DLL, two PN sequences pnr(t + Tc/2 + t) and pnr(t - Tc/2 + t) are delayed from each other by one time chip (Tc). The Early and Late outputs are the evaluation of the autocorrelation function at two timepoints: ce = Ra(t - Tc/2) cl = Ra(t + Tc/2).

When t is positive, the feedback signal Y(t) instructs the chip clock generator (NCO = numerical controlled oscillator) to increase its frequency, thereby forcing t to decrease. When t is negative, the feedback signal Y(t) instructs the numerical controlled oscillator (NCO) to decrease its frequency, thereby forcing t to increase.

75

10. Multiple Access

Code Division Multiple Access (CDMA) is a method of multiplexing (wireless) users by distinct (orthogonal) codes. All users can transmit at the same time, and each is allocated the entire available frequency spectrum for transmission. CDMA is also know as spread-spectrum multiple access SSMA. CDMA does not require the bandwidth allocation of FDMA, nor the time synchronization of the indivividual users needed in TDMA. A CDMA user has full time and full bandwidth available, but the quality of the communication decreases with an increasing number of users (BER-). In CDMA each user: has its own PN code uses the same RF bandwidth transmits simultaneously (asynchronous or synchronous)

Correlation of the received baseband spread spectrum signal rxb with the PN sequence of user 1 only despreads the signal of user 1. The other users produce noise Nu for user 1.

Only that portion of the noise produced by the other users falling in the information bandwidth [-Rs, Rs] of the receiver, will cause interference with the desired signal.

76

The set of PN codes must have the following properties: autocorrelation for good synchronization low crosscorrelation (orthogonal codes) for low MAI Useful codes are: Gold codes, Kasami codes (asynchronous CDMA) Hadamard-Walsh codes (synchronous CDMA)

Multiple Access Interference (MAI)

The detector receives a signal composed of the sum of all users signals, which overlap in time and frequency. Multiple access interference (MAI) refers to the interference between directsequence users and is a factor which limits the capacity and performance of DS-CDMA systems. In a conventional DS-CDMA system, a particular users signal is detected by correlating the entire received signal with that users code waveform. The conventional detector does not take into account the existence of MAI. Because of the interference among users, however, a better detection strategy is one of multi-user detection. Information about multiple users is used jointly o better detect each individual user.

Near-Far problem
Suppose: . Wireess channel Multi-users (transmitters) using the same channel One receiver Each user is a source of interference for the other users, and if one is received with more power, han that user generates more interference for the other users. It is important that the receiver ets the same power from each transmitter. The use of power control ensures that all users rrive at about the same power Prx at the receiver, and therefore no user is unfairly isadvantaged relative to the others. The signal-to-noise interference power ratio at the receiver inputfor Nu simultaneous users is:

77

11. Multipath Channels

In wireless channels there exists often multiple path propagation: there is more than one path rom the transmitter to the receiver. Such multipaths may be due to: . Atmospheric re flection or refraction Reflections from ground, buildings or other objects

Multipaths may result in fluctuations in the received signal level (fading). Each path has its own ttenuation and time delay. It is important to keep the direct path and reject the others. Assume that the receiver is synchronized to the time delay and RF phase of the direct path. The ignals at the receiver can be from: the direct path, other paths, white noise, interference. Suppose two discrete paths: a direct path and only one non-direct path (delayed by a time t compared to the direct path).

= differential time delay between the two paths 0< < T = random angle phase between the carrier of the direct and the non-direct path = attenuation of the secundary path
The signal at the receiver can be expressed as :

78

For the receiver, synchronized to the direct path signal, the output of the correlator, can be ritten as:

The PN sequence has an autocorrelation function with the property: pn(t) pn(t) = 1 pn(t) pn(t-t) 1 Multipath signals that are delayed by a chip period or longer relative to the desired signal outdoor reflections) are essentially uncorrelated and do not contribute to multipath fading. The SS system effectively rejects (mitigation) the multipath interference like in the case of CDMA.

with n0 = noise and multipath interference The PN code that arrives from the non-direct channel(s) is not synchronized to the PN code of the direct path and is rejected.

12. Jamming
The goal of a jammer is to disturb the communication of his adversary. The goals of the communicator are to develop a jam-resistant communication system under the following assumptions: Complete invulnerability is not possible The jammer has a priori knowledge of most system parameters, frequency bands, timing, traffic, ... The jammer has no a priori knowledge of the PN spreading code Protection against jamming waveforms is provided by purposely making the information-beating signal occupy a bandwidth far in excess of the minimum bandwidth necessary to transmit it. This has the effect of making the transmitted signal assume a noise-like appearance so as to blend into background. The transmitted signal is thus enabled to propagate though the channel undetected by anyone who may be listening. Spread spectrum is a method of camouflaging the information-bearing signal.

13. ISM Bands

ISM (Industrial, Scientific,and Medical) frequency bands are reserved for (unlicensed) spread pectrum applications.

Properties for higher frequencies: - higher path loss, shorter distance - higher implementation cost + less interference + more channels, higher throughput

79

Regulations for the 2.4 GHZ ISM band:

FHSS = Frequency Hopping Spread Spectrum

20 non-overlapping channels (hopping posistions) dwell time/channel 400 ms each channel occupied at least once during 4.(#channels).(dwell time/hop)

DSSS = Direct Sequence Spread Spectrum

Spread spectrum modulation that does not satisfy the constraints of the FHSS specification.

14. Evaluation of SS
Positive
1. Signal hiding (lower power density, noise-like), non-interference with conventional systems and other SS systems 2. Secure communication (privacy) 3. Code Division Multiple Access CDMA (multi-user) 4. Mitigation (rejection) of multipath, hold only the direct path 5. Protection to intentional interference (Jamming) 6. Rejection of unintentional interference (narrowband) 7. Low probability of detection and interception (LPI) 8. Availability of licence-free ISM (Industrial, Scientific and Medical) frequency-bands

Negative
9. No improve in performance in the presence of Gaussian noise 10. Increased bandwidth (frequency usage, wideband receiver) 11. Increased complexity and computational load

80

Spread Spectrum (SS) applications ir. J. Meel

jme@denayer.wenk.be Studiedag Spread Spectrum - 6 okt. 99

81

1. Spread Spectrum Applications

1.1 WLAN IEEE 802.11
IEEE 802.11 is the first internationally recognized standard for Wireless Local Area Networks (WLAN), introducing the technology of mobile computing. 1.1.1 Network Topology Ad-hoc Network An Ad-hoc network or Independent Basic Service Set (IBSS) is a simple network where communications are established between two or more wireless nodes or Stations ( STAs) in a given coverage area without the use of an Access Point (AP) or server. The STAs recognize each other and communicate directly with each other on a peer-to-peer level.

Infrastructure Network
An Infrastructure network (or client/server network) is a more flexible configuration in which each Basic Service Set (BSS) contains an Access Point (AP). The AP forms a bridge between the wireless and wired LAN. The STAs do not communicate on a peer-to-peer basis. Instead, all communications between STAs or between an STA and a wired network client go through the AP. APs are not mobile and form part of the wired network infrastructure. The Extended Service Set (ESS) consists of a series of BSSs (each containing an AP) connected together by means of a Distribution System (DS). Although the DS could be any type of network (including a wireless network), it is almost invariably an Ethernet LAN. Within an ESS, STAs can roam from one BSS to another and communicate with any mobile or fixed client in a manner which is completely transparant in the protocol stack above the MAC sublayer. The ESS enables coverage to extend well beyond the range of the WLAN radio.

82

1.1.2 Physical Layer (Radio Technology) Spreading and Modulation

EEE 802.11 defines three variations of the Physical Layer: Infrared (IR) and two RF transmissions in the unlicensed 2.4 GHz ISM -band, requiring spread spectrum modulation: DSSS (Direct Sequence Spread Spectrum) and FHSS (FrequIency Hopping Spread Spectrum). Only the RF transmission has significant presence in the market.

DSSS
The DSSS physical layer uses an 11-bit Barker sequence to spread the data before it is transmitted. This sequence gives a processing gain of 10.4 dB, meeting the minimum requirements of FCC 15.247 and ETS 300 328. The 11 Mcps baseband stream is modulated onto a carrier frequency (2.4 GHz ISM band, with 11 possible channels spaced with 5 MHz) using: DBPSK (Differential Binary Phase Shift Keying): data rate = 1 Mbps DQPSK (Differential Quaternary Phase Shift Keying): data rate = 2 Mbps

FHSS
In the FHSS physical layer the information is first modulated using: 2-GFSK (2-level Gaussian Frequency Shift Keying): data rate = 1 Mbps 4-GFSK (4-level Gaussian Frequency Shift Keying): data rate = 2 Mbps Both modulations result in a symbol rate of 1 Msps. The carrier frequency (2.4 GHz ISM band, with 79 possible channels spaced with 1 MHz) hops from channel to channel in a prearranged pseudo-random manner (hop pattern). There are 78 different hop patterns (subdivided in 3 sets of 26 patterns). The FCC and ETS regulations require a minimum hop rate of 2.5 hops/s or a channel dwell time of less than 400 ms .

83

Spectrum
The spectrum of the transmitted signals determines the network packing.

DSSS
With a symbol rate of 11 Mbps the channel bandwidth of the main lobe is 22 MHz. There are 11 channels identified for DSSS systems, but there is a lot of overlap (only 5 MHz spacing). All IEEE 802.11 DSSS compliant products utilize the same PN code. Since there is not a set of codes available the DSSS network cannot employ CDMA. When multiple APs are located in close proximity, it is recommended to use frequency seperations of at least 25 MHz. Therefore the 2.4 GHz ISM band will accommodate 3 non-overlapping channels. Only 3 networks can operate collocated.

FHSS
When the hop patterns are selected well, several APs can be located in close proximity with a fairly low probability of collision on a given channel. Up to 13 FHSS networks can be collocated before the interference is to high. This is based on the probability of collisions where two of the nets choose the same one of 79 channels at the same time. When the probability of collisions gets to high, network throughput suffers.

84

Comparison of DSSS and FHSS

1.2 GPS (Global Positioning System )

GPS is a satellite navigation system, funded by and controlled by the U.S. Department of Defense (DOD). The GPS system consists of three building blocks: the Space Segment (SS), the User Segment and the Control Segment (CS).

85

Space Segment (SS)

The Space Segment of the GPS system consists of the GPS satellites. These Space Vehicles (SVs) send radio signals to the User Segment and the Control Segment. The nominal GPS operational constellation consists of 24 satellites that orbit the earth in 12 hours. The satellite orbits have an altitude of 20.200km and an inclination of 55 degrees with respect to the equatorial plane. There are six orbital planes (with nominally four SVs in each), equally spaced (60 degrees apart). The satellite orbits repeat almost the same ground track once each day (4 minutes earlier each day).

Control Segment
Monitor stations measure signals from the SVs which are incorporated into orbital models for each satellite. The models compute precise orbital data (ephimeris) and SV clock corrections for each satellite. The Master Control station uploads ephimeris and clock data to the SVs. The SVs then send subsets of the orbital ephimeris data to GPS receivers (User Segment).

User Segment
The GPS User Segment receivers convert SV signals into position, velocity and time estimates. Four satellites are required to compute the four dimensions of X,Y,Z (position) and time. Authorized users with cryptographic equipment and keys and specially equipped receivers use the Precise Positioning System (PPS). PPS Predictable Accuracy (95%): 22 meter horizontal accuracy 27.7 meter vertical accuracy 100 nanosecond time accuracy Civil users worldwide use the Standard Positioning System (SPS) without charge or restrictions. Most receivers are capable of receiving and using the SPS signal. The SPS accuracy is intentionally degraded by the DOD by the use of Selective Availability. SPS Predictable Accuracy (95%): 100 meter horizontal accuracy 156 meter vertical accuracy 340 nanoseconds time accuracy

86

GPS Satellite Signals

The SVs transmit two microwave carrier signals. The L1 frequency (1575.42 MHz) carries then avigation message and the SPS code signals. The L2 frequency (1227.60 MHz) is used to rneasure t he ionospheric delay by PPS equipped receivers. Three binary codes shift the L1 and/or L2 carrier phase. The C/A Code (Coarse Acquisition) modulates the L1 carrier phase. The C/A code is a repeating 1.023 Mchip/s Pseudo Random Noise (PRN) Code. This noise-like code modulates the L1 carrier signal, "spreading" the spectrum over a 1 MHz bandwidth. The C/A code repeats every 1023 chips (one millisecond). This chip length Nc of 1023 chips results in a processing gain of 30 dB. Thats why GPS receivers dont need big satellite dishes to receive the GPS signal. There is a different C/A code PRN for each SV. GPS satellites are identifed by their PRN number, the unique identifier for each pseudo-random-noise code. This code-division-multiplexing technique allows the identification of the SVs even though they all transmit at the same L1-band frequency. A low cross-correlation gives a minimum of interference between the SV signals at the receiver side. The C/A code that modulates the L1 carrier is the basis for the civil SPS. The P-Code (Precise) modulates both the L1 and L2 carrier phases. The P-Code is a very long (seven days period = 6.19.1012 chips) 10.23 Mchip/s PRN code. In the Anti-Spoofing (AS) mode of operation, the P-Code is encrypted into the Y-Code. The encrypted Y-Code requires a classified AS Module for each receiver channel and is for use only by authorized users with cryptographic keys. The P (Y)-Code is the basis for the PPS. The Navigation Message (NAV data) also modulates the L1-C/A code signal. The Navigation Message is a 50 bps signal consisting of data bits that describe the GPS satellite orbits, clock corrections, and other system parameters (1500 bits = 30 sec).

The Long code (P or Y code) is identical for each satellite. The Short code or C/A code is a Gold code with the generator shown below.

87

The C/A code generator produces a different 1023 chip sequence for each phase tap setting. The C/A codes are defined for 32 satellite identification numbers (PRN ID).

88

Measuring the distance d between the SV and the RX is based on measuring the travel time t d of the radio signal (L1/L2) send by the SV and the propagation speed c of the signal: d = c . Td The travel time td is measured by synchronizing the C/A code (or P(Y) code) of the receiver to the C/A code in the signal received from the SV. The start time of this synchronized C/A code in the receiver gives the Time Of Arrival (TOA) of the C/A code of the SV at the receiver. The start time t1 of the C/A code in the SV is known (time information is included in the Navigation Message). The travel time td can be calculated from t1 and TOA. Because c = 300000000 m/s, the time must be measured very accurate:

On the Space Vehicle (SV), timing is almost perfect because they have precise atomic clocks on board. A low-cost GPS receiver cannot have an atomic-accuracy clock. The receiver clock time t RX shows an offset toff from the SVs GPS time t GPS: Trx = Tgps Toff

Due to this inaccuracy the TOA is called the pseudo-range.

If the receiver clock is perfect, than all the satellite (SV) ranges would intersect at a single point (which is the position of the receiver). Three perfect measurements can locate a point in 3- dimensional space. With imperfect receiver clocks, a fourth measurement (done as a cross-check), will not intersect with the first three. Since any offset from GPS time will effect the four measurements in an equal way, the receiver must look to a single correction factor (timeoffset to) that it can substract from all its timing measurements that would cause them all to intersect at a single point. Making four satellite measurements gives accurate position and time information.

89

1.3 IS-95
IS-95 CDMA is a digital cellular radio system for mobile voice communication as well as many new services like mobile fax and data transmission. n the US, the initial standards were the Telecommunications Industry Association/ Electronic ndustry Association (TIA/EIA) Interim Standard 95 (IS-95) and related versions for base station and obile performance (IS-97 and IS-98, respectively). The IS-95 system operates in the same frequency band as the analog cellular system AMPS (Advanced Mobile Phone System).

1.3.1 Network Architecture Mobile Station (MS)

The Mobile Station (MS) is the subscribers interface with the CDMA network. Both hand-held MS units having a low-power radio transmitter and vehicle-mounted MS units are permitted. The manufacturer assignes a unique 32-bit Electronic Serial Number (ESN) to each MS. It is a permanent and private identification code of the mobile terminal.

Base Station Subsystem (BSS)

Each Base Station has a unique pilot PN-offset, a delay applied to a random number sequence (PN Short Code) at the base station. This sequence is applied to forward direction transmissions that enables the terminals in a cell to decode the desired signal and reject the signals from other base stations. Pilot PN offsets ensure that the received signal from one cell does not correlate with the signal from a nearby cell. It is possible for adjacent cells to use the same CDMA radio channel frequency (f 1). Reusing the same frequency in every cell eliminates the need for frequency planning in a CDMA system. Pilot PN-offset planning must be done in stead. In an area where the ranges of two cells overlap, there is an increased interference, but this only reduces the number of users that can share the radio channel.

90

Base Transceiver Station (BTS)

The BTS comprises several base radio transceivers. Each transceiver consists of a transmitter and a receiver which has a duplicated front end to match up with the two receiving antennas used in the base antenna assembly.

Base Station Controller (BSC)

The BSC comprises control logic, data communication facilities and multiplexing and demultiplexing equipment. The BSC can control the radio power levels of the various transceivers in the BTS, and also can autonomously control the mobile stations radio transmitter power levels. A single BSC can control several BTS radio equipment transmitters.

GPS Receiver
CDMA soft handover (an MS establishes contact with a new base station before giving up its radio link to the original base station) requires base stations to operate in synchronism with one another. Therefore each base station contains a GPS receiver.

Mobile Switching Centre (MSC)

The MSC is a switching network that interconnects calls between Mobile Stations and between Mobile Stations and the Public Switched Telephone Network (PSTN). The MSC is also needed for automatic roaming capabilities.

1.3.2 Forward Link Radio Transmission

The forward link is by convention the transmission from Base Station to Mobile Station (MS).

Traffic Channel
The traffic channel can accept data rates of 9600 bps, 4800 bps, 2400 bps and 1200 bps comming from a variablebit-rate speech coder (QCELP = Qualcomm Code Excited Linear Prediction). Check digits and tail bits (convolutional encoder tail sequence to drive the convolutional encoder into a known state at the end of each frame) are included. The signals are processed in frames of duration of 20 ms. A convolutional code, with constraint length K=9 and rate protects each signal. When the rate is less than 19200 bps, the transmitter repeats code bits (factor 1, 2, 4 or 8) to bring the rate up to 19200 bps, corresponding to 384 bits in a frame of 20 ms. An interleaver permutes the code bits in each frame. This will spread the influence of burst errors, typical for wireless communications. The baseband sequence is scrambled by the PN sequence derived from a Long Code Generator (a PN sequence with length 242 1 at a rate of 1.2288 Mbps) and Long Code Mask (a time -offset determined by the ESN of the MS for traffic channels).

91

The Long Code period is:

To match the rate of the Long Code sequence to the 19200 bps baseband rate, a decimator extracts 1 bit out of 64 bits of the Long Code sequence.

The baseband symbol stream is spread by multiplication with a Walsh sequence of length 64, thus creating a baseband chip rate of 1.2288 Mcps. There are 64 orthogonal Walsh sequences of length 64, certain of which are assigned to different users of the channel. All users transmissions occur synchronously from the base station, so these transmissions are also synchronized at any individual subscribers receiver (synchronous CDMA). The use of a set of orthogonal sequences thus allows perfect rejection of other-user interference associated with any given transmission path within the cell. The same baseband sequence is duplicated on the I and Q channels of an IQ-modulator. Then they are spread with different pilot sequences on the I and Q channels. This pilot sequence or Short Code sequence has a length of 215 chips. A PN-offset in the pilot sequences is assigned to each base station and is synchronized to Universal Coordinated Time (UCT). To demodulate a received signal, an MS synchronizes its receiver with the assigned base station and generates I-channel and Q-channel pilot sequences with the value of PN-offset assigned to the local base station. Signals received from other base stations, with different values of PN-offset, appear as low-level noise in the receiver of the MS, due to the correlation properties of the sequences. There are 512 possible PNoffsets, with offset i corresponding to a time delay of 64i chips (a delay of 64 chips @ 52 ms @ 15km). Since the period of the sequence is 215 chips, there are 2^15 / 2^6 = 2^9 = 512 possible offsets.

92

The bandwidth of a CDMA signal is 1.23 MHz. The bandwidth of an AMPS channel (using the same frequency band) is 30 kHz. Therefore the bandwidth of a CDMA signal corresponds to an aggregate bandwidth of 41 AMPS channels .

Pilot Channel
The pilot channel uses WALSH 0, a sequence of all 0s (or 1s). The channel contains no information, only the PN pilot sequence. It provides the MS with a beacon, timing and phase reference (for coherent detection). The I and Q channels of the traffic channels (containing the same information) can be despread independently to determine the amplitude of the channels. The pilot sequence can be employed for channels sounding purposes to determine the amplitudes and phases of various multipath components received at the MS (RAKE receiver).

Sync Channel
The sync channel uses WALSH 32, a sequence of 32 0s, followed by 32 1s. It provides the MS with critical time synchronization data: system time (obtained from GPS), the PN-offset of the pilot sequence and the rate of the base station paging channels (4.8 kbps or 9.6 kbps).

Paging Channel
A CDMA signal carries up to 7 paging channels and using WALSH 2 to WALSH 7. The paging channels transmit information to terminals that do not have calls in progress.

93

1.4 W-CDMA
The target of the third-generation (3G) mobile communication systems (cellular) is the introduction of multimedia capabilities. ETSI (European Telecommunications Standards Institute) has been responsible for UMTS (Universal Mobile Telecommunications System) standardization since the early 1990s. In January 1998 (historical milestone) the basic technology for the UMTS Terrestrial Radio Access (UTRA) system was selected: For the paired bands 1920 1980 MHz and 2110-2170 MHz wideband CDMA (W-CDMA) shall be used in frequency-division duplex (FDD) operation. The bearer capability targets have been defined as: 384 kbps for full area coverage ( Internet access) 2 Mbps for local coverage ( video/picture transfer) A variety of data services from low to very high bitrates must be supported.

Traffic data (voice) is multiplexed with control information (pilot bits, transmit power control bits, rate information, ... ). The serial-to-parallel converter maps the 60 kbps to the I and Q branch of the QPSK modulator. This produces a 30 ksps symbol rate. The I and Q branches are then spread to the 3.840 Mcps chip rate with the same Orthogonal Variable Spreading Factor (OVSF) code. Since the spreaded bandwidth is same for all users, multiple-rate transmission needsmultiple Spreading Factors (SF). The OVSF has an SF of 128 in this case (length of thespreading code). This results in the relation: 2 (QPSK) x datarate x SF = chip rate The OVSF code is the channelization code. Next a scrambling code is applied, which is unique to the Base Station within the geographic area. Baseband filtering is done with a Square Root Raised Cosine (SRRC) with roll-off 0.22. Different physical channels in the same cell use different channelization codes. Several downlink physical channels can be transmitted in parallel on one connection using a grouped channelization code (with lower SF and thus less transmission quality) in order to achieve higher channel bit rates.

94

Orthogonal Variable Spreading Factor (OVSF) Codes

The OVSF codes preserve mutual transmit orthogonality between different downlink physical channels, even if they use different Spreading Factors and thus offer different channel bit rates. The use of OVSF codes is thus a key factor in the high degree of service flexibility of the WCDMA air interface. Let CN be a matrix of size NxN and denote the set of N binary spreading codes of N chip length, CN(i) is the row vector of N elements and N = 2^n. The matrix CN is generated from CN/2:

These variable length codes can be generated from a tree structure as shown in the figure below. Starting from C1(1) = 1, a set of 2n spreading codes with the length of 2n chips are generated at the nth layer. The generated codes from the same layer constitute a set of Walsh functions and they are orthogonal, although the rows of CN are not in the same order of HN. Any two codes of different layers are also orthogonal except for the case that one of the two codes is a mother code of the other. For example C2(2) is a mother code for C4(3), C4(4), C8(5), C8(6), C8(7), C8(8),... , so these codes are not orthogonal against C2(2). A mother code is mapped on all the codes in the sub-tree produced by that code (they all start with that code). In other words, a code can be used in a channel if and only if no other code on the path from the specific code to the root of the tree or the sub-tree produced by the specific code is used in the same channel. For example if C8(1) is assigned to a user, all the codes {C16(1), C16(2), C32(1), , C32(4), ... } generated from this code cannot be assigned to other users requesting lower rates; in addition, mother codes {C2(1), C4(1)} of C8(1) cannot be assigned to users requesting higher rates. The OVSF code C8(1) has a Spreading Factor (SF) of 8. With a given (fixed) symbol rate of 3.840 Msps, this results in a data rate:

2bits/symbol(QPSK) x 3.840 Msps / (SF = 8 ) = 960 kbps

C8(1) is utilizing 12.5% of the available code space (channel capacity). The OVSF code C4(2) gives a datarate of 2.048 Mbps and uses 25% of the channel capacity. These restrictions are imposed in order to maintain orthogonality.

95

96