Hybrid automatic repeat request

• Hybrid automatic repeat request (Hybrid ARQ or HARQ) is a combination of forward

error-correcting coding and error detection using the ARQ error-control method. In
standard ARQ, redundant bits are added to data to be transmitted using an error-detecting
code such as cyclic redundancy check (CRC). In Hybrid ARQ, forward error correction
(FEC) bits are added to the existing Error Detection (ED) bits (such as Reed-Solomon
code, Convolutional code or Turbo code) to correct a subset of all errors while relying on
ARQ to detect uncorrectable errors. As a result Hybrid ARQ performs better than
ordinary ARQ in poor signal conditions, but in its simplest form this comes at the expense
of significantly lower throughput in good signal conditions. There is typically a signal
quality cross-over point below which simple Hybrid ARQ is better, and above which
basic ARQ is better.

Simple Hybrid ARQ

The simplest version of HARQ, Type I HARQ, adds both ED and FEC information to each
message prior to transmission. When the coded data block is received, the receiver first decodes
the error-correction code. If the channel quality is good enough, all transmission errors should be
correctable, and the receiver can obtain the correct data block. If the channel quality is bad, and
not all transmission errors can be corrected, the receiver will detect this situation using the error-
detection code, then the received coded data block is rejected and a retransmission is requested
by the receiver, similar to ARQ .

In a more sophisticated form, Type II HARQ, the message originator alternates between
message bits along with error detecting parity bits and only FEC parity bits. When the first
transmission is received error free, the FEC parity bits are never sent. Also, two consecutive
transmissions can be combined for error correction if neither is error free (Comroe/Costello
1984, pp. 474-5).

To understand the difference between Type I and Type II Hybrid ARQ, consider the size of ED
and FEC added information: error detection typically only adds a couple bytes to a message,
which is only an incremental increase in length. FEC, on the other hand, can often double or
triple the message length with error correction parities. In terms of throughput, standard ARQ
typically expends a few percent of channel capacity for reliable protection against error, while
FEC ordinarily expends half or more of all channel capacity for channel improvement.

In standard ARQ a transmission must be received error free on any given transmission for the
error detection to pass. In Type II Hybrid ARQ, the first transmission contains only data and
error detection (no different than standard ARQ). If received error free, it's done. If data is
received in error, the second transmission will contain FEC parities and error detection. If
received error free, it's done. If received in error, error correction can be attempted by combining
the information received from both transmissions.
Only Type I Hybrid ARQ suffers capacity loss in strong signal conditions. Type II Hybrid ARQ
does not, because FEC bits are only transmitted on subsequent retransmissions as needed. In
strong signal conditions, Type II Hybrid ARQ performs with as good capacity as standard ARQ.
In poor signal conditions, Type II Hybrid ARQ performs with as good sensitivity as standard
FEC.

[edit] Hybrid ARQ with soft combining

In practice, incorrectly received coded data blocks are often stored at the receiver rather than
discarded, and when the retransmitted block is received, the two blocks are combined. This is
called Hybrid ARQ with soft combining (Dahlman et al., p. 120). While it is possible that two
given transmissions cannot be independently decoded without error, it may happen that the
combination of the previously erroneously received transmissions gives us enough information
to correctly decode. There are two main soft combining methods in HARQ:

• Chase combining: every retransmission contains the same information (data and parity
bits). The receiver uses maximum-ratio combining to combine the received bits with the
same bits from previous transmissions. Because all transmissions are identical, Chase
combining can be seen as additional repetition coding. One could think of every
retransmission as adding extra energy to the received transmission through an increased
Eb/N0.
• Incremental redundancy: every retransmission contains different information than the
previous one. Multiple sets of coded bits are generated, each representing the same set of
information bits. The retransmission typically uses a different set of coded bits than the
previous transmission, with different redundancy versions generated by puncturing the
decoder output. Thus, at every retransmission the receiver gains extra knowledge.

Several variants of the two main methods exist. For example, in partial Chase combining only a
subset of the bits in the original transmission are retransmitted. In partial incremental
redundancy, the systematic bits are always included so that each retransmission is self-
decodable.

An example of incremental redundancy HARQ is HSDPA: the data block is first coded with a
punctured 1/3 Turbo code, then during each (re)transmission the coded block is usually
punctured further (i.e. only a fraction of the coded bits are chosen) and sent. The punctuation
pattern used during each (re)transmission is different, so different coded bits are sent at each
time. Although the HSDPA standard supports both chase combining and incremental
redundancy, it has been shown that incremental redundancy almost always performs better than
chase combining, at the cost of increased complexity, though[1].

HARQ can be used in stop-and-wait mode or in selective repeat mode. Stop-and-wait is simpler,
but waiting for the receiver's acknowledgment reduces efficiency. Thus multiple stop-and-wait
HARQ processes are often done in parallel in practice: when one HARQ process is waiting for
an acknowledgment, another process can use the channel to send some more data.
There are other forward error correction codes that can be used in an HARQ scheme besides
Turbo codes, e.g. extended irregular repeat-accumulate (eIRA) code and Efficiently-Encodable
Rate-Compatible (E2RC) code, both of which are Low Density Parity Check Code

Applications
HARQ is used in HSDPA and HSUPA which provides high speed data transmission on
downlink and uplink respectively for mobile phone networks such as UMTS, and in the IEEE
802.16-2005 standard for mobile broadband wireless access, also known as "mobile WiMAX".It
also has been used in 3GPP Long Term Evolution.

Type I Hybrid ARQ is also used in ITU-T G.hn, a high-speed Local area network standard that
can operate at data rates up to 1 Gbit/s over existing home wiring (power lines, phone lines and
coaxial cables). G.hn uses CRC-32C for Error Detection, LDPC for Forward Error Correction
and Selective Repeat for ARQ.

HARQ is generally implemented in hardware, rather than in software.[citation needed]

Layer 2 structure for DL

Layer 2 structure for UL