Sie sind auf Seite 1von 6

BASIC ORBITAL MECHANICS Introduction

Our concern with orbital mechanics is based on a very particular perspective. We need to understand how the motion of an earth-orbiting spacecraft affects the design parameters of a remotesensing system. The objective of this lecture is to relate orbital parameters to sensor performance metrics, such as SNR. We will adopt a frame of reference fixed, so to speak, at the sensor. Some Nomenclature An orbit is described by five independent parameters. A sixth parameter is necessary to locate the satellite along its orbit at any particular time. These orbital parameters are: eccentricity, e, inclination angle, i, semi-major axis, a [sometimes replaced by the semi-latus rectum, p = a 1 e 2 ], longitude of the ascending node, , argument of periapsis (perigee), , and the time of periapsis passage, T.2 See Figures A1(a) and (b) for a diagram that illustrates these parameters. Orbit Types There are several orbit types that you are likely to encounter. The most important type of orbit, given our current emphasis on an orbiting remote-sensing instrument, is the sun-synchronous orbit. This is a more sophisticated version of a polar orbit. In addition, there is also a geosynchronous or geostationary orbit (GEO). This latter type is more appropriate to communication or meteorological satellites or sensors whose FOV encompasses an entire half of a planet. Our interest in a polar orbit is motivated by the specification that all points on the planet need to be accessible by the satellite. A polar orbit has an inclination of 90. Consequently, as the satellite moves in its orbit, the earth rotates underneath it (Reference 1 has some very illuminating illustrations of ground coverage by polar satellites). A typical orbital period is on the order of 100 minutes (more on this later). If total coverage is required, the sensor FOV (swath width, for instance) must cover the distance that the earth rotates in that time along the equator, i.e.,
SW = RE 100 = 6,378 25

180

) 2,800 km ,

where we have slyly substituted the mean equatorial radius of the earth, RE = 6,378.145 km. Now consider a satellite at an altitude of 700 km. The above swath width requirement translates into a FOV of 127. No mean feat as we have already discovered in our discussions of optics options. The polar orbit suffers from a potentially serious short-coming: As the earth moves in its orbit around the Sun, the solar irradiance angle along the satellite trace continues to vary during the year [see Fig. 2(a)]. A better scenario would maintain a more uniform solar irradiance angle throughout the missions duration. This uniformity makes it possible to stitch separate swaths together to while eliminating effects due to varying atmospheric path (in the case of reflectance measurements) and BRDF. With a proper choice of the orbits inclination, the effects of earths oblateness may be put to use. The equatorial bulge pulls on the satellite and causes its orbit to precess (Figure 1). The inclination angle is adjusted until the precession rate matches the movement of the earth about the sun, i.e., about a 1 per day. This way the satellite crosses the equator at about the same local solar time on each pass. Figure A2 presents a plot that traces precession rate vs. inclination angle. Note that LANDSAT, for instance, takes advantage of this phenomenon with an orbital inclination of 98.2. In addition to taking advantage of natures offerings, the orbit of the satellite may be adjusted by

means of thrusters. Thrusters may be used to counter the effects of atmospheric drag on the spacecraft, for instance (see Fig. A3).

Figure 1. The oblate earth and forces on an orbiting satellite. Figure adapted from Ref. 2.

Figure 2 schematically shows the difference between a polar and a sun-synchronous orbit. A sun-synchronous orbit is usually described by its equator-crossing time. For example, in the case of LANDSAT, this time is 9.45 am. The orbital precession, as already indicated, ensures that this crossing time remains approximately the same throughout the year. Another type of orbit is the geosynchronous (or geostationary) orbit. In this case, the satellite is positioned over some point on the planet at an altitude of 35,800 km. If the point is located at the equator, the trace of the orbit on the planet is really a point. If the point on the planet is located at some non-zero latitude, the orbital trace on the planet becomes a figure-8 (see Fig. A4).2 A good overview of several orbit types can be also found in Ref. 1.

(a) Figure 2. Polar (a) and sun-synchronous (b) orbit orientations.

(b)

Key Parameters We will begin at the remote-sensing system and then work back to how what parameters of the satellites motion relate to SNR.

Point Source First, consider the abbreviated SNR-equation derivation for a point-source target (more detailed derivation to follow in future lectures),
SNR = , NEP

(1)

where is the power on the detector and NEP is the noise-equivalent power, already familiar to you from earlier lectures. A point-source target means that its image in the focal plane is smaller than the detector; it is a sub-resolution target. If we now begin at the power-transfer equation and substitute all the relevant areas, efficiencies, transmittances, ranges, and detector D , we get
*

SNR =

As Ao R
2

Ad f

a o

Ls D* .

(2)

The terms under the integral all exhibit spectral dependence. The integration is performed over the spectral range associated with the measurement. This range may be narrow, as in the case of a spectrometer, or wide, as in the case of a broad-band (visible, 3 - 5 m, etc.) imager. Extended Source Second, consider the SNR equation for an extended-source target (derivation, again, to be found in previous lectures),

SNR =

Ad Ao f Ad f
2

a o

Ls D* .

(3)

An extended-source target is one whose image on the focal plane exceeds the size of a detector element. The key consequence of this definition is that the range-to-target disappears from the SNR equation as presented in Eq. (2). Also, the source area has now been replaced by the detector area, Ad . Several of the parameters in Eqs. (2) and (3) are determined by the satellite orbit. Minimum and Maximum Range to Target Orbits are described by conics. A conic obeys the general form,
r= p , 1 + e cos

(4)

where p is the semi-latus rectum and e is termed the eccentricity of the orbit. The parameter e = 0 for a circular orbit, 0 < e < 1 for an elliptical orbit, e = 1 for a parabolic orbit, and e > 1 for a hyperbolic orbit. We can calculate the minimum and maximum values of r and thus obtain the minimum and maximum ranges by replacing = 0 or replacing = 180, respectively. From Eq. (4) and assuming that the eccentricity is very small, we obtain the following approximate relation,
rp ra =

(1 e) 1 e 2 1 2 e ( ) (1 + e )

rp = (1 2 e) ra .

(5)

Recall that r = ( RE + R ) . This approximate relation is valid to within a relative error of 1% to an eccentricity of 0.068. For reference, consider the first U.S. surveillance satellite; the eccentricity of its orbit was 0.0196. Note that satellite altitudes may be specified in nautical miles: 1 n mi = 6080 ft = 1.852 km. The typical assumption is that the orbit is circular (e = 0). In reality, orbits are elliptical and the minimum and maximum R can be calculated given the eccentricity of the orbit, e [see Eq. (5)]. The minimum and maximum ranges are often denoted by perigee and apogee, respectively (if the earth is being orbited; otherwise we may speak of perihelion and apohelion, perijove and apojove, etc.). More generally, the terms periapsis and apoapsis may be used. Ground Velocity and Dwell Time The satellite must maintain a certain speed in order to remain in orbit vs. plummeting to the ground. All the relations that we will develop in this section are based on the following equality (deja vu all over again)
G mE msat = msat r 2 , r2

(6)

where ms denote masses of the Earth and the satellite, G is the constant of gravitation, r is the distance separating the centers of the orbiting bodies, and is the angular orbital velocity. The circular velocity, vc , is obtained from Eq. (6),
2 G mE 2 G mE = vc = . r RE + R 1 1

(7)

Upon plugging in the constants (G = 6. 672 10 11 m3 kg-1 sec-2 and mE = 5. 9742 1024 kg), we get

3.986 105 2 vc = [km/sec]. 6,378 + R [ km]


Return for a moment to Eq. (7) and note that in our sorts of applications, 200 km. Consequently, we can make the following approximation,
2 G mE 2 G mE = vc = R E (1 + R R E ) RE + R = 7905 . 4 6 .197 10 4 R [ m]
1 1

(8)

R < RE . For example, R = <

G mE 1 2 RE

G mE R= , RE 3

(9)

where this time the units of the orbital altitude, i.e., the range R, are meters. The velocity is in meters/sec. Equation (7) is plotted versus orbital altitude in Fig. 3. Note the linear nature of the relationship between vc and R.

Figure 3. Plot of orbital speed (km/sec) vs. orbital altitude R (km). relationship which leads to Eq. (9).

Note the linearity of this

A similar linearized relationship may be obtained for orbital period as a function of orbital altitude,
PR P0 1 +

3 R 2 RE

) = 1. 40815 + 3.311 10

R [ km] ,

(10)

where P0 is the orbit of a somewhat theoretical satellite with zero altitude. The period is calculated in hours in Eq. (10). This approximation is good within 1% relative error out to a range of 1200 km.3 Orbital circular velocity translates to a ground-track speed. applying the ratio
vg = RE v . ( RE + R ) c

This is simply a matter of

(11)

Ground-track speed, combined with a desired ground-sampling distance (GSD), translates into a dwell time. The dwell time is obtained simply by
t dwell = GSD . vg

(12)

The dwell time represents an upper limit on integration time, t int . In practice, the integration time will need to be shorter than the dwell time to allow for readout of the sensor. The sensor bandwidth is given by f = ( 2 t int )
1

. For example, for a GSD = 5 m and a ground-track velocity of 7.5 km/sec, the

maximum integration time is 670 sec. Compare this to the time between video fields, 16 msec. For low-eccentricity orbits, the following approximate rule, similar to Eq. (2), relates the ground speed at perigee ( Rmin ) and apogee ( Rmax ) ,
v P = (1 + 2 e ) v A .

(13)

This rule is valid (relative error of 1%) for eccentricities less than 0.073. What if we measure reflectance? If a sensor measures reflected solar radiance (VNIR, SWIR, and to some extent MWIR), the solar-illumination geometry is more relevant than if the sensor measured only emitted thermal radiation. Varying solar-irradiance angle will mean: (a) varying reflected radiance and (b) possible reflectance variation due to BRDF. The reflected radiance is affected through a cos projection. The irradiance on the ground decreases by the cosine of the solar-irradiance angle. Furthermore, if the reflecting material is not Lambertian in reflectance, that too may vary. Consequently, studies that measure environmental changes, for instance, are more severely affected by such inconsistency. Summer Solstice Consider the following scenario; you wish to image a reflectance target in Nevada or New Mexico from the satellite. The sensor will be pointed at nadir, looking straight down at the target. In designing this experiment, you would also like to have the solar radiance pass through the approximately the same atmosphere through which the sensor views the target. Furthermore, the solar irradiance ought to be normally incident on the target. There is a time window during which you can execute just such an experiment. This time window occurs around June 21, when the sun appears overhead at a latitude of 23.5. This event is celebrated in certain circles as the summer solstice. If you have a more extensive budget, you can execute the experiment more accurately by performing it in Mazatlan. For an example application, see Ref. 5. References:

1. http://www.jpl.nasa.gov/basics/bsf-toc.htm 2. R.R. Bate, D.D. Mueller, J.E. White, Fundamentals of Astrodynamics (Dover Publications, New
York, 1971).

3. V.G. Szebehely, Adventures in Celestial Mechanics (University of Texas Press, Austin, 1989). 4. When the jargon begins to fly, you may wish to reach for the Longman Illustrated Dictionary of
Astronomy & Astronautics, (York Press, Beirut, 1987).

5. A.F.H. Goetz, B. Kindel, J.W. Boardman, Subpixel target detection in HYDICE data from Cuprite,
Nevada, Imaging Spectrometry II, SPIE 2819, pp. 7-14 (1996).

Das könnte Ihnen auch gefallen