U.S. patent application number 09/799723 was filed with the patent office on 2002-01-10 for high accuracy, high integrity scene mapped navigation.
This patent application is currently assigned to Winged Systems Corporation. Invention is credited to Loss, Keith R., Nicosia, Joseph M., Taylor, Gordon A..
Application Number | 20020004692 09/799723 |
Document ID | / |
Family ID | 22952034 |
Filed Date | 2002-01-10 |
United States Patent
Application |
20020004692 |
Kind Code |
A1 |
Nicosia, Joseph M. ; et
al. |
January 10, 2002 |
High accuracy, high integrity scene mapped navigation
Abstract
An aircraft including an approach and landing system, including
a navigation unit for providing navigation information, a weather
radar unit for providing radar information, a processor which
receives navigation information from the navigation unit and
information from the weather radar unit, the processor unit
providing an output representing information concerning the
aircraft in accordance with the provided navigation information and
radar information, a memory for storing information representing a
scene, the processor unit correlating the stored scene information
with the output representing information concerning the aircraft to
provide a mapped scene, a display unit for displaying the output of
said processor and the mapped scene, and a steppable frequency
oscillator for providing a signal which is stepped in frequency to
the weather radar unit, thereby providing an increased range
resolution.
Inventors: |
Nicosia, Joseph M.;
(Carlsbad, CA) ; Loss, Keith R.; (Escondido,
CA) ; Taylor, Gordon A.; (Escondido, CA) |
Correspondence
Address: |
SUGHRUE, MION, ZINN, MACPEAK & SEAS
2100 Pennsylvania Avenue, N.W.
Washington
DC
20037
US
|
Assignee: |
Winged Systems Corporation
|
Family ID: |
22952034 |
Appl. No.: |
09/799723 |
Filed: |
March 7, 2001 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
09799723 |
Mar 7, 2001 |
|
|
|
09419767 |
Oct 18, 1999 |
|
|
|
6219594 |
|
|
|
|
09419767 |
Oct 18, 1999 |
|
|
|
08880362 |
Jun 23, 1997 |
|
|
|
6018698 |
|
|
|
|
08880362 |
Jun 23, 1997 |
|
|
|
08251451 |
May 31, 1994 |
|
|
|
5654890 |
|
|
|
|
Current U.S.
Class: |
701/16 ; 340/947;
701/17 |
Current CPC
Class: |
G01S 7/412 20130101;
G08G 5/0091 20130101; G01S 13/953 20130101; G01S 19/15 20130101;
G01S 13/86 20130101; G08G 5/0021 20130101; G08G 5/0086 20130101;
G01C 21/165 20130101; G01S 19/52 20130101; G08G 5/025 20130101;
G01S 13/286 20130101; G08G 5/0065 20130101; Y02A 90/10 20180101;
G01S 13/90 20130101; G01S 13/913 20130101; G01C 23/005 20130101;
G05D 1/0676 20130101 |
Class at
Publication: |
701/16 ; 701/17;
340/947 |
International
Class: |
G08G 005/00 |
Claims
What is claimed is:
1. An aircraft including an approach and landing system, said
system comprising: at least one navigation unit for providing
navigation information; a weather radar unit for providing radar
information; a processor unit coupled to receive navigation
information from said at least one navigation unit, and to receive
radar information from said weather radar unit, said processor unit
operable for providing an output representing information
concerning an aircraft in accordance with the provided navigation
information and radar information; a memory, coupled to said
processor unit, for storing information representing a scene, said
processor unit operable for correlating the stored scene
information with the output representing information concerning the
aircraft to provide a mapped scene; a display unit, coupled to
receive the output of said processor and coupled to said memory,
for displaying the output of said processor and the mapped scene;
and a steppable frequency oscillator for providing a signal which
is stepped in frequency to said weather radar unit, thereby
providing an increased range resolution.
2. The approach and landing system as defined in claim 1, wherein
said processor unit operates according to a Generalized Hough
Transform map-match routine to provide the mapped scene.
3. The approach and landing system as defined in claim 1, wherein
said signal provided by said oscillator to said weather radar unit
is randomly stepped in frequency.
4. The approach and landing system as defined in claim 3, where
said signal provided by said oscillator to said weather radar unit
comprises a series of pulses.
5. The approach and landing system as defined in claim 4, wherein
said a first predetermined number of said pulses are randomly
stepped in frequency so that each of said predetermined number of
pulses is at a different frequency.
6. The approach and landing system as defined in claim 5, wherein
the predetermined number of pulses is 160.
7. The approach and landing system as defined in claim 1, wherein a
frequency step size of said oscillator is 250 KHz.
8. The approach and landing system as defined in claim 1, wherein
said at least one navigation unit includes an Inertial Navigation
System (INS).
9. The approach and landing system as defined in claim 7, wherein
said at least one navigation unit further includes a Global
Positioning System (GPS).
10. A method of improving a range resolution of an aircraft
approach and landing system including a weather radar unit, the
method comprising the steps of: generating a waveform which is
stepped in frequency; and applying the generated waveform to the
weather radar unit.
11. The method as defined in claim 10, wherein said generating step
includes the step of randomizing the frequency steps so that the
waveform applied to the weather radar unit is randomly stepped in
frequency.
12. The method as defined in claim 11, wherein said generating step
includes generating a series of pulses as the waveform.
13. The method as defined in claim 12, wherein a predetermined
number of said pulses is randomly stepped in frequency so that each
pulse is at a different frequency.
Description
FIELD OF THE INVENTION
[0001] The present invention is directed to an autonomous precision
approach and landing system (APALS) for enabling low visibility
landings at airports.
BACKGROUND OF THE INVENTION
[0002] Current industry practice for low-visibility landings is
dependent on airport ground equipment and inertial navigation
equipment. These techniques are limited to landings at those
runways which are equipped with highly reliable transmitters of
radio frequency localizer and glide slope information. These
existing systems either land the aircraft using an automatic pilot
or aid the pilot in landing the aircraft by providing the pilot
with autopilot control commands displayed on a Head Up Display
(HUD).
[0003] It has been suggested that future systems make use of
information received from the Global Positioning System (GPS) in
conjunction with on-board Inertial Navigation systems (INS) to
generate the necessary precise navigation for landing. However, in
addition to the external satellites required for GPS, these systems
are currently envisioned to require ground stations at known
locations near the runway for the differential precision necessary
for landing. Other proposed systems provide the pilot with a real
time image of the runway scene as derived from millimeter wave
(MMW), X-Band, or infrared (IR) frequencies.
[0004] The following are further examples of navigation systems
known in the art.
[0005] U.S. Pat. No. 5,136,297 to Lux et al discloses an autonomous
landing system. The Lux patent includes a navigation unit employed
in the system which includes a sensor, flight position data, an
image correction unit, a segmentation unit, a feature extraction
unit and a comparison unit. The Lux patent discloses that a
comparison is conducted as to whether or not a sequence of features
in the overflight path image pattern agrees with features found in
a reference store, such as map data which is stored in the system.
Further, Lux discloses the use of a radar navigation system for use
as a sensor in the system.
[0006] U.S. Pat. No. 4,698,635 to Hilton et al discloses a radar
guidance system coupled to an inertial navigation apparatus. The
system includes a master processor, a radar altimeter, a video
processor, a memory and a clock. The memory has stored therein
cartographic map data.
[0007] U.S. Pat. No. 4,495,580 to Keearns is cited to show a
navigation system including a radar terrain sensor and a reference
map storage device for storing data representing a terrain
elevation map.
[0008] U.S. Pat. No. 4,910,674 to Lerche discloses a navigation
method which includes a correlator for comparing terrain reference
data with processed altitude data obtained with a wave sensor.
[0009] U.S. Pat. No. 4,914,734 to Love et al is cited to show a
map-matching aircraft navigation system which provides navigational
updates to an aircraft by correlating sensed map data with stored
reference map data.
[0010] U.S. Pat. No. 4,891,762 to Chotiros is cited to show a
pattern recognition system for use in an autonomous navigation
system.
[0011] The above-mentioned prior systems suffer from one or more of
the following problems:
[0012] 1) Reliance on ground-based systems for precise terminal
landing information severely reduces the number of runways
available for Cat III a and b landings (currently 38 runways in the
U.S.).
[0013] 2) Reliance on GPS' and differential ground transmitters for
GPS creates a need for currently rare ground equipment and a lack
of reliability (based on the military nature of GPS). The GPS is a
military program owned, operated, and paid for by the United States
Air Force originally intended for military navigational purposes
and is designed so that civilian use can be made of it but at a
reduced accuracy. The military uses a very special code which gives
them better accuracy, that is called the P code. The normal
civilian code is called the C code which is good to about 30 m in
accuracy; however, the military retains the right to disable the C
code to the point where the accuracy goes down to about no better
100 m. This is what the military refers to as "selective
availability" so that in time of conflict they can turn on
selective availability and deny the enemy the ability to navigate
better than 100 m. There are a number of schemes for getting around
the inaccuracies imposed by the military. However, the Air Force
has maintained a position that they are against any of these
schemes which improve the accuracy when they are trying to make it
inaccurate.
[0014] The lack of reliability is also a result of the fact that,
in order to be accurate, at least four satellites must be present
in the overhead view; and, if one of the four satellites fails,
then the accuracy will be degraded. Thus, the reliability is not
just based on the on-board equipment, i.e., the GPS receiver, but
it is also based on the reliability of the satellites
themselves.
[0015] 3) Additional sensors such as MMW and IR, currently
envisioned for systems to provide pilots with the "situational
awareness" necessary to successfully land in low visibility
conditions are expensive additions to the on-board flight equipment
and are marginal in performance. MMW real-beam radars provide
"grainy" low resolution images which are difficult to interpret and
IR systems cannot penetrate in many types of fog that cause the
"low visibility" in the first place.
SUMMARY OF THE INVENTION
[0016] It is therefore an object of the present invention to
overcome the problems associated with the prior approach and
landing systems.
[0017] It is another object of the invention to provide an approach
and landing system which provides low visibility take-off and
landing assistance for several classes of aircraft.
[0018] It is another object of the invention to provide safe
landing of general aviation and transport aircraft (covered by
parts 25, 91, 121 and 125 in the Code of Federal Regulation) in low
visibility conditions [Category II, IIIa, and IIIb defined by the
Federal Aviation Administration (FAA)] without dependence on high
reliability ground transmitting equipment.
[0019] These and other objects are accomplished by the present
invention which provides an Autonomous Precision Approach &
Landing System that makes use of radar echoes from ground terrain
and cultural (man made) targets to provide the on-board Inertial
Navigation System with accurate aircraft position and velocity
updates. According to the invention, these measurements come from a
modified standard X-band, low-resolution weather radar.
BRIEF DESCRIPTION OF THE DRAWINGS
[0020] FIG. 1 is a block diagram of the APALS system according to
the invention.
[0021] FIG. 2 is a waveform diagram according to the invention.
[0022] FIG. 3 is a circuit diagram of a modified weather radar
device according to the invention.
[0023] FIGS. 4(a)-4(e) illustrate steps of APALS Synthetic Aperture
Radar (SAR) processing according to the invention.
[0024] FIG. 5 shows a reference scene and a corresponding Radar Map
according to the present invention.
[0025] FIG. 6 illustrates a Generalized Hough Transform Map-Match
Algorithm employed in the present invention.
[0026] FIG. 7 illustrates a Navigation Solution according to an
example of the invention.
[0027] FIG. 8 illustrates a Head-up Display (HUD) according to the
present invention.
DETAILED DESCRIPTION OF A PREFERRED EMBODIMENT
[0028] Several of the important features of the APALS system
according to the invention are set forth below:
[0029] A. Modified Weather Radar: The modification to a
conventional weather radar allows the modified weather radar to
make high resolution synthetic aperture maps of overflown
terrain.
[0030] B. Area Correlation: This refers to the application of
matching synthetic aperture radar maps with previously stored
references to locate specific spots on the ground near a path to a
specific runway.
[0031] C. Range/Range Rate Measurements Integrated Into Kalman
Filter: This refers to the application of using high resolution
radar range and velocity measurements of specific, but not
augmented, spots on the ground of known location to update a
navigation system using Kalman filtering.
[0032] D. Situational Awareness Display Format: This refers to the
application of precise navigational information to provide the
pilot with a "situational awareness" display of sufficient accuracy
to allow the pilot to land the aircraft in low visibility
conditions in the same manner as if (s)he were using his/her
judgement to land the plane in good visibility conditions.
[0033] Each of the above features is discussed in detail below.
[0034] FIG. 1 is a block diagram of the APALS system according to
the present invention. The other NAV Aids 2 refers to the
navigation aids that are conventionally employed on any aircraft
and include, for example, a VOR (VHF omnidirectional radar), DME
(distance measuring equipment) receiving equipment, the artificial
horizon, the vertical gyro, the airspeed indicator, and the
altimeter (which is a barometric altimeter). The APALS processor 16
will make use of this information in order to monitor the
reasonableness of the APALS estimate concerning the state of the
vehicle, which is the output {circumflex over (X)} from the
processor 16. The above-discussed elements all interface to the
system over a standard interface bus such as the known ARINC 429
bus.
[0035] The INS (inertial navigation system) or IMU (inertial
measurement unit) 4 are inertial instruments that measure the
translational accelerations and the angular rates. There are
several different IMU's that can be employed in APALS, one of which
is, for example, a Bendix unit known as the Bendix mini-tact
IMU.
[0036] The GPS receiver 6 is a special receiver that is designated
to receive the satellite signals and deduce from those satellite
signals the position and velocity of the aircraft. There are
several models that can be used for this, but there is only one or
two at present that have passed the FAA requirements for primary
navigation equipment on-board an aircraft.
[0037] The weather radar 8 which is also equipment that will
already be on-board the aircraft is, according to the invention, as
will be discussed below modified. For example, the Honeywell Primus
870 made by Honeywell may be employed. This radar is a non-coherent
radar so it would have to be modified with a new receiver and
transmitter to make it coherent. The weather radar 8 provides a
range R, and range rate R' outputs to the Processor. The weather
radar 8 receives a radar frequency control signal from the
Processor 16 which will be discussed below in connection with the
radar modification shown in FIG. 3.
[0038] The scene data base 10 is a data base created by going to
different airports that will use the system and making flights
during which the radar signatures of the ground returns are
measured. Further, aerial photographs are taken to use together
with the radar data to make references which would then be used to
compare against the radar returns that will occur when the actual
low visibility landing is taking place.
[0039] The display generator 12 and the display 14 are typically
supplied by the manufacturer of the device known as a Head-Up
Display (HUD), which is what the APALS uses as a see through device
that allows the pilot to view the outside world, and see the APALS
display in front of him or her. The pilot will see a virtual runway
even when the actual runway is obscured by, for example, clouds or
fog. Suitable HUD's are currently built by GEC Avionics (Great
Britain), Flight Dynamics, Inc. (Portland, Oregon) and Sextant
Avionique (France). The actual APALS output is a vector labeled
{circumflex over (X)} and consists of the position, velocity and
attitude information of the aircraft as best determined by the
APALS system. The display generator typically takes that
information and generates what the outside world scene would look
like from the currently estimated state of the aircraft,
{circumflex over (X)}.
[0040] The processor 16 receives inputs from elements 2, 4, 6, 8
and 10 and outputs vector {circumflex over (X)}. There are a number
of known processors that can be used for APALS.
A. Modified Weather Radar
[0041] The radar modification consists of applying randomized
stepped frequency pulse compression to allow a range resolution of
4 meters (even though a pulse length of 2 .mu.sec would normally
limit range resolution to 300 meters). The waveform consists of a
series of pulses at the normal higher Pulse Repetition Frequency
(PRF) of the weather radar (.about.3000 Hz). The first 160 pulses
are randomly stepped in frequency so that each pulse is at a
different frequency. Any one pulse, however, stays at a constant
frequency for its entire 2 .mu.s duration. This is important
because it allows the precision measurements to be made without
modifying the band-pass characteristics of the radar receiver. The
frequencies are such that there are 160 different frequencies
spanning 40 Mhz in 250 Khz steps. Over the time of each set of 160
pulses, the 40 Mhz spectrum is completely filled. The order of the
steps is randomized to avoid ambiguities. A diagram of the waveform
is shown in FIG. 2. The step size is 250 Khz which corresponds to a
4 .mu.s or 600 meter "coarse" range bin. This wider (than 2 .mu.s)
coarse bin was chosen to eliminate any ambiguities from adjacent
pulse "spillover" energy. The waveform can be as long as necessary
to integrate returns for a precise Doppler measurement.
[0042] The waveform is extended to multiples of 160 pulses because
160 is the number of pulses required to cover the 40 MHz bandwidth
needed for 4 meters range resolution. In this case the integration
time is limited to 0.25 seconds since, at X-band, it will yield a
velocity resolution of 0.07 m/sec. which is a sufficient accuracy
to update the navigation Kalman filter.
[0043] FIG. 3 shows a typical implementation of generating the
waveform by adding a steppable frequency oscillator 17 to the
weather radar. As shown in FIG. 3, the majority of the circuits of
radar (transmitter 18, frequency multipliers, dividers and mixers
20, receiver 22, duplexer 24) remain unchanged. The integration of
the modified waveform into the weather radar from each of the
different radar manufacturers will be unique.
[0044] Processing the waveform to achieve the desired resolution (4
m in range and 0.07 m/sec. in Doppler) is accomplished in a highly
efficient manner because the image is being taken of just one short
segment of range (where the beam intersects the ground). The
"picture" or map will extend 160 meters or 40 pixels in range and
therefore is contained in one 600 meter "coarse range". This is in
effect "zoom processing" of the region which is very efficient. The
application of zoom processing to this unique waveform allows very
high resolution to be achieved with very minor physical
modifications to a normally low resolution radar.
[0045] Motion Compensation: The Synthetic Aperture Radar (SAR) map
that is required for this system to work well covers a small area
and the accuracy of the vehicle motion required is within the
bounds of the knowledge of the system. This is because the
navigation portion of the system will have very precise knowledge
of the state of the vehicle's motion relative to the earth as will
be discussed below.
[0046] The following delineates the steps required for the two
dimensional zoom processing of APALS.
[0047] As described in the waveform of FIG. 2, during the
integration of 0.25 seconds there are 4000 pulses. This large
integration time is broken down into 25 sub-intervals or "words" of
160 pulses each (FIG. 2). During each sub-interval, the full
bandwidth of 40 MHz is transmitted by having each pulse at a
different frequency taken, at random, from a set of 160 frequencies
spaced 250 KHz apart. If the lowest frequency were 9 GHz, the
sequence of frequencies would be: 9.000 GHZ, 9.00025 GHz, 9.005
GHz, 9.00075 GHz, 9001 GHz . . . 9.040 GHz. The order is scrambled
randomly for reasons which will be explained below.
[0048] FIG. 4(a ) is a wave diagram for explaining a coarse range
bin. With a 2.mu.s pulse (1), if the receive signal (2) and (3) is
sampled the same time delay after the transmit pulse, those returns
will all represent targets or ground clutter from the same range.
Since the pulse is 2 .mu.s wide, the energy at the time of the
sample will come from 150 meters in front of to 150 meters behind
the point on the ground with a time delay of the sample center. The
processing chosen covers a 600 meter region centered at the time of
the central return. While there should be no return in any area
beyond .+-.150 meters, there may be spill-over from other bright
reflectors and by processing the wider coarse bin, the possibility
of ambiguous foldover is eliminated.
[0049] To simplify the explanation, a "linear" rather than a random
frequency sequence is examined. In FIG. 4(b) it is seen that the
samples, each being from a different pulse in the chain of 4000
pulses, range in frequency from f.sub.1 to f.sub.160 and then
f.sub.1 to f.sub.160 is repeated for the next 160 Pulse Repetition
Intervals (PRI's) and so on for 25 sub-intervals until 4000 pulses
have been transmitted and 4000 receive samples have been gathered.
As shown from FIG. 4(b), processing the 4000 samples into a range
profile of fine 4 meter bins is nothing more than summing the
sample values that come from the same frequency (there are 25 of
them) and using the sum as one of the inputs to an Inverse Digital
Fourier Transform (IDFT), and representing that process for all 160
frequencies. A Fourier transform is a process of taking samples in
time of a waveform and determining how much energy there is at each
frequency and the inverse of the process is taking samples of
energy content at different frequencies and producing what the
waveform looks like as a function of time (time is equivalent to
range for a radar echo).
[0050] The example given above and in FIG. 4(b) is a simplification
that would work well if there were no motion between the radar and
ground. In order to describe what is necessary for APALS to
accommodate motion, it is necessary to introduce the concepts of
phase and phase compensation.
[0051] The phase of a radar signal depends on two items, the
frequency or wavelength of the signal and the distance from the
transmitter. This is shown in FIG. 4(c). Radar waves are variations
in local electric and magnetic fields which can be represented by
the sine wave shown in FIG. 4(c).
[0052] The distance from one peak to another is called the
wavelength and is determined by the frequency of the transmitted
signal. FIG. 4(c) shows a Receiving Object whose distance is 51/4
wavelengths away from the Transmitter. The whole number of
wavelengths is not important to phase but the remainder or
fractional part is the phase difference between what is sent and
what is received. In FIG. 4(c), the phase difference is 1/4 of one
wavelength or 90.degree. (one wavelength is characterized by one
full cycle of 360.degree.). If the receiving object simply
reflected the signal back to a Receiver co-located with the
transmitter, as is the case with radar, the distance and,
therefore, the phase shift is doubled to 180.degree..
[0053] The phase of the returns from different samples but off of
the same stationary object will change with a frequency hopped
radar such as APALS. FIG. 4(d) shows the effect of changing
wavelength on phase. In FIG. 4(d), even though the transmitter and
the receiving object are the same distance apart as they are in
FIG. 4(c), the phase has increased to 180.degree., one way. In FIG.
4(d) there are 51/2 wavelengths in the single path-length.
[0054] As the frequency of the pulses increases (FIG. 4(b)), the
wavelength gets shorter and the phase difference increases. It is
precisely this change in phase as a function of frequency that
allows the IDFT to discern the ranges of object from the frequency
content of the return samples. The samples, by their nature,
contain both a measure of the energy and a measure of the phase
difference of the return from a pulse of a particular
frequency.
[0055] Relative motion between the Transmitter and the Reflecting
Object causes a phase shift with time which causes a phase shift
from pulse to pulse as shown in FIG. 4(e).
[0056] This phase shift as a function of time is known as the
Doppler effect. The measurement of this rate of change of phase or
Doppler is what allows APALS to update range rate as well as range
for the inertial system after each map-match. It is also what
creates the need for phase compensation.
[0057] It is important to note that the phase changes due to
increasing frequency have the same characteristic as the phase
changes due to increasing distance between the transmitter and the
reflecting object. In both cases, the phase changes will increase
steadily with time. This is the ambiguity that was mentioned
earlier. As long as the frequencies are stepped in order from pulse
to pulse, the IDFT will not be able to distinguish between distance
of the Reflecting Object and the speed of the Reflecting Object.
This is because the distance information is contained in the phase
differences of the reflections off a single object at different
transmit frequencies.
[0058] To obviate this ambiguity problem, the frequencies are not
stepped in order of increasing frequency as shown in FIG. 4(b), but
rather randomly. This breaks the linearity of the phase changes
with time due to frequency shifting so that it can be separated
from the always linear changing phase that is due to constant
velocity motion. It is still necessary to present the sampled
values of the return signal to the IDFT in order of increasing
frequency so the order of frequencies transmitted must be kept
track of. This is accomplished in APALS by using a pre-stored
pseudo-random frequency order which is 4000 elements long.
[0059] Once the relationships between distance, phase, and velocity
are understood in the context of the APALS waveform as described
above, the phase compensation and processing for APALS can be
concisely explained in the following steps:
[0060] 1) The received waveform is converted to a set of digital
samples which preserves both signal strength and phase difference.
This process is well known in the art as in-phase and quadrature
sampling or I & Q sampling. The digital samples are stored
temporarily and tagged both with their order in time of reception
and with their frequency order.
[0061] 2) The coarse range of interest is identified by the system
based on the desired map area, and the samples which come from the
corresponding delay are singled out for processing.
[0062] 3) The Doppler frequencies are determined for the desired
map area, and the center frequencies for the Doppler bins to be
processed are determined.
[0063] 4) For each Doppler bin, the set of samples is arranged in
order of the transmit frequency which generated it, and presented
for phase compensation prior to being sent to the IDFT.
[0064] 5) For each Doppler bin the phase rotation for each transmit
frequency and each receive time is calculated and that phase is
subtracted from each sample according to its time order and
adjusted for its wavelength based on its transmitted frequency. The
net effect is that motion is taken out of the samples that are
moving at the precise velocity that is the designated center of the
Doppler bin or filter. Objects that are moving faster or slower
will not "add up" because the phases of their samples will not be
recognized by the IDFT.
[0065] In order to prevent smearing, due to accelerations which
change the velocity during the 0.25 second dwell, the compensating
phase rotations must be calculated based, not on a constant
velocity, but on a velocity modified by the aircraft's
accelerations. These acceleration values are readily available in
the APALS system because they are part of the accurate state vector
which is calculated by the navigation filter.
B. Area Correlation
[0066] The APALS system uses the Scene Data Base 10 for pre-stored
scenes as references with which to compare the radar maps that are
produced through the weather radar. The radar maps can be thought
of as comprising resolution "cells" whose dimensions are range
resolution in the down range direction and range rate resolution in
the cross range dimension. Down range direction is simply the
radial distance from the aircraft. In a radar system the normal way
of mapping with a radar is to cut the return up into pieces that
are returns coming from different ranges. This is because the radar
is capable of measuring range by the time delay of the return. The
down range dimension is always the distance radially away from the
radar. The present system is typically looking at 45.degree. right
or left and so the down range dimension is a line going 45.degree.
off the nose of the aircraft. The cross range dimension is the
dimension that is directly orthogonal or at 90.degree. to the down
range dimension. It is not always exactly 90.degree., in the
present system it is measured by changes in the doppler frequency
of the return. The frequency of the return is dependent on the
relative velocity in the direction of that return.
[0067] The contextual information in the radar map is compared to
that of the reference. When a match is found for each point of
ground represented by a cell in the reference, the range and range
rate of the sensed scene are known with respect to the aircraft.
Since the location of at least one point in the scene is known
precisely with respect to the desired touch down point, by simple
vector subtraction, the range rate to the touch down point is
calculated. There are two aspects to generating this important
information:
[0068] 1) Generating a reference which will allow a locally unique
match to the radar map.
[0069] 2) Using a correlation algorithm that efficiently
"fine-tunes" the match point to a 1-cell accuracy and provides a
"measure of goodness" or confidence in the match.
[0070] The references for APALS are generated from aerial
photographs that have been digitized or scanned into a computer and
from SAR maps. The SAR maps are taken in two swaths, one on either
side of the final approach trajectory, that are centered 1 mile
offset of the aircraft's trajectory (ground projection). Software
is used to match points in the aerial photo with coordinates of a
pre-stored navigation grid so that the location of any point in the
photo is known relative to the runway touch down point (no matter
how far the scene is away from the runway). The key features of
these references are that they are simple and that they rely on
prominent cultural and natural features which produce consistent
radar returns that are distinguishable as lines with a unique
shape. The two types of features to have these characteristics
consistently are the corners made by a building face and the
ground, and roads.
[0071] FIG. 5 shows a typical reference and the corresponding radar
map. In this case the dots represents a specific pattern of a
highway crossing. Such simple references are found to work well
when used with the map matching algorithm well known in the art as
the "generalized Hough transform" which is described below.
[0072] The correlation algorithm used for map matching in the APALS
system is the well known generalized Hough transform. The Hough
transform is incorporated in several image processing techniques in
use today, especially in military applications. In general, the
Hough transform is a computer method typically used to find a line
or other simple shapes/patterns in a complex picture. This scene
matching algorithm is advantageous in that:
[0073] a) It requires very few points to be compared, (i.e., much
less than the total in the scene).
[0074] b) It requires the computer to perform only the mathematical
operation of adding and avoids the other more time consuming
mathematical operations.
[0075] In FIG. 6(A) a simple reference is shown to the left and a
very sparse sensed scene (just two points) is shown to the right.
The algorithm works such that every point in the sensed scene is
operated on in the following manner:
[0076] 1) Each point in the reference is tried as the particular
sensed point.
[0077] 2) As each point in the reference is tried, the position
that the "nominated point" (black point in the reference) occupies
in the sensed scene is recorded. This is shown in the sequence of
scenes in FIG. 6b. After all the points in the reference are used,
the set of recorded "nominated" points in the sensed scene is an
"upside-down and backwards" replica of the reference scene, rotated
about the "nominated point". This reversed image is shown above the
last block of FIG. 6b.
[0078] 3) As all the points in the reference are operated on, the
point in the scene with the most accumulated nominations is
designated as the match point. This is illustrated in FIG. 6c.
C. Range/Range Measurements Integrated Into Kalman Filter
[0079] The measurements being made by the radar are the magnitude
of the range vector and the magnitude of the range-rate vector from
the aircraft to a specific point in the map match scene. If at
least three of these measurements were being made simultaneously,
one could solve for the three elements of aircraft velocity
explicitly. This solution is shown in FIG. 7. The sequence of
measurements being made in FIG. 7 are the range and range-rate to
known points on the ground. The way in which these are used is
exactly analogous to the way the global positioning system works
with satellite measurements. For example, consider three satellites
that are displaced in the sky angularly. If the range and the range
rate to those satellites are known, then the components of both the
position vector and the velocity vector of the position relative to
those three satellites can be solved. Beyond that, it is necessary
to depend on information stored on the satellites and transmitted
down so that it can be determined where they are. Then the position
can be deduced. The difference in APALS is that the APALS system
takes pictures of the actual ground and compare the taken pictures
to stored maps. Once it compares them to stored maps and finds the
match point, the match point is immediately known in terms of its
range and range rate at that point of time. As APALS progresses it
measures its range and range rate relative to known points on the
ground. Once it measures three different points, it can form a
deterministic solution of where it is, in what direction it is
heading and how fast it is going. The sequence in APALS is a little
different in that it does not obtain a good geometric case in close
time proximity. Rather, it flies along mapping from side to side
but not getting that third one. It repetitively gets the range and
range rates on each side, and over time forms a very accurate
solution iteratively to the equation that allows it to know its
position vector and velocity vector.
[0080] Since, however, the measurements being made by radar are
separated in time by as much as 4 seconds, it is necessary to solve
for the components of the vectors recursively, over time, through
the use of a Kalman statistical filter. The Kalman filter uses data
from an inertial navigation system INS, or an inertial measurement
unit (IMU) 4(FIG. 1) to determine the motion of the aircraft
between measurement times. The INS or IMU 4 is more than just the
inertial instruments, but the complete collection of inertial
instruments and computer that result in a navigation solution
including position and velocity. An INS is typically only used on
commercial aircraft today that traverse the ocean.
D. Situational Awareness Display Format
[0081] The raw output of the APALS system is a very accurate
estimate of the "state vector" of the aircraft in a coordinate
system that has its origin at the desired touch down point on the
particular runway that is targeted. This knowledge of position,
velocity and attitude are provided as a "situational awareness"
display which the pilot can effectively use to safely land the
aircraft. This is accomplished primarily by displaying a conformal,
properly positioned runway outline in proper perspective to the
pilot on a Head-up Display (HUD). In clear weather the image will
overlay that of the actual runway edges as the pilot views the
runway through the wind screen. The appropriate touch-down zone
will also be displayed (conformably), thereby providing the pilot
situational awareness such that (s)he may land his/her aircraft in
the same manner as (s)he would in visual meteorological conditions
(VMC). The use of the HUD allows the pilot the earliest possible
view of the actual visual scene on the way to touchdown. The
precise navigational knowledge of the APALS system together with
the radar altimeter allows for the generation of a "flare cue" to
tell the pilot when and how to flare for a precise, slow
descent-rate touchdown.
[0082] The key aspect in being able to land using situational
awareness is the display of the conformal runway symbol and
extended center-line in a context which also includes conformal
symbols of the horizon line, flight path vector, and 3.degree.
glide slope indicator. The display of these symbols can be derived
from APALS navigation knowledge or from other aircraft instruments.
FIG. 8 shows a particular symbol set with the addition of the
synthetic runway image.
[0083] True ground speed information in the APALS system is
sufficient to generate moving segments in the extended center-line
to create a sensation of "speed" for the pilot.
[0084] A secondary aspect of APALS is that the X-band radar
together with APALS enhanced resolution can detect runway
incursions prior to landing in low visibility conditions. This is
accomplished with a broad sweeping ground map just prior to landing
which is similar to the "ground map mode" of a conventional weather
radar. The notable exception is that the range resolution is two
orders magnitude sharper than that of the conventional weather
radar. This allows large objects, such as a taxiing aircraft to be
resolved into more than one pixel. As a result, the APALS is able
to correctly distinguish and separate larger and smaller objects
from each other. When this is coupled with the precise navigational
knowledge of APALS, any radar returns can be related to their
precise location in the airport scene to determine if they are a
hazard.
[0085] As set forth above, the APALS system does not depend on
ground equipment installed at a particular airport. It therefore
offers the potential of low-visibility landings at many airports
than are currently unavailable for such landings because they do
not have the ground equipment of sufficient reliability to support
the automatic landing systems.
[0086] Further, APALS does not require the addition of any new
vision sensors on the aircraft or installations on the ground and
therefore installation costs are minimum. The accuracy and
reliability of the display can be checked and verified during
normal visual operations. It can also be routinely used for
training at any airport during normal visual operations. In
addition, it can detect runway obstacles prior to landing without
adding any sensors.
[0087] The display for APALS can be either head-up or
head-down.
[0088] The waveform can be varied in PRF (Pulse Repetition
Frequency), pulse width, bandwidth, and integration time to affect
changes in resolution and processing dynamic range. The Pulse
Repetition Frequency is the number of pulses per second that the
radar transmits. This is important because the PRF determines the
amount of average power that the radar receives. It also determines
what kind of ambiguities there are in range.
[0089] Those skilled in the art will understand that variations and
modifications can be made to system described above, and that such
variations and modifications are within the scope of the invention.
For example, different scene match correlation algorithms and
different navigation filters (other than Kalman) such as neural net
"intelligent" estimators can be used without changing the nature or
concept of the present invention.
* * * * *