FIELD OF THE INVENTION

This invention relates to methods and apparatus for providing precision information for airplanes during initial and terminal phases of flight, and more specifically, methods and apparatus for performing precision guidance using a global positioning system, and an inertial reference system.

BACKGROUND OF THE INVENTION

As will be understood from the following description, the present invention was developed for increasing the availability of precision approach landings using a Global Positioning System (GPS) Landing System (GLS) and an Inertial Reference System (IRS) at airports anywhere in the world.

The Automatic Landing Systems (i.e., autopilots) on today's commercial airplanes receive their guidance from a ground-based Instrument Landing System (ILS). In low weather minimums, the integrity and continuity of the ILS transmissions are absolutely crucial to the safety of the airplane during the final phase of approach, touchdown and roll-out. (“Integrity” is the probability that the signals are not hazardously misleading. “Continuity” is the probability that the signals remain present and usable during the approach). The integrity is assured by a set of near-field and far-field monitors, ready to shut down the ILS should the ILS signals move outside allowed tolerances. The continuity of the signals is assured by a backup transmitter. The backup transmitter comes on-line if the primary transmitter fails or is shut down. A key feature of today's systems is that the ground station has the sole responsibility for ensuring the integrity and continuity of its own transmissions. Because ILS equipment is costly due to initial purchase price and maintenance costs, ILSs are only practical at airfields that have large incomes generated by commercial traffic or government funding. Also, ILS signals are sensitive to local building construction and even vehicle movement. This sensitivity increases operating costs, because the ILS operators, such as the Federal Aviation Administration (FAA) in the US, must continually ensure each ILS is producing an accurate signal. Therefore, global implementation is not practical.

The GLS has been proposed as a replacement for ILS. GLS is attractive, because satellite signals are present everywhere in the world, at no cost to airports or other authorities responsible for providing airplane approach information.

In present GLSs, airplane position signals, determined from GPS signals sent by orbiting satellites, are augmented in the airplane by differential corrections (differential GPS) received from a local ground station. The differentially corrected GPS signals are referenced to an intended approach path received by the airplane from the same ground station. The ground station is also responsible for monitoring each satellite and providing airplanes with the integrity status of each satellite. The integrity and continuity of the received airplane position signals depend on the number of satellites in the airplane's field of view, the satellites' positions in the sky (their “geometry”), and the data received from the ground station. The airplane's on-board equipment must determine that the signals being received from satellites and ground station will provide a level of integrity and continuity compatible with the prevailing approach weather minimum for the duration of the approach about to be performed. There will be times and places in the world where the satellites in view cannot support the required continuity and integrity for certain approaches, such as FAA Category 3 (Cat. 3) approaches.

Even when the satellite geometry supports the required continuity and integrity, the signals received by airplanes are subject to environmental threats, such as electromagnetic interference (EMI) (both accidental and malicious), lightning and ionospheric scintillation (i.e., brown-outs associated with sunspot activity). There is also the threat of random satellite failures and satellites setting over the horizon. These threats can affect the reception of some or all of the available satellite signals, resulting in degradation or loss of guidance. Some of the threats are not well understood, and will remain so for several years.

Several methods of enhancing GLS for providing acceptable signals for Cat. 3B and 3C (autoland) approaches have been proposed. One method is to enhance the satellite constellation by making use of another country's satellite system, such as the Russian GLONASS system. This approach places an added burden on the airborne equipment and has complex political implications. Another method uses so-called “pseudolites,” ground-based transmitters, located on or near the airport, which mimic satellites by providing additional range information to the airplane. Similar to ILS, this approach is impractical, because it entails large equipment expenditures and maintenance costs in addition to those of the differential GPS ground station. Also, neither of these approaches adequately addresses the environmental and other threats described above, which may produce unreliable GLS data for an indefinite period of time.

Accordingly, a need exists for a low-cost, low-maintenance, worldwide useful, airplane precision approach guidance system that is highly accurate and reliable. The present invention combines the best features of GLS and IRS to fulfill this need.

SUMMARY OF THE INVENTION

In accordance with this invention, an airplane precision approach guidance system and method are provided. The airplane precision approach guidance system includes: (i) GPS Landing System (GLS) components for receiving and processing signals from GPS satellites and a differential GPS ground station and generating a first set of velocities; (ii) an inertial reference system for generating a second set of velocities; and (iii) guidance software for determining a cross-runway velocity and a lateral distance from runway centerline based on received runway centerline coordinates and the generated first and second set of velocities. The airplane precision approach guidance system also includes flight instruments and an autopilot system for receiving and processing the information generated by the guidance software and for guiding the airplane through approach, touchdown, and rollout. The runway centerline coordinates may be stored at the ground station or in local memory. The ground station can also provide differential GPS information and satellite health status information.

In accordance with other aspects of this invention, the guidance software can be included in (or spread between) the global positioning system, the inertial reference system or the autopilot system. That is, for example, the guidance software may be executed by the autopilot processor, i.e., the guidance software may take the form of a subroutine or program included in the autopilot and carried out by the autopilot processor. Alternatively, the guidance software may be executed by a separate, stand-alone processor.

In accordance with still other aspects of this invention, the guidance software includes a first filter for generating a first velocity in a first predefined direction based on velocities in the first predefined direction from the first and second set of velocities, a second filter for generating a second velocity in a second predefined direction based on velocities in the second predefined direction from the first and second set of velocities, and a third filter for generating a cross-runway velocity and a lateral distance from runway centerline based on received runway centerline information and the generated first and second velocities. The third filter further generates the lateral distance from runway centerline based on a lateral distance from runway centerline received from the global positioning system.

In accordance with further aspects of this invention, the first and second filters estimate and store the velocity and acceleration biases in the first and second predefined directions of the velocities received from the inertial reference system based on the velocities received from the global positioning system.

In accordance with still further aspects of this invention, the first and second filters generate the first and second velocities based on the estimated and stored velocity and acceleration biases when the velocities received from the global positioning system become unavailable or unusable.

In accordance with yet other aspects of this invention, the first, second, and third filters are Kalman filters.

In accordance with still other aspects of this invention, the first, second, and third filters are complementary filters.

In accordance with yet still other aspects of this invention, the first and second filters are second order complementary filters and the third filter is a first order complementary filter.

In accordance with other further aspects of this invention, redundant inertial reference systems, each of which produces a second set of velocities, are provided. The guidance software includes first and second filters for each second set of velocities and decides which second set of velocities to use when determining a cross-runway velocity and a lateral distance from runway centerline. Preferably, three inertial reference systems and, thus, three second sets of velocities are provided and the decision is based on which second set of velocities lies between the other two second sets of velocities.

As will be readily appreciated from the foregoing summary, the invention provides inertially smoothed GLS guidance information and the ability to coast on (use only) IRS guidance information with its velocity and acceleration biases corrected using the previously known GLS information.

BRIEF DESCRIPTION OF THE DRAWINGS

The foregoing aspects and many of the attendant advantages of this invention will become more readily appreciated as the same becomes better understood by reference to the following detailed description, when taken in conjunction with the accompanying drawings, wherein:

FIG. 1

is a schematic diagram that illustrates components external to an airplane required by the preferred embodiment of the present invention;

FIG. 2

is a block diagram illustrating the components internal to an airplane required by the preferred embodiment of the present invention;

FIG. 3

is a flow diagram illustrating a process performed by the components illustrated in

FIGS. 1 and 2

;

FIG. 4

is a block diagram illustrating filters included in the guidance software of the inertially augmented GLS illustrated in

FIG. 2

;

FIGS. 5A and 5B

are control law diagrams of the second order complementary filters illustrated in

FIG. 4

;

FIGS. 5C and 5D

are graphs of velocity information received by the inertially augmented GLS illustrated in

FIG. 2

;

FIG. 6

is a control law diagram of the first order complementary filter illustrated in

FIG. 4

;

FIG. 7

is a graph of inertial reference unit generated guidance information (velocities); and

FIG. 8

is a diagram of a voting arrangement having some undesirable features that could be used when an airplane employing the invention has redundant inertial reference systems;

FIG. 9

is a diagram of a voting arrangement, usable when an airplane employing the invention has redundant inertial reference systems, that avoids the undesirable features of the voting arrangement shown in FIG.

8

.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT

In accordance with this invention, an airplane precision approach guidance system that includes a Global Positioning System (GPS) Landing System (GLS) augmented with an Inertial Reference System (IRS) is provided. The airplane precision approach guidance system includes a GLS that includes components both internal and external to an airplane, an IRS, flight instruments, an autopilot system capable of performing automatic landings and roll-out, and a guidance process for supplying signals to the flight instruments and autopilot system. The guidance process is a program, preferably a software program. Preferably, the guidance software physically resides in the GLS receiver, the IRS or the autopilot and is executed by the processor(s) included in the system in which the software resides. That is, the guidance software may take the form of a program or subroutine included in the program that controls the operation of the GLS receiver, the IRS or the autopilot. Alternatively, the guidance software could execute on a separate, stand-alone processor. For ease of illustration and description, the guidance software is depicted and described as a separate element or component of the invention. As shown in

FIG. 1

, the components of the GLS external to an airplane

10

are a set of satellites

30

and a differential GLS groundstation

32

. As shown in

FIG. 2

, the components of the GLS that are internal to the airplane

10

include a GPS receiver

14

and a GLS processor

16

for receiving and processing signals from the satellites

30

and the differential GLS groundstation

32

. The other internal airplane components include a radio altimeter

18

, an IRS

20

, memory

22

, the guidance software

24

, flight instruments

26

, and an autopilot system

28

.

Some of today's autopilots operating within an Instrument Landing System (ILS) environment have the ability to “coast” solely on navigation data supplied by the IRS (i.e., inertial guidance) for both vertical and lateral airplane control, but only for periods of a few seconds during an ILS transmitter switch-over or a temporary interruption of the signal sent by the ground-based ILS. The present invention uses IRS data to augment an airplane's GLS so that the autopilot has the ability to coast accurately for a much longer period of time. The GLS is subject to periods of time when the GLS guidance signals are out of acceptable limits for continuing a precision approach. The causes of such an occurrence might be satellite failure, lack of an acceptable number of satellites sending signals, accidental or malicious electromagnetic interference (EMI), lightning, ionospheric scintillation (i.e., brown-outs associated with sunspot activity), or other unknown interruptions, all collectively called environmental and other threats. The period of time that the GLS may produce unacceptable guidance signals can be several minutes, thereby requiring the airplane's autopilot to coast for a period up to a minute on an approach. If the autopilot were to receive only uncorrected or partially corrected guidance information from the IRS during the period of time the GLS is producing unacceptable guidance signals, the precision of the autopilot approach would be unacceptable for landing and rollout. The precision of the autopilot approach, landing, and rollout would be unacceptable because the uncorrected or partially corrected IRS data is not sufficiently accurate to be used for the period of time that the GLS can produce unacceptable signals. This is due to the normally occurring errors in the IRS which tend to be sinusoidal and have a period of oscillation of about 84 minutes (known as the Schuler error). In summary, the uncorrected or partially corrected IRS signals, if they were to be used for the minute or so that the GLS guidance signal is unacceptable, would provide unacceptable guidance control signals to the autopilot, thereby causing the airplane to land off or roll off the runway. The problem is more acute in lateral control than in vertical control, because the autopilot must coast on inertial guidance from the time the GLS is abandoned (typically below 200 feet height above the runway) all the way through roll-out, whereas for vertical control the autopilot need only coast until the airplane's radio altimeter takes over at approximately 50 feet above the runway. While the present invention when applied to precision landing approaches is preferably applied to an airplane's lateral control, the present invention can also be applied to vertical control during all flight regimes, e.g., landings, takeoffs, weapons release, etc.

The GPS receiver

14

within the GLS

12

receives position information from the satellites

30

and receives differential position information and runway approach path coordinates from the differential GPS groundstation

32

. If the received position information is determined to be acceptable for guidance, the GLS processor

16

processes the received position information and generates GLS guidance signals. The guidance software

24

processes: (i) the GLS guidance signals, (ii) guidance information generated by the IRS

20

, and (iii) guidance information generated by the radio altimeter

18

. Based on the information it receives, the guidance software generates guidance information for the flight instruments

26

and the autopilot system

28

. When the received GLS position information is determined to be unacceptable, the lateral control guidance information reverts to IRS generated information adjusted using bias estimates calculated during the period of time when the GLS signal was known to be healthy. The type of GLS and IRS guidance information and the type of guidance information produced by the guidance software executed by the processor of the GLS receiver

12

, the IRS

20

or the autopilot system

38

, or a stand-alone guidance processor, is described in more detail below.

FIG. 3

is a functional flow diagram that illustrates the process performed by the components illustrated in

FIG. 2

to provide approach information for a precision approach and roll-out. The guidance information generated by the present invention also may be used in other flight scenarios, such as low weather minimum takeoffs. At block

100

, the airplane begins a precision approach to landing. At block

104

, the IRS

20

velocity and acceleration errors are estimated by a filter, preferably a Kalman filter or a second order complementary filter, using the GLS guidance signals. This estimation, which may take from ten (10) seconds to a minute depending upon the specific type of equipment used, is illustrated in FIGS.

4

-

6

and described in more detail below. Thereafter, the GLS processor

16

continuously determines if the satellite produced position information is acceptable. See decision block

108

. If the position information is determined to be acceptable, i.e., the integrity of the GPS position information (i.e., the satellite signals) is determined to be above a required or acceptable level, the guidance software

24

continues processing the GLS and IRS guidance information together to provide guidance information to the flight instruments

26

and autopilot system

28

until landing and rollout. See blocks

112

and

113

. However, if at decision block

108

, the GLS position information becomes unacceptable, the system determines if the position information became unacceptable before a point on the approach at which a safe landing and roll-out can be completed with only the corrected IRS guidance information. See decision block

116

.

If the position information became unacceptable after the point on the approach at which a safe landing and roll-out can be completed with corrected IRS guidance information only, the process continues the approach with only the corrected IRS guidance information and continues to test the GLS position data. See blocks

124

and

125

. As long as the position information remains unacceptable, processing remains in the loop formed by blocks

124

and

125

until landing and rollout occur. If GLS position data becomes acceptable again (block

125

), processing cycles to block

112

.

If, at decision block

116

, the position information becomes unacceptable before the point on the approach at which a safe landing and roll-out can be completed with only the corrected IRS guidance information, the approach is discontinued and a go-around is executed. The point on the approach at which safe landings and roll-outs occur using only corrected IRS guidance information can vary. The point on the approach is chosen based on the time it takes to complete the roll-out safely. The point on the approach or the time to complete roll-out is calculated by analyzing landing and roll-out lateral deviation from the centerline caused by worst case divergence of the corrected IRS guidance information. The point could be as low as 100 feet above the runway for a low precision IRS or as high as 500 feet above the runway for a high precision IRS.

FIG. 4

is a block diagram that illustrates a set of filters that are included in the guidance software

24

for determining precise lateral guidance information. As noted above and as will be readily appreciated by those skilled in information processing, rather than forming part of guidance software

24

executed on a separate processor, as might be suggested by

FIG. 3

, the filters shown in

FIG. 4

may be included in software executed by other system components, such as the GLS

12

, the IRS

20

or the autopilot system

28

.

The set of filters includes North/South (N/S), East/West (E/W) and runway filters. While various types of Kalman filters can be used, preferably the N/S and E/W filters are second order complementary filters

180

and

184

and the runway filter is a first order complementary filter

188

.

The N/S filter

180

receives GLS and IRS north/south velocities and the E/W filter

184

receives GLS and IRS east/west velocities (V

N

GLS

, V

E

GLS

, V

N

IRS

, V

E

IRS

). The GLS and IRS north/south and east/west velocities are generated by the GLS

12

and IRS

20

, respectively. The output of the N/S filter

180

is a north/south complementary velocity (V

N

C

) and the output of the E/W filter

184

is an east/west complementary velocity (V

E

C

). Suitable N/S and E/W filters

180

and

184

are shown in

FIGS. 5A and 5B

and described in more detail below.

The north/south and east/west complementary velocities (V

N

C

, V

E

C

) are received by the runway (first order) complementary filter

188

. The runway filter

188

also receives lateral distance from the runway centerline from the GLS (Y

X-RWY

) and runway (centerline) true heading (&PSgr;

TRWY

) from the GLS groundstation

32

or from memory

22

. The runway filter

188

produces the airplane's compensated lateral distance (Y

C

X-RWY

) from the runway centerline and a cross-runway velocity (V

X-RWY

) based on the received information. Y

C

X-RWY

and V

X-RWY

are sent to the autopilot system

28

and the flight instruments

26

(shown in FIG.

2

). A suitable runway filter

188

is shown in FIG.

6

and described in more detail below.

FIGS. 5A and 5B

illustrate the components of suitable N/S and E/W filters

180

and

184

as shown in FIG.

3

. As noted above, the N/S and E/W filters are second order complementary filters. It is to be understood that other types of filters, such as Kalman filters, can be used, if desired. For ease of illustration, the N/S and E/W second order complementary filters are shown in control law form. As will be readily appreciated by those familiar with airplane control systems, while often depicted in control law form for ease of illustration, complementary and other airplane control system filters are actually implemented in software designed for execution by the processors included in modem commercial and other airplanes.

The N/S second order complementary filter

180

, shown in

FIG. 5A

, receives the GLS and IRS north/south velocities, V

N

GLS

and V

N

IRS

. V

N

GLS

and V

N

IRS

are ground referenced velocities. V

N

GLS

is applied to the positive input of a subtractive combiner

200

. The output of the subtractive combiner is applied to a switch

204

. The switch

204

is opened if the received GPS position information is determined to be unacceptable and closed if the received GPS position information is determined to be acceptable. The GPS position information acceptability determination is performed by the GLS

12

based on preset requirements pertaining to received GPS position information. When switch

204

is closed, the output of the subtraction combiner is delivered to two paths. The first path includes a first gain K

1

. The second path includes a second gain K

2

and a first bipolar integrator

208

. Typical ranges for K

1

and K

2

are 0.05 to 1.0 and 0.001 to 0.5, respectively. Preferred values are 0.4 and 0.04, respectively. Values for the first and second gains K

1

and K

2

depend upon the airplane type and the types of inertial and GLS equipment used in an actual embodiment of the invention. The two paths terminate at a first additive combiner

212

. The output of the first additive combiner

212

passes through a second integrator

216

to one input of a second additive combiner

220

. V

N

IRS

is applied to the other input of the second additive combiner

220

. Thus, the second additive combiner sums the result of the second bipolar integrator

216

and the received V

N

IRS

. The output of the second additive combiner

220

is the complementary north/south velocity, V

N

C

. V

N

C

is fed back to the negative input of the subtractive combiner

200

. Thus, the subtracting combiner subtracts V

N

C

from V

N

GLS

.

Since the E/W second order complementary filter

184

shown in

FIG. 5B

is identical in form to the north/south second order complementary filter

180

, it is not described in detail, except to note that similar reference numbers with the suffix “a” are used to identify the various components and to note that the inputs are the GLS and IRS east/west velocities V

E

GLS

, and V

E

IRS

, and the result is the complementary east/west velocity V

E

C

. The K

1

and K

2

values are the same.

If the GPS position information becomes unacceptable, the switches in the N/S and E/W second order complementary filters are opened and the only input to the N/S and E/W second order complementary filters

180

and

184

becomes the IRS velocities V

N

IRS

and V

E

IRS

. Because V

N

IRS

and V

E

IRS

originate from signals generated by inertial sensor devices within the IRS, they exhibit a dominant error whose period of oscillation is approximately equal to 84 minutes. As shown in

FIG. 5C

, which is a knots (kts) versus time (t-minutes) graph for either V

N

GLS

/V

N

IRS

or V

E

GLS

/V

E

IRS

, this error, called the Schuler error, appears sinusoidal, as shown by curve

302

.

The first integrator

208

estimates and stores an acceleration bias and the second integrator

216

estimates and stores a velocity bias. The stored velocity and acceleration biases correct V

N

IRS

to be equivalent to V

N

GLS

when the received GPS information is determined to be unacceptable. This adjustment occurs as follows. As soon as the switch

204

is open, the first integrator

208

freezes at the last determined acceleration bias value. The output of the second integrator

216

, i.e., the last known velocity bias value, ramps based on the acceleration bias value stored in the first integrator

208

. Then the ramped velocity bias value is applied to the received V

N

IRS

to generate V

N

C

. As will be readily appreciated by those skilled in control law implementation, when the functions performed by the N/S and E/W filters

180

and

184

are implemented in computer code, the velocity and acceleration biases can be stored in computer memory, e.g. RAM.

The following example, using

FIG. 5C

, illustrates the operation performed by the N/S and E/W filters

180

and

184

. In this example V

N

GLS

remains steady at some predetermined speed, e.g., 200 kts, represented by a straight line

300

in FIG.

5

C. V

N

IRS

, which exhibits the Schuler error, is represented by the sinusoidally oscillating curve

302

in FIG.

5

C. If V

N

GLS

is not received because the GPS position information is unacceptable, in order to maintain accuracy, V

N

IRS

must be adjusted to equal what V

N

GLS

would have been. In order to adjust V

N

IRS

to equal V

N

GLS

, the difference in amplitude between the V

N

IRS

curve

302

and the V

N

GLS

curve

300

, and the slope of the V

N

IRS

curve

302

must be determined, prior to a loss of the GLS signal. Since the slope of the V

N

IRS

curve

302

is not a constant value, the results of adjusting V

N

IRS

over long periods of time eventually becomes inaccurate for the purpose of precision approaches. It is this inaccuracy that requires that a point (altitude) on the approach be determined above which to discontinue a precision approach and below which to continue the approach using only the adjusted IRS information. See decision block

116

,

FIG. 3. A

minute or two has been found in a worst case IRS error scenario to provide acceptably accurate information. The approximate time period to safely complete the landing roll-out using only the corrected IRS information is about {fraction (1/84)}th that of one full V

N

IRS

Schuler oscillation. The time period represented by the space between vertical lines

306

in

FIG. 5C

is the approximate maximum period of time for completing a precision approach using only adjusted IRS information.

FIG. 5D

, which is an enlargement of

FIG. 5C

between vertical lines

306

, illustrates the time period for a final approach using only corrected IRS data as defined between lines

306

shown in

FIG. 5C

using the 200 kts exemplary speed. The approach time period is about 120 seconds. Before the GLS guidance signal becomes unacceptable, the difference (amplitude) between the V

N

IRS

curve

302

and the V

N

GLS

curve

300

(i.e., velocity bias) and the slope of the V

N

IRS

curve

302

(i.e., acceleration bias) are continuously being determined. As long as acceptable GPS signals are being received, the slope and amplitude of the V

N

IRS

curve

302

are accurately determinable. When the GLS signals become interrupted or unacceptable, the last known amplitude and slope of the V

N

IRS

curve

302

are used to adjust V

N

IRS

to create an accurate V

N

C

value. Since the slope of the V

N

IRS

curve

302

is relatively constant for about 1 to 2 minutes, an accurate output velocity can be determined. For example, if the GPS signals become interrupted or unacceptable at t=12 min in

FIG. 5C

, the V

N

IRS

is about 202.5 kts and is changing at about 1 kt/min. Thus, 202.5 kts and 1 kt/min are the estimated values that are then used to adjust V

N

IRS

for the remainder of the approach if and when the GLS signals become invalid. The error or bias estimations are done in earth-fixed axes rather than airplane-fixed, because the Schuler errors remain fixed in earth axes as the airplane turns. This allows the complementary filters to settle (i.e., the bias estimates to be generated) before the airplane has turned onto final approach. Also, the slope for a V

E

IRS

curve and the amplitude difference between the V

E

IRS

curve and a V

E

GLS

curve are determined in the same manner as described above for the V

N

IRS

and V

N

GLS

curves.

The high frequency noise (frequency 0.1 to 10 Hz) that is typically present in GLS signals can cause an autopilot system receiving only GLS guidance signals to send erroneous (noisy) flight control correction signals to the flight control surfaces. This is avoided by the present invention. When the GPS position information is acceptable, the relatively steady IRS guidance information inertially smoothes the GLS information making the complementary displacement and velocity signals, Y

X-RWY

and V

X-RWY

, less noisy. The cleaner, less noisy signals cause the autopilot system

24

to provide smoother signals to the flight control surfaces.

As shown in

FIG. 6

, the runway filter

188

is a Kalman filter, specifically a first order complementary filter, that computes “cross-runway velocity”, i.e., the component of the airplane's velocity that is perpendicular to the runway's true heading based on the complemented outputs of the N/S and E/W filters

180

and

184

(

FIG. 4

) and the runway true heading &PSgr;

TRWX

. The runway true heading is the precise true heading of the approach runway and is supplied by the GLS groundstation

32

or the airplane system memory

22

. Runway true heading is of extremely high integrity and is implicit in the approach path coordinates supplied by the ground station or memory. The runway filter

188

also generates a compensated lateral distance from runway centerline value (Y

C

X-RWY

) based on the cross-runway velocity (V

X-RWY

) and a lateral distance from runway centerline value (Y

X-RWY

) generated by the GLS

12

.

Returning to

FIG. 6

, the runway (first order complementary) filter

188

performs a cross-runway velocity process

350

. The V

X-RWY

process

350

operates on the N/S and E/W filter outputs V

N

C

and V

E

C

and generates the cross-runway velocity V

X-RWY

based on the following equation:

V

X-RWY

=V

E

C

COS&PSgr;

TRWY

−V

N

C

SIN&PSgr;

TRWY

where

&PSgr;

TRWY

=Runway True Heading.

The lateral distance from runway centerline Y

XRWY

is received from the GLS

12

. Y

X-RWY

is applied to the positive input of a subtractive combiner

356

. The output of the subtractive combiner

356

passes through a switch

358

and is multiplied by a gain factor K

3

. A typical range for K

3

is 0.1 to 1.0, preferably 0.4. The result is summed with the cross-runway velocity, V

X-RWY

, in an additive combiner

360

. The output of the additive combiner is integrated by a bipolar integrator

362

and becomes the compensated lateral distance from runway centerline, V

C

X-RWY

. The compensated lateral distance from runway centerline (Y

X-RWY

) is applied to the negative input of the subtractive combiner

356

. Thus, the subtractive combiner subtracts Y

C

X-RWY

from Y

X-RWY

. When switch

358

is closed the compensated lateral distance from runway centerline (Y

C

X-RWY

) is an inertially smoothed lateral displacement value. The switch

358

, like switch

204

, opens when GPS information becomes unacceptable. When this occurs, the compensated lateral distance from runway centerline (Y

C

X-RWY

) is an integration of the cross-runway velocity (V

X-RWY

) determined with biases removed. The integrators of the N/S and E/W filters

180

and

184

store the most recently determined bias information, thereby making available the bias information when needed. As will be readily appreciated by those skilled in airplane control signals, the filtered lateral displacement and the cross-runway velocity are used by the autopilot's lateral control laws and the flight instruments in a conventional manner.

When, on an approach, the GLS data becomes invalid and the switches

204

,

204

a

and

358

open, the performance of the filters described above is transient-free and transparent to the autopilot's control laws. As a result, the airplane's response to wind, thrust asymmetry, and rudder inputs is unchanged. What has changed is that velocity and lateral distance from centerline are determined using only IRS information with biases removed. Because the adjusted IRS information was determined with a high degree of precision before the GLS was disconnected, the airplane is able to coast for many tens of seconds with no significant increase in touchdown dispersion.

A typical IRS

20

includes redundant inertial reference units (IRUs) and also typically, when multiple IRUs are present, a vote of the IRUs is performed to determine the IRS guidance information to use. One type of voting algorithm is a mid-value select, i.e., the mid-channel IRU velocity value becomes the selected value.

FIG. 7

illustrates right(R), left(L) and center(C) channel velocities generated by three redundant IRUs displayed over time. As can happen with redundant IRUs, and as shown in

FIG. 7

, the IRU that generates the mid-channel velocity can change over time. In this example, the L channel velocity is replaced by the C channel velocity because the C channel velocity has become the mid-channel velocity. When the change occurs, the acceleration bias for the old selected channel velocity (L) will be different from the newly selected channel velocity (C). If, as shown in

FIG. 8

, this voting

399

occurs prior to the application of the IRS velocity to the N/S and E/W filters, a discontinuity in the input to the runway filter can occur. As shown in

FIG. 9

, the invention avoids this undesirable result by having the voting performed after the N/S to E/W filters, before the runway filter. More specifically, each IRS “channel” includes its own N/S or E/W filter

400

,

402

or

404

. The filters may receive separate GLS velocity inputs, if redundant GLS data is available, or the same GLS data. The source of GLS data is not important since GLS data do not exhibit the kind of errors that are characteristic of IRS data. In any event, the outputs of the N/S or E/W filters

400

,

402

or

44

are voted

406

, and the center value is chosen as the data to be applied to the runway filter. In summary, each IRU's velocity and acceleration bias is estimated and corrected independently of the other IRUs. The resulting corrected velocities from the N/S and E/W filters are then voted to produce the selected corrected output. The end result of the voting arrangement shown in

FIG. 9

is a smooth transition between the old and new selected channel velocities.

The present invention is equally applicable to other flight operations, such as guidance during takeoff. The velocity complementary filters can estimate the IRS biases prior to takeoff before turning onto the active runway. If, during the takeoff roll, the GLS guidance becomes unusable, the switch-over to IRS only information, described above, is made and the autopilot or flight director steers towards an accurate adjusted centerline track. Also, as noted above, the invention can be applied to vertical as well as lateral airplane position during a precision approach landing.

While the preferred embodiment of the invention has been illustrated and described, it will be appreciated that various changes can be made therein without departing from the spirit and scope of the invention.