CN106601032B - A kind of multipath landform integrality detection method based on lower view sensor - Google Patents

Abstract

The invention discloses a multi-path terrain integrity detection method based on a downward-looking sensor, which belongs to the field of avionics systems. In order to make pilots believe that the synthetic vision system can provide accurate external scenes, certain integrity monitoring and warning capabilities are required. . The method disclosed by the invention detects in real time the degree of consistency between the airborne terrain database and the actual terrain data measured by the downward-looking sensor, and sends a warning message to the pilot when the deviation between the two exceeds a set threshold. Specifically include: airborne terrain database sampling algorithm, sensor measurement data synthetic terrain algorithm, multi-path detection algorithm and statistical testing algorithm. The method can simultaneously detect the horizontal error and the vertical error of the terrain data, and the detection accuracy is not affected by the attitude of the aircraft, the detection accuracy is high, and the real-time performance is good, which is of great significance to flight safety.

Description

A Multi-path Terrain Integrity Detection Method Based on Downward-Looking Sensor

technical field

The invention belongs to the field of avionics systems, and in particular relates to a multi-path terrain integrity detection method based on a downward-looking sensor in a synthetic vision system.

Background technique

In order for pilots to trust that the Synthetic Vision System (SVS) can provide an accurate external view without dangerous misleading information, the system requires certain integrity monitoring and warning capabilities. If the integrity check finds that the onboard terrain database does not match the actual terrain, the display system will warn the pilot that the synthetic view is unreliable and cannot be used, so as to reduce the risk of misleading terrain information provided to the pilot.

The purpose of terrain integrity detection is to detect in real time the consistency between the airborne terrain database and the actual terrain data measured by the downward-looking sensor, rather than the correctness of the airborne terrain database. The key technologies are, first, the establishment of the sampling model of the airborne terrain database. The model should be as close as possible to the measurement characteristics of the radar altimeter to ensure the detection accuracy. The second is how to test the vertical error and the horizontal transfer error at the same time. The third is to reduce the alarm time , and send warning information to the pilot in time to ensure flight safety.

Contents of the invention

The purpose of the present invention is to solve the deficiency of the existing terrain integrity detection method in the synthetic vision system, and proposes a multi-path terrain integrity detection method based on a downward-looking sensor.

The multi-path terrain integrity detection method based on downward-looking sensors of the present invention comprises the following steps:

Step 1: Airborne terrain database sampling;

Step 2: Calculate the synthetic terrain height;

Step 3: Calculate the statistical test quantity;

Step 4: multipath detection;

Step 5: Hypothesis testing, if the inconsistency between the airborne terrain database and the actual terrain data measured by the down-looking sensor exceeds the set threshold, a warning message will be sent to the pilot.

The advantages of the present invention are:

(1) In the existing airborne terrain database sampling algorithms, most of the sampling models assume that the radar altimeter measures the vertical height directly below the aircraft. This limitation causes sampling errors in the pitch and roll attitudes of the aircraft, and This kind of error is easy to appear in the situation where the integrity of the terrain is high, such as the take-off and landing of the aircraft. The sampling model proposed in the present invention overcomes this limitation and maintains high-precision terrain integrity detection no matter what attitude the aircraft is in;

(2) In the present invention, when the airborne terrain database is sampled, intersection detection and bilinear difference are used, and the efficiency is maximized while ensuring the sampling accuracy;

(3) Most of the existing terrain integrity detections use single-path detection, which can only detect vertical errors in the terrain database, especially when the terrain is flat, and cannot detect horizontal transfer errors in the terrain database. The multi-path detection used in the present invention provides A method for detecting horizontal errors in terrain database is proposed, which detects vertical errors and horizontal errors simultaneously, and improves the detection ability of terrain integrity.

Description of drawings

Fig. 1 is a flow chart of the present invention;

Fig. 2 is a schematic diagram of the sampling model of the airborne terrain database;

Fig. 3 is a schematic diagram of synthetic terrain measurement;

Fig. 4 is an event tree of terrain integrity detection in synthetic vision system.

Detailed ways

The present invention will be further described in detail with reference to the accompanying drawings and embodiments.

The multi-path terrain integrity detection method based on the downward-looking sensor of the present invention detects in real time the degree of consistency between the airborne terrain database and the actual terrain data measured by the downward-looking sensor through multi-path detection and statistical algorithms. When the threshold is set, a warning message is sent to the pilot.

A kind of multi-path terrain integrity detection method based on down-looking sensor of the present invention, as shown in Figure 1, comprises the following steps:

Step 1: Airborne terrain database sampling.

As shown in Figure 2, in the synthetic vision system, the installation position of the radar altimeter is taken as the origin, within the range of the radar altimeter beam, the ray is sent along the direction of the radar altimeter antenna, and the intersection detection is performed with the terrain. The closest point to the aircraft is used as a sampling point for comparison with the measured value of the radar altimeter, and the height value h _DEM (lat _DGPS (t _i ), lon _DGPS (t _i )) of the sampling point is calculated using the bilinear interpolation method, And then get the terrain profile of the terrain database.

Among them, the ray is emitted along the direction of the radar altimeter antenna, and the sampling point closest to the aircraft is selected instead of the sampling point in the vertical direction of the aircraft, in order to make the sampling model closer to the measurement characteristics of the radar altimeter. To calculate the height value of sampling points, in addition to the bilinear interpolation algorithm, the nearest neighbor interpolation method and the cubic convolution interpolation method can also be selected, but the accuracy of the nearest neighbor interpolation method is low, and the calculation amount of the cubic convolution interpolation method is too large. In order to meet the requirements of performance and precision, the present invention chooses bilinear interpolation.

Step 2: Calculate the synthetic terrain height.

As shown in Figure 3, the synthetic terrain profile is the ground height above sea level:

h _synth (t _i )＝h _DGPS (t _i )-(h _radicalt (t _i )+h _antenna )

Among them, h _DGPS is the altitude value of the aircraft above sea level provided by DGPS; h _antenna is the height difference between the top of the GPS antenna and the low end of the radar altimeter antenna; h _radalt is the measured value of the radar altimeter.

Step 3: Calculate the statistical test quantity.

(1) The absolute error p(t _i ) between the synthetic terrain profile measured by the radar altimeter and the sampled terrain profile of the terrain database is defined as:

p(t _i )＝h _SYNT (t _i )-h _DEM (lat _DGPS ,lon _DGPS )

Among them, h _DEM (lat _DGPS , lon _DGPS ) is the height of the terrain in the terrain database, and h _SYNT (t _i ) is the synthetic height of the terrain.

(2) Carry out Kalman filtering on the absolute error p(t _i ), reduce the rated noise in the sensor and Digital Elevation Model (DEM) data and estimate the deviation between the sensor and DEM, the specific steps are as follows:

1) Initialize the predicted value and error variance

in, According to the assumption that the system model is 0, The value range of is ((15) ² ,(20) ² );

Compute the Kalman filter gain:

Among them, K _k is the Kalman filter gain, is the error variance of the previous state, H _k is the unit field transformation matrix, R _k is the measured value of the error variance, which is a constant;

2) Update the estimated value of the sampled value:

in, is the estimated value at time t _k , is the estimated value of the last state, z _k is the measured value at time t _k ;

3) Calculate the error variance of the estimate

where P _k is the error variance of the estimated value;

4) Calculate the predicted value

in, is the predicted value at time t _k+1 , Φ _k is the unit state transition matrix, is the prediction error variance at time t _k+1 , Q _k is the system noise variance, which is called the tuning parameter of the filter and is a constant;

5) Return to 2) and repeat the above process until the system is running.

(3) Define the statistical test volume:

Among them, T is the statistical test quantity, P is the estimated value of the error variance, N is the number of sampling points, is the absolute error after Kalman filtering.

Step 4: Multipath detection.

(1) Simulate the horizontal DEM transfer error, and shift the flight path horizontally to obtain multiple flight paths. According to the method in step 3, calculate the multi-path statistical inspection quantity:

Among them, T(m,n) is the statistical test quantity;

lat _DGPS ＝lat _DGPS (t _i )+lat_off _m m∈(1,M);

lon _DGPS = lon _DGPS (t _i ) + lon_off _n n∈(1,N).

(2) Calculate the statistical test quantity T in 1, and the offset path with the minimum T value is considered to be the current horizontal position of the aircraft in the terrain database, which is defined as:

Among them, the min() function returns the estimated position, is the minimum value of the statistic T.

Define T _V (t _i ) = T _min as the vertical error inspection quantity. T _min is the minimum value of the test quantity.

(3) The estimated position is defined as the estimated position lon _DBP and lat _DBP for the terrain database; the real position is defined as lon _true and lat _true , and the horizontal error between the two is calculated:

p _H ＝dist([lon _DBP ,lat _DBP ],[lon _true ,lat _true ])

Among them, dist() returns the distance between two points in the space; lon _true and lat _true are the measured positions of DGPS.

(4) Calculate the statistical test quantity of horizontal error.

where _σp is the standard deviation.

Step 5: Carry out hypothesis testing, and the event tree is shown in Figure 4. If the inconsistency between the airborne terrain database and the actual terrain data measured by the downward-looking sensor exceeds a set threshold, a warning message will be sent to the pilot.

(1) Put forward two competing hypotheses:

Null hypothesis H ₀ : the system works in the rated state or no operating errors are found;

in, Obey the normal distribution with a mean of 0 and a standard deviation of _σp ;

Alternative hypothesis H ₁ : Errors are found during system operation;

in, It obeys the normal distribution with mean value μ _B and standard deviation σ _p .

(2) According to the three different applications of the synthetic vision system as auxiliary, key elements, and tactical elements, three safety thresholds are set for the probability of missed alarms to determine the probability of false alarms P _FFD and missed alarms P _MD . According to the requirements of different grades, the probability of missing alarms is respectively set as: less than 10 ^-3 , 10 ^-4 ～10 ^-7 , 10 ^-6 ～10 ^-9 , where

P _FFD = P(detected error|H ₀ )·P(H ₀ )

P _MD =P(no error detected|H ₁ )·P(H ₁ )

Through the actual flight experiment, the test threshold _TVD and T _HD values of the appropriate vertical error test amount and horizontal error test amount are taken to meet the requirements of P _FFD and P _MD .

(3) Test the vertical error test quantity and the horizontal error test quantity respectively and make a decision. If it is smaller than the vertical test threshold T and horizontal test threshold T _H in step 3 and step 4, then accept H ₀ , otherwise, reject it H ₀ .

(4) When H ₀ is rejected, the synthetic vision system sends a warning message to the pilot, and when H ₀ is received, it means that the system is operating normally.

The specific embodiments described above have further described the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above descriptions are only specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims (2)

1. A multi-path terrain integrity detection method based on downward looking sensor, comprising the following steps: Step 1: Airborne terrain database sampling; In the synthetic vision system, the installation position of the radar altimeter is taken as the origin, within the beam range of the radar altimeter, the ray is emitted along the direction of the radar altimeter antenna, and the intersection detection is performed with the terrain, and the closest point to the aircraft is taken as Comparing the sampling point with the radar altimeter measurement value, calculate the sampling point height value h _DEM (lat _DGPS (t _i ), lon _DGPS (t _i )), where, lat _DGPS (t _i ) represents the latitude measurement at time t _i value, lon _DGPS (t _i ) represents the longitude measurement value at time t _i ; Step 2: Calculate the synthetic terrain height; The synthetic terrain profile is the ground height above sea level: h _synth (t _i )＝h _DGPS (t _i )-(h _radicalt (t _i )+h _antenna ) Among them, h _DGPS is the height value of the aircraft above sea level; h _antenna is the height difference between the top of the GPS antenna and the low end of the radar altimeter antenna; h _radalt is the measured value of the radar altimeter; Step 3: Calculate the vertical statistical test quantity; (1) Obtain the absolute error p(t _i ) between the synthetic terrain profile measured by the radar altimeter and the sampled terrain profile of the terrain database: p(t _i )＝h _SYNT (t _i )-h _DEM (lat _DGPS ,lon _DGPS ) Among them, h _DEM (lat _DGPS , lon _DGPS ) is the height of the terrain in the terrain database, and h _SYNT (t _i ) is the synthetic height of the terrain; (2) Kalman filtering is performed on the absolute error p(t _i ); (3) Obtain the vertical statistical test volume: Among them, T is the vertical statistical test quantity, P is the estimated value of the error variance, N is the number of sampling points, is the absolute error after Kalman filtering; Step 4: multipath detection; (1) Simulate the horizontal DEM transfer error, shift the flight path horizontally to obtain multiple flight paths, and calculate the multi-path statistical inspection quantity: Among them, T(m,n) is the statistical test quantity; M and N represent the maximum value of latitude and longitude offset respectively; lat _DGPS ＝lat _DGPS (t _i )+lat_off _m m∈(1,M); lon _DGPS ＝lon _DGPS (t _i )+lon_off _n n∈(1,N); (2) Let the offset path of the minimum T value be the current horizontal position of the aircraft in the terrain database: Among them, the min() function returns the estimated position, is the minimum value of the statistic T; (3) The latitude and longitude of the estimated position are the latitude and longitude of the estimated position of the terrain database, which are defined as lon _DBP and lat _DBP respectively; the longitude and latitude of the real position are respectively lon _true and lat _true , and the level between the two is calculated error: p _H ＝dist([lon _DBP ,lat _DBP ],[lon _true ,lat _true ]) Among them, dist() returns the distance between two points in the space; (4) Calculate the statistical test quantity of horizontal error; Among them: σ _p is the standard deviation; Step 5: Carry out hypothesis testing, if the inconsistency between the airborne terrain database and the actual terrain data measured by the downward-looking sensor exceeds the set threshold, a warning message will be sent to the pilot; (1) Put forward two competing hypotheses: Null hypothesis H ₀ : the synthetic vision system works in the rated state or no operating errors are found; in, Obey the normal distribution with a mean of 0 and a standard deviation of _σp ; Alternative Hypothesis H ₁ : The Synthetic Vision System found an error during operation; in, Obey the normal distribution with mean value μ _B and standard deviation σ _p ; (2) Determine the false alarm P _FFD and the probability of missing alarm P _MD of the synthetic vision system: P _FFD = P(detected error|H ₀ )·P(H ₀ ) P _MD =P(no error detected|H ₁ )·P(H ₁ ) (3) Test the vertical error test quantity and the horizontal error test quantity respectively and make a decision. If it is smaller than the vertical test quantity and horizontal test quantity in step 3 and step 4, then accept H ₀ , otherwise, reject H ₀ ; (4) When H ₀ is rejected, the synthetic vision system sends a warning message to the pilot, and when H ₀ is received, it means that the synthetic vision system is operating normally. 2. a kind of multi-path terrain integrity detection method based on downward looking sensor according to claim 1, (2) in described step 3 is specifically: 1) Initialize the predicted value and error variance According to the assumption that the system model is 0, The value range of is ((15) ² ,(20) ² ); Compute the Kalman filter gain: where K _k is the Kalman filter gain, is the error variance of the previous state, H _k is the unit domain transformation matrix, and R _k is the measured value of the error variance; 2) Update the estimated value of the sampled value: in, is the estimated value at time t _k , is the estimated value of the last state, z _k is the measured value at time t _k ; 3) Calculate the error variance of the estimate where P _k is the error variance of the estimated value; 4) Calculate the predicted value in, is the predicted value at time t _k+1 , Φ _k is the unit state transition matrix, is the prediction error variance at time t _k+1 , Q _k is the system noise variance, which is called the tuning parameter of the filter; 5) Return to step 2) until the end.