CN114627687B - Helicopter ground proximity warning method for predicting escape trajectory based on neural network - Google Patents

Abstract

The invention discloses a helicopter ground proximity warning method based on escape trajectory prediction of a neural network, which comprises the following steps of 1, substituting the current manipulated variable, flight state parameters and escape transformation modes of a helicopter into a neural network model, and predicting a plurality of escape trajectories of the helicopter in different transformation directions in real time; step 2, predicting potential collision threats of forward-looking terrains by combining a terrain elevation database based on the escape tracks predicted in real time to generate a terrain envelope; step 3, making a decision for modifying and judging the potential ground collision threat by comparing whether the escape tracks intersect with the corresponding terrain envelope lines or not; and 4, carrying out alarm prompt based on the decision-making result. The method solves the problems that HTAWS is installed on helicopters flying at low altitude and ultra-low altitude, and the success rate of alarm is low, the false alarm rate is high and the like easily caused in the flying process; the problems that the escape track is difficult to realize online accurate prediction, the flight dynamics model is complex to build and the like are solved.

Description

A Helicopter Ground Proximity Warning Method Based on Neural Network Prediction of Escape Trajectories

technical field

The invention belongs to the technical field of helicopter flight safety, and in particular relates to a helicopter ground proximity warning method based on neural network prediction escape trajectory.

Background technique

During the flight of the aircraft, the crash accident caused by the pilot's failure to perceive the dangerous approach of the surrounding terrain or obstacles in time is called Controlled Flight into Terrain (CFIT). CFIT has always been a modern One of the main causes of aircraft accidents.

In the 1970s, in order to prevent the occurrence of CFIT accidents, the industry launched the Ground Proximity Warning System (GPWS) suitable for civil aviation airliners. In the 1980s, after civil aviation airliners were required to be equipped with GPWS, the number of CFIT accidents decreased significantly, but it was still the main cause of aviation accidents. However, GPWS has exposed some problems in the process of using, and there is room for further improvement. In order to eliminate the shortcomings of GPWS, the industry launched the Terrain Awareness Warning System (TAWS) in 1998, also known as the Enhanced Ground Proximity Warning System (EGPWS). While maintaining the original advantages of GPWS, TAWS adds new functions such as forward-looking terrain warning and terrain display. Since the launch of TAWS, the number of CFIT accidents worldwide has been further reduced every year. Statistics show that TAWS can effectively prevent CFIT accidents.

Helicopters often fly in low-altitude and ultra-low-altitude areas with complex geographical environments, and CFIT is also an important cause of flight accidents. With the successful application of TAWS in civil aviation aircraft, people began to consider transplanting this system to helicopters. However, compared with civil aviation airliners, helicopters are very different in terms of mechanical structure, maneuvering mode, and flight performance. If TAWS suitable for civil aviation airliners is directly installed on helicopters, it will not be able to effectively provide terrain collision avoidance warning. , On the contrary, it will bring a series of problems such as excessive false alarm rate. Therefore, it is necessary to study a reasonable and effective warning method according to the flight performance and flight characteristics of the helicopter.

In this context, European and American avionics manufacturers represented by Honeywell of the United States have launched the Helicopter Terrain Awareness Warning System (HTAWS). The principle of HTAWS forward-looking warning is similar to that of TAWS. According to the flight performance of the helicopter, a virtual warning boundary is generated in the space of its forward direction. The warning boundary consists of four parts: forward-looking boundary, downward-looking boundary, upward-looking boundary and side boundary. The terrain data information in front of the helicopter is obtained according to the terrain elevation database, and the spatial position relationship between the warning boundary and the front terrain is compared in real time. When the warning boundary intersects the front terrain, an alarm is triggered, and the system gives an alarm prompt at the same time.

The introduction of HTAWS has greatly reduced the frequency of helicopter CFIT accidents. However, in order to consider the versatility of the current helicopter forward-looking warning algorithm, the design of the warning boundary is relatively conservative, and a high safety threshold is usually set. Low success rate and high false alarm rate. When helicopters, especially military armed helicopters, perform low-altitude and ultra-low-altitude flight missions, while ensuring their own flight safety, they also need to ensure the smooth execution of flight missions. Therefore, HTAWS is not suitable for helicopters with strong maneuverability, especially military armed helicopters, which greatly limits the performance of helicopters. Therefore, it is necessary to optimize the design of the warning method in combination with the flight performance and flight requirements of the helicopter.

At present, the helicopter ground proximity warning method is often designed based on the escape trajectory, and the accurate prediction of the helicopter escape trajectory is the key factor affecting the success of the warning method. There are two methods for calculating the escape trajectory of the helicopter. One is the curve fitting method based on the flight test data. The common curve fitting methods include the elliptical trajectory method and the parabolic trajectory method. Situations with certain roll angles are unpredictable. The other is the integral method of motion equations based on the flight dynamics model. This method can calculate the escape trajectory of the helicopter in any flight state, but this method has high requirements for the accuracy of the flight dynamics model and the speed of trim calculation. Fixed-wing aircraft The dynamics model of is relatively simple, and pre-applied manipulation can obtain good real-time trajectory prediction effect. However, the complex control system of the helicopter itself, coupled with external environmental factors, makes the flight dynamics model to be built very complicated, and it is impossible to predict the high-precision escape trajectory of the helicopter in real time. Secondly, for different types of helicopters, the flight dynamics models are quite different, resulting in poor portability and versatility of the models.

Contents of the invention

In view of the above defects, the present invention provides a helicopter ground proximity warning method based on neural network prediction of escape trajectory, aiming to solve the problem that HTAWS is installed in low-altitude and ultra-low-altitude helicopters, which may easily cause low alarm success rate and false alarm during flight. issues such as higher rates.

In addition, the method for predicting the escape trajectory of the helicopter based on the neural network proposed by the present invention,

A kind of helicopter ground proximity warning method based on neural network prediction escape track, comprises the following steps,

Step 1, training the neural network model, substituting the current control amount of the helicopter, flight state parameters and escape recovery methods into the trained neural network model, and predicting several escape trajectories of the helicopter under different recovery directions in real time with a certain frequency;

Step 2, based on the escape trajectory predicted in step 1 in real time, combined with the terrain elevation database, predict the potential collision threat of the forward-looking terrain within a certain range, and generate a terrain envelope;

Step 3, by comparing whether several escape trajectories intersect with the corresponding terrain envelope to judge the potential ground collision threat;

Step 4: Based on the recovery decision result in step 3, an alarm prompt is given in three ways: sound, light, and display.

Preferably, the neural network model is trained based on real flight test data or flight simulation data of the helicopter, and error and accuracy are selected as evaluation indicators.

Preferably, the error and accuracy indicators are respectively:

Traj_E=ABS(Traj_Model−Traj_True),

Traj_Acc=(1-Traj_E/Traj_True)*100%,

Among them, Traj_E represents the absolute value error between the escape trajectory Traj_Model predicted based on the neural network model and the escape trajectory Traj_True obtained based on the real flight test data of the helicopter, and Traj_Acc represents the relative accuracy between Traj_Model and Traj_True.

Preferably, the escape trajectory of the helicopter is related to the recovery maneuver adopted, and the recovery maneuver includes direct pull-up recovery and roll-up recovery, which correspond to the direct pull-up recovery trajectory and roll-up recovery respectively. The recovery track, wherein the roll-up recovery track includes a left roll-up recovery track and a right roll-up recovery track.

Preferably, step 2 is specifically:

Step 2.1, determine the terrain scanning range;

Step 2.2, within the scanning range, combined with the terrain elevation database data, real-time elevation extraction is performed on the terrain below the helicopter escape trajectory at a certain frequency, and the maximum value of the terrain elevation under a fixed step size is obtained, and used as the terrain under this step size Elevation value, generate terrain elevation profile;

In step 2.3, the vertical safety threshold is superimposed on the basis of the terrain elevation profile to finally generate the terrain envelope.

As a preference, the vertical safety threshold should not be a fixed value, and it needs to take into account the trajectory prediction uncertainty, navigation positioning uncertainty and terrain database vertical uncertainty, and its calculation formula is:

SCAN_Vert=NAV+(DEM+TPA)/2

Among them, NAV is the uncertainty of navigation positioning, DEM is the vertical uncertainty of terrain database, and TPA is the uncertainty of trajectory prediction.

Preferably, step 3 is specifically as follows: sequentially compare the escape trajectory with the terrain envelope below the trajectory, and if the comparison results of several escape trajectories in different recovery directions and the terrain envelope do not meet the requirements, then the recovery decision-making judgment There is a potential risk of the helicopter hitting the ground, otherwise the helicopter will continue to perform the current flight mission.

Preferably, when the escape trajectory intersects the terrain envelope, the direction corresponding to the escape trajectory is no longer considered to be an effective terrain avoidance option, and the comparison result is determined to be unqualified.

Preferably, step 4 specifically includes: giving a warning prompt based on the recovery decision and judgment obtained in step 3, and the warning prompt includes a light warning, a voice warning and a display warning.

The invention discloses a helicopter ground proximity warning system based on neural network prediction escape trajectory, which includes an escape trajectory prediction module, a collision threat prediction module, a recovery decision module and an alarm prompt module. The escape trajectory prediction module is based on the current parameters of the helicopter and In the way of escape and recovery, the neural network model is substituted into the escape trajectory of the helicopter and input into the recovery decision-making module. The collision threat prediction module scans the terrain below the escape trajectory of the helicopter to obtain the terrain elevation profile, and generates a terrain package on this basis. line and input it into the recovery decision-making module; the recovery decision-making module compares in real time whether the escape trajectory intersects with the corresponding terrain envelope to judge the potential ground collision threat, and prompts the warning through the warning prompt module. Change the results of decision-making and judgment.

The beneficial effects of the present invention are:

(1) The helicopter ground proximity warning method based on neural network prediction escape trajectory proposed by the present invention can be applied to helicopters performing low-altitude and ultra-low-altitude flight missions, and while ensuring its flight safety, the helicopter is allowed to perform the current flight mission to the greatest extent;

(2) Compared with the HTAWS forward-looking warning method, the warning success rate of the helicopter warning method proposed by the present invention is as high as 99.95%, and the false alarm rate is as low as 0.97%, which is more suitable for helicopters performing low-altitude and ultra-low-altitude flight tasks;

(3) Compared with the traditional trajectory prediction method, the method of predicting escape trajectory based on neural network has higher calculation accuracy, fewer operations, lower complexity, better real-time performance, and good trajectory prediction accuracy and algorithm versatility. In the present invention, the escape trajectory predicted by the neural network is consistent with the real flight test data, and the prediction accuracy of the escape trajectory is relatively good, and the trajectory prediction accuracy can basically reach 95% within the range of the cruise speed of the helicopter;

(4) Compared with the traditional helicopter escape trajectory prediction method, the present invention solves the problems that the escape trajectory is difficult to realize online accurate prediction and the flight dynamics model is complicated to build.

Description of drawings

In order to illustrate the technical solutions in the embodiments of the present invention more clearly, the following will briefly introduce the drawings that are used in the embodiments.

Fig. 1 is the schematic diagram of the helicopter ground proximity warning method based on the neural network prediction escape trajectory of an embodiment of the present invention;

Fig. 2 is a neural network frame diagram applicable to escape trajectory prediction according to an embodiment of the present invention;

Fig. 3 is an error and accuracy test result diagram of predicting the escape trajectory based on the neural network under different airspeeds according to an embodiment of the present invention;

Fig. 4 is a schematic diagram of collision threat prediction according to an embodiment of the present invention;

Fig. 5 is a schematic diagram of a recovery decision judgment according to an embodiment of the present invention.

Detailed ways

The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

Embodiments of the present invention provide a helicopter ground proximity warning method based on neural network prediction escape trajectory, as shown in Figure 1, which is a schematic diagram of the alarm method of the present invention, including an escape trajectory prediction module (101), a collision threat prediction module (102 ), recovery decision-making module (103), warning prompt module (104). Among them, the escape trajectory prediction module (101) is substituted into the neural network model to calculate the helicopter escape trajectory in real time at a certain frequency according to the current control amount of the helicopter, the flight state parameters and the determined escape recovery method; the collision threat prediction module (102) uses the terrain elevation database Scan the terrain below the escape trajectory of the helicopter to obtain the terrain elevation profile, and superimpose the vertical safety threshold on this basis to generate the terrain envelope; the recovery decision module (103) detects whether the escape trajectory and the terrain envelope below the trajectory are real-time with a certain frequency. Intersect to carry out decision-making judgment on the potential ground collision threat; the warning prompt module (104) predicts the collision prediction result and the decision-making for recovery through the prompt mode of light warning (105), voice warning (106) and display warning (107) A warning is recommended.

A kind of helicopter ground proximity warning method based on neural network prediction escape trajectory that the present invention proposes, specifically comprises the following steps:

Step 1, in the escape trajectory prediction module (101), according to the current control amount of the helicopter, the flight state parameters and the determined escape recovery mode, the neural network model is substituted to solve the helicopter escape trajectory. During the flight of the helicopter, the escape trajectory of the helicopter is calculated in real time at a certain frequency. The escape trajectory refers to the trajectory of the helicopter after taking certain maneuvering measures based on the current flight state. There are two common ways for the helicopter to recover from the maneuver. Pull out and roll up to recover out. The present invention determines three escape and recovery modes, and simultaneously calculates three escape and recovery trajectories according to the current flight state parameters of the helicopter, that is, directly pulling up the recovery trajectory, rolling to the left and pulling up the recovery trajectory, and rolling to the right to pull up the recovery trajectory. track.

As shown in FIG. 2 , it is a neural network frame diagram suitable for escape trajectory prediction proposed by the present invention. First, the sample data required by the neural network model is obtained through the real flight test data of the helicopter, and then the data is screened to determine the input and output of the neural network model, and finally the sample data is segmented and normalized to obtain the training sample data and test sample data. The input of the neural network model is mainly composed of helicopter control input and flight state parameters, and the output of the neural network model is mainly composed of relative positioning information of escape trajectory prediction points. In particular, if the real flight test data of the helicopter cannot be obtained during the calculation process, the flight simulation data calculated by the helicopter flight dynamics model can be used instead.

Use the pre-processed training sample data to train the neural network model, select the appropriate learning rate of the neural network model, the number of hidden layer nodes and the learning step size, and make the neural network model converge through iterative operations, so that the helicopter can be accurately and quickly predicted A neural network model of escape trajectories.

Select error and accuracy as the evaluation index of neural network model, test the neural network model that has been trained by test sample data, test the accuracy of predicting the escape trajectory of helicopter based on neural network model, the different airspeeds of the present invention based on neural network The error and accuracy test results of network prediction escape trajectory are shown in Figure 3. It can be seen that the prediction effect is better in the range of cruising speed, and the prediction accuracy can basically reach 95%. The calculation formulas (1) and (2) of the escape trajectory error and accuracy based on the neural network are as follows:

Traj_E＝ABS(Traj_Model-Traj_True) (1)

Traj_Acc＝(1-Traj_E/Traj_True)*100% (2)

Among them, Traj_E represents the absolute value error between the escape trajectory Traj_Model predicted based on the neural network model and the escape trajectory Traj_True obtained based on the real flight test data of the helicopter, and Traj_Acc represents the relative accuracy between Traj_Model and Traj_True.

Step 2, in the collision threat prediction module (102), based on the three helicopter escape trajectories calculated by the escape trajectory prediction module (101), combined with the terrain elevation database, the terrain contour in front of the helicopter is extracted by the terrain scanning method, and the terrain envelope is generated , a schematic diagram of collision threat prediction according to an embodiment of the present invention is shown in FIG. 4 . The terrain scanning method reads the terrain elevation data within a certain range ahead based on the current position of the helicopter to determine potential dangerous terrain areas. Among them, the key of the terrain scanning method lies in the determination of the terrain scanning range. Helicopter flight environment, navigation positioning uncertainty, and trajectory prediction uncertainty are all important factors affecting the terrain scanning range.

According to the determined terrain scanning range, combined with the terrain elevation database, the terrain below the three helicopter escape trajectories calculated by the escape trajectory prediction module (101) is extracted in real time at a certain frequency, and the escape trajectory prediction point at a fixed step is calculated. The maximum value of the terrain elevation within a certain range is determined as the terrain elevation value under this step, and the terrain elevation profile is generated.

There is a certain vertical uncertainty in the terrain elevation database, and the vertical safety threshold should be superimposed on the terrain scanning results to ensure the safety of helicopter flight to the greatest extent. Considering the real flight situation of the helicopter and the requirements of combat missions, the vertical safety threshold should not be a fixed value, and it needs to take into account the trajectory prediction uncertainty, navigation positioning uncertainty and terrain database vertical uncertainty. The terrain vertical safety threshold calculation formula (3) is as follows:

SCAN_Vert＝NAV+(DEM+TPA)/2 (3)

Among them, NAV is the uncertainty of navigation positioning, DEM is the vertical uncertainty of terrain database, and TPA is the uncertainty of trajectory prediction.

The terrain envelope in the helicopter's forward-sight area is generated according to the terrain elevation after superimposing the vertical safety threshold.

Step 3, in the recovery decision-making module (103), the potential ground collision is determined by detecting whether the three helicopter escape trajectories calculated by the escape trajectory prediction module (101) intersect with the terrain envelope generated by the collision threat prediction module (102) Threats make recovery decisions and judgments, and a schematic diagram of recovery decisions and judgments in an embodiment of the present invention is shown in FIG. 5 . Among them, it is determined that when the three escape trajectories predicted by the helicopter intersect with the terrain envelope below the trajectory, the system will give an alarm, so as to ensure flight safety and not affect the helicopter's current flight mission as much as possible. During the recovery decision-making and judgment process, the three escape trajectories and the terrain envelopes below them are updated in real time at a certain frequency, and the recovery decision-making judgment is also carried out in real time at a corresponding frequency.

Considering that the recovery decision is based on the prediction of three possible escape trajectories, and the possibility of collision of the three predicted escape trajectories is judged, a relatively safe path is selected as the recovery decision suggestion. When any of the three escape trajectories intersects its corresponding terrain envelope, that direction is no longer considered a valid terrain avoidance option and the helicopter continues its current flight mission. When two escape trajectories are no longer valid options, the helicopter will still be allowed to fly the current mission because the third escape trajectory still provides the helicopter with an effective means of terrain avoidance. Only when the last escape trajectory intersects the terrain envelope, the recovery decision judges that the helicopter has a potential risk of collision with the terrain.

Step 4, in the warning prompt module (104), based on the recovery decision-making suggestion given by the recovery decision-making module (103), through the light warning (105), voice warning (106), display warning (107) mode to the collision The prediction result and the direction of recovery will be alerted. Among them, while the recovery decision-making module gives suggestions for the recovery decision, the alarm prompt module issues an alarm through a combination of sound, light, and display alarms.

When the warning prompt module (104) receives a potential collision risk, the light warning prompts the prediction result of the potential collision. If the three predicted escape trajectories all intersect with the terrain envelope below it, the warning light displays red, otherwise it displays green. The voice warning prompts the recovery direction of the helicopter to avoid the terrain, such as "pull up directly", "roll to the left" and "roll to the right". Display the warning by displaying the avoidance indicator arrow on the multi-function display (MFD) or head-up display (HUD). If it is judged that there is a potential risk of collision with the ground, the arrow representing the escape trajectory of the last warning will be displayed on the screen.

The route is randomly generated based on different test geographical environments and latitude and longitude intervals, and the ground proximity warning method is simulated and tested within the helicopter cruising speed range. The statistical analysis results are averaged as the final test results. Table 1 is the simulation test result of the helicopter ground proximity warning method based on neural network prediction escape trajectory proposed by the present invention, as can be seen, compared with the traditional HTAWS forward-looking warning method, the helicopter ground proximity warning method proposed by the present invention has Higher alarm success rate, lower false alarm rate.

Table 1

Finally, it should be noted that: the above is only a preferred embodiment of the present invention, and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, for those skilled in the art, it still The technical solutions recorded in the foregoing embodiments may be modified, or some technical features thereof may be equivalently replaced. 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 (8)

1. A helicopter ground proximity warning method based on a neural network prediction escape trajectory is characterized by comprising the following steps:

step 1, training a neural network model, substituting the current operation amount, flight state parameters and escape changing modes of the helicopter into the trained neural network model, and predicting a plurality of escape tracks of the helicopter in different changing directions in real time at a certain frequency, wherein the escape tracks of the helicopter are related to the adopted changing maneuvering modes, the changing maneuvering modes comprise direct pull-up changing and roll-pull changing, and respectively correspond to direct pull-up changing tracks and roll-pull changing tracks, and the roll-pull changing tracks comprise left roll-pull changing tracks and right roll-pull changing tracks;

step 2, predicting potential collision threats of forward-looking terrains in a certain range by combining a terrain elevation database based on the escape tracks predicted in real time in the step 1, and generating a terrain envelope, specifically: step 2.1, determining a terrain scanning range;

step 2.2, in a scanning range, combining terrain elevation database data, performing elevation extraction on the terrain below an escape track of the helicopter at a certain frequency in real time to obtain the maximum value of the terrain elevation under a fixed step length, and taking the maximum value as the terrain elevation value under the step length to generate a terrain elevation profile;

step 2.3, superposing a vertical safety threshold on the basis of the terrain elevation profile to finally generate a terrain envelope;

step 3, making a decision for modifying and judging the potential ground collision threat by comparing whether the escape tracks intersect with the corresponding terrain envelope lines or not;

and 4, carrying out alarm prompt based on the decision changing result of the step 3.

2. The helicopter ground proximity warning method based on neural network prediction escape trajectory of claim 1, characterized by that, the neural network model is trained based on helicopter real flight test data or flight simulation data, and selects error and accuracy as evaluation indexes.

3. A helicopter ground proximity warning method based on neural network prediction escape trajectory according to claim 2, characterized by that, the error and accuracy indexes are respectively:

Traj_E＝ABS(Traj_Model-Traj_True)，

Traj_Acc＝(1-Traj_E/Traj_True)*100％，

wherein, traj _ E represents the absolute value error between the escape trajectory Traj _ Model predicted based on the neural network Model and the escape trajectory Traj _ True obtained based on the real flight test data of the helicopter, and Traj _ Acc represents the relative precision between the Traj _ Model and the Traj _ True.

4. A helicopter ground proximity warning method based on neural network prediction escape trajectory as claimed in claim 3 wherein said vertical safety threshold should not be a fixed value, it needs to take into account the trajectory prediction uncertainty, navigation positioning uncertainty and terrain database vertical uncertainty, and its calculation formula is:

SCAN_Vert＝NAV+(DEM+TPA)/2

the NAV is the navigation and positioning uncertainty, the DEM is the vertical uncertainty of a terrain database, and the TPA is the track prediction uncertainty.

5. The helicopter ground proximity warning method based on escape trajectory prediction of neural network of claim 4, characterized in that said step 3 specifically is: and sequentially comparing the escape tracks with the terrain envelope lines below the escape tracks, if the comparison results of the escape tracks and the terrain envelope lines in different departure directions do not meet the requirements, making a decision to judge that the potential ground collision risk exists in the helicopter, otherwise, continuing to execute the current flight task.

6. The helicopter ground proximity warning method based on neural network prediction escape trajectory of claim 5, characterized by that, when the escape trajectory intersects with the terrain envelope, the direction corresponding to the escape trajectory is no longer considered as an effective terrain evasion selection, and the comparison result is determined to be not satisfactory.

7. The helicopter ground proximity warning method based on escape trajectory prediction of neural network of claim 6, characterized in that said step 4 specifically is: and (4) carrying out alarm prompt based on the decision-making judgment obtained in the step (3), wherein the alarm prompt comprises light alarm, voice alarm and display alarm.

8. The helicopter ground proximity warning system based on neural network prediction escape trajectory according to any one of claims 1, 5, 6, and 7, comprising an escape trajectory prediction module, a collision threat prediction module, a decision-making modification module and a warning prompt module, wherein the escape trajectory prediction module substitutes a neural network model according to the current parameters and escape modification mode of a helicopter to obtain a helicopter escape trajectory and inputs the helicopter escape trajectory into the decision-making modification module, and the collision threat prediction module scans the terrain below the helicopter escape trajectory to obtain a terrain elevation profile, generates a terrain envelope on the basis of the terrain elevation profile, and inputs the terrain envelope into the decision-making modification module; the extraction decision module compares whether the escape track intersects with the corresponding terrain envelope line in real time to carry out extraction decision judgment on potential ground collision threats, and prompts the extraction decision judgment result through the alarm prompt module.

Images