Flight behavior identification method based on aircraft intention in autonomous operation mode

Abstract

The invention discloses a flight behavior identification method based on an aircraft intention in an autonomous operation mode. Firstly, aircraft behavior characteristics are extracted from the aspects of time and space; secondly, constructing a flight intention model of the aircraft in the typical flight scene; and thirdly, carrying out clustering analysis on the behavior characteristics of the aircraft based on a BIRCH clustering model, dividing typical behaviors when the aircraft flies, and finally, calculating a real-time behavior characteristic value based on a flight intention model, and realizing online behavior identification of aircraft flight. According to the method, the behavior characteristics of the aircraft are constructed from the time-space perspective, so that the behavior of the aircraft is more concrete, secondly, the flight intention model of the aircraft can adapt to a dynamic change scene and is more flexible, and finally, the BIRCH model can adapt to the incremental clustering process and meets the real-time performance of online behavior recognition.

Description

Technical Field

[0001] The invention relates to a flight behavior recognition method based on aircraft intention in an autonomous operation mode, belonging to the field of behavior recognition. Background Art

[0002] In autonomous operation mode, aircraft flight behavior identification is one of the key technologies to ensure flight safety and efficiency. There are large differences in aircraft flight behavior at different stages. In the cruising stage, aircraft usually fly in a straight line at a constant speed. In the approach stage, there are more turns, large speed changes, and a surge in operational complexity. Autonomous operation increases the freedom of flight, and the risk also increases accordingly. Therefore, flight behavior identification can provide guidance for aircraft safety management and control, and its behavior prediction can lay the foundation for trajectory prediction, thereby providing safety guidance for the safe flight of autonomous aircraft.

[0003] At present, aircraft behavior feature recognition is usually oriented to terminal flight scenarios, and there is rarely any full-range flight behavior recognition, which is not consistent. In addition, most behavior recognition is based on the track similarity calculated through spatial distortion to obtain behavior similarity, ignoring the impact of the time characteristics of aircraft flight on behavior. Moreover, the mainstream behavior recognition of aircraft in the civil aviation field is offline calculation, which cannot meet the real-time requirements of autonomous operation mode. Therefore, it is urgent to propose a behavior recognition method that integrates the behavior characteristics of spatiotemporal characteristics, which can adapt to the dynamically changing airspace environment, is suitable for full-range flight, and ensures the timeliness of data updates. Summary of the invention

[0004] The present invention proposes a flight behavior recognition method based on aircraft intention in autonomous operation mode. This method can dynamically update the aircraft flight intention and identify the current aircraft flight behavior type. This method can be used as a pre-algorithm for track prediction to improve prediction accuracy and provide a reference for flight guidance in autonomous operation mode. The steps are as follows:

[0005] Step 1: Construct an aircraft flight behavior characteristic system from the temporal and spatial levels, and process the ADS-B data of the airport OD pair.

[0006] Step 2: Construct a variable flight intention model of the aircraft in normal and abnormal scenarios to obtain the next key intention point.

[0007] Step 3: Use the BIRCH clustering algorithm to cluster the flight behaviors based on the aircraft's spatiotemporal characteristics and classify the typical flight behaviors of aircraft.

[0008] Step 4: Calculate the online flight behavior feature value of the aircraft based on the aircraft flight intention model, and use the adaptive BIRCH clustering algorithm to achieve real-time identification of aircraft behavior.

[0009] Step 1 specifically includes:

[0010] Step 1.1: Construct an aircraft flight behavior feature system from the temporal and spatial levels, and select turning angle, current speed, and yaw distance as the aircraft flight temporal and spatial features. The turning angle and yaw distance are defined as follows:

[0011] ① Turning angle: the spatial angle formed by an aircraft flying through two consecutive flight segments.

[0012]

[0013] Where: —Angle between consecutive flight segments;

[0014] —Previous segment position vector;

[0015] —Position vector of the next flight segment.

[0016] ② Deviation distance: the distance at which the aircraft can quickly enter the planned route.

[0017]

[0018] Where: d o is the projection distance of the aircraft’s intercept segment, α is the angle between the aircraft’s heading and the intercept segment, and p o represents aircraft, p M Indicates the intersection of the extended aircraft heading line and the intercepted flight segment;

[0019] Step 1.2: Obtain flight plan information, perform equal-interval processing on the ADS-B timestamps of the airport OD pairs, and use Kalman filtering to smooth the track point features.

[0020] Step 1.3: In order to eliminate the impact of different dimensions between features, Z-score standardized features are used to speed up the convergence speed. The formula is as follows:

[0021]

[0022] Among them, x norm is the feature normalization value, x int is the initial value of the feature, x is the feature mean, σ is the feature standard deviation x int .

[0023] Step 2 specifically includes:

[0024] Step 2.1: Use the binary tree search algorithm to obtain the waypoint W2 closest to the current aircraft and the two waypoints before and after it, W1 and W3.

[0025] Step 2.2: Establish the aircraft flight intention model in the conventional flight scenario, and obtain the next key point of the aircraft through the angle θ. First, determine the inside and outside of the flight segment WP1-WP2-WP3, and calculate the cross product of the two groups of vectors to determine the inside and outside of the current aircraft. The calculation is as follows:

[0026]

[0027] Step 2.3: Determine the flight segment and the intended point. When the aircraft is outside the flight segment, the next key point is calculated as follows:

[0028]

[0029] When the aircraft is inside the segment, the next key point is calculated as follows:

[0030]

[0031] Among them, L1 is the vector formed by the aircraft and the nearest waypoint, L2 is the vector formed by the aircraft and the next waypoint of the nearest waypoint, τ1 and τ2 are the constraint angles for determining the next key point;

[0032] Step 2.4: Establish an aircraft flight intention model under abnormal flight scenarios, establish a new flight plan based on the trajectory planning triggered by the abnormal movement, and update the next key point as the trajectory planning point.

[0033] Step 3 specifically includes:

[0034] Step 3.1: Obtain n m-dimensional ADS-B flight data samples X = {X1, X2, …, X n}, construct its clustering feature tree.

[0035] Step 3.2: On the cluster feature tree, construct the cluster feature CF triple (N, LS, SS)

[0036]

[0037] Where: X i is the feature vector of the i-th sample in CF, x im is the mth feature of the ith sample, and N represents the sample size.

[0038] Step 3.3: There are no nodes on the initial clustering feature tree. Read the aircraft samples one by one and add them to the CF triples.

[0039] Step 3.4: When the maximum capacity limit of the leaf node is exceeded or the branching factor is exceeded, the node combination and splitting operations will be performed.

[0040] Step 3.5: After all samples are read, the cluster feature tree completes the preliminary clustering. Then the K-means algorithm is used to perform further secondary clustering on the triplets in these preliminary clusters.

[0041] Step 3.6: Use the silhouette coefficient to determine the optimal number of clusters, calculated as follows:

[0042]

[0043] Where: a(i) is the mean distance between sample i and other sample points in the cluster. b(i) is the mean distance between sample i and each sample point in the nearest cluster.

[0044] Step 4 specifically includes:

[0045] Step 4.1: Add new flight data points in real time, obtain the next key point based on the aircraft flight intention model, use this point as the last point of the turning angle, and calculate the aircraft's online flight behavior characteristic value.

[0046] Step 4.2: Input the calculated feature values ​​into the clustering feature tree for dynamic adjustment, adaptive BIRCH clustering algorithm, further train the model, realize incremental clustering, and achieve real-time recognition of aircraft behavior.

[0047] The beneficial effects achieved by the present invention are as follows: establishing aircraft flight behavior characteristics from both time and space levels to more comprehensively summarize the time and space information of aircraft flight; establishing a variable intention model of aircraft flight behavior under different scenarios to adapt to the dynamically changing airspace environment and ensure the effectiveness of recognition; using a BIRCH clustering algorithm to comprehensively divide aircraft behavior characteristics to obtain typical flight behaviors, combining the flight intention calculation online feature adaptive incremental clustering mode to achieve the update of dynamic behavior recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 is a flow chart of the method of the present invention;

[0049] Figure 2 This is a schematic diagram for calculating the turning angle;

[0050] Figure 3 This is a schematic diagram for calculating the yaw distance;

[0051] Figure 4 Searching schematics for aircraft segment matching;

[0052] Figure 5It is a diagram for judging the inside and outside of an aircraft segment, in which the upper left shows that the aircraft is outside the segment and has not flown over the nearest waypoint, the upper right shows that the aircraft is outside the segment and has flown over the nearest waypoint, the lower left shows that the aircraft is inside the segment and has not flown over the nearest waypoint, and the lower right shows that the aircraft is inside the segment and has flown over the nearest waypoint;

[0053] Figure 6 Capture visualizations to capture key points for example;

[0054] Figure 7 Schematic diagram of intention change in abnormal flight scenario

[0055] Figure 8 This is a schematic diagram of the BIRCH clustering feature tree;

[0056] Fig. 9 Schematic diagram of BIRCH leaf node splitting and combination;

[0057] Fig.10 is the result of silhouette coefficient calculation;

[0058] Fig.11 The typical flight behavior classification based on BIRCH, where (a) is right turn and deceleration, (b) is constant speed and straight flight, (c) is left turn and constant speed, (d) is deceleration and straight flight, (e) is right turn and acceleration, (f) is right turn and constant speed, (g) is acceleration and straight flight, (h) is left turn and acceleration, and (i) is left turn and deceleration; Specific embodiments

[0059] In order to make the purpose, technical solution and key points of the present invention clearer, the OD pair ADS-B data from Chengdu Shuangliu Airport ZUUU to Hangzhou Xiaoshan Airport ZSHC is selected for flight behavior recognition. The present invention is further described below in conjunction with the accompanying drawings.

[0060] The present invention proposes a flight behavior recognition method based on aircraft intention in autonomous operation mode, such as Figure 1 The specific implementation steps are as follows:

[0061] Step 1: Construct an aircraft flight behavior characteristic system from the temporal and spatial levels, and process the ADS-B data of the airport OD pair.

[0062] Step 1.1: Select the turning angle, current speed, and yaw distance as the aircraft flight time and space characteristics, where the turning angle and yaw distance are calculated as follows:

[0063] ① Turning angle: It is the spatial angle formed by the aircraft flying through two consecutive segments, such as Figure 2 .

[0064]

[0065] Where: —Angle between consecutive flight segments;

[0066] —Previous segment position vector;

[0067] —Position vector of the next flight segment.

[0068] ② Deviation distance: The distance at which the aircraft can quickly cut into the planned route, such as Figure 3 .

[0069]

[0070] Where: d o is the projection distance of the aircraft’s intercept segment, α is the angle between the aircraft’s heading and the intercept segment, and p o represents aircraft, p M Indicates the intersection of the extended aircraft heading line and the intercepted flight segment;

[0071] Step 1.2: Obtain the flight plan information from ZUUU to ZSHC, and process the ADS-B timestamps of the OD pairs of the two airports at equal intervals of 15 seconds. Use Kalman filtering to smooth the longitude, latitude, altitude, speed, and heading characteristics of the track points to improve the accuracy of behavior recognition.

[0072] Step 1.3: In order to eliminate the impact of different dimensions between features, Z-score standardized features are used to speed up the convergence speed. The formula is as follows:

[0073]

[0074] Among them, x norm is the feature normalization value, x int is the initial value of the feature, is the characteristic mean, σ is the characteristic standard deviation x int .

[0075] Step 2: Construct a variable flight intention model of the aircraft in normal and abnormal scenarios to obtain the next key intention point.

[0076] Step 2.1: Use the binary tree search algorithm to obtain the waypoint W2 closest to the current aircraft and the two waypoints before and after it, W1 and W3, as follows: Figure 4 As shown, within the blue area, the aircraft is shown in the flight segment WP1-WP2-WP3. Within this range, the aircraft's closest point is always WP2.

[0077] Step 2.2: Establish the aircraft flight intention model in the conventional flight scenario, and obtain the next key point of the aircraft through the angle θ. Determine the inner and outer sides of the current aircraft by calculating the cross product of two sets of vectors, as follows:

[0078]

[0079] Step 2.3: Determine the current flight segment and destination point of the aircraft, such as Figure 5 When the aircraft is outside the flight segment, the selected division angle is 90°, and the next key point is calculated as follows:

[0080]

[0081] When the aircraft is inside the flight segment, the division angle is selected as 45° and the next key point calculation is as follows:

[0082]

[0083] Wherein, L1 is the vector formed by the aircraft and the nearest waypoint, and L2 is the vector formed by the aircraft and the next waypoint of the nearest waypoint.

[0084] Implementation example Figure 6 shown.

[0085] Step 2.4: When entering an abnormal flight scenario, a new flight plan is established based on the trajectory planning triggered by the abnormal movement, and the next key point is updated as the trajectory planning point, such as Figure 7 .

[0086] Step 3: Use the BIRCH clustering algorithm to cluster the flight behaviors based on the aircraft's spatiotemporal characteristics and classify the typical flight behaviors of aircraft.

[0087] Step 3.1: There are 60286 3D behavior feature data samples X = {X1, X2, X3}, and its clustering feature tree CFTree is constructed, as follows: Figure 8 .

[0088] Step 3.2: Construct clustering feature CF triples (N, LS, SS) on CF Tree

[0089]

[0090] Where: X i Represents the 3D feature vector of each flight feature sample in CF, x im represents the mth feature of the i-th sample, and N represents the number of samples in CF.

[0091] Step 3.3: There is no node on the initial CF Tree. Read the feature samples one by one and add them to the CF triples.

[0092] Step 3.4: When the maximum number of leaf nodes L or the branching factor B is exceeded, the node combination is split, such as Fig. 9 shown.

[0093] Step 3.5: After reading the samples, CF Tree completes the first clustering of the samples to form several CFs, and then uses Kmeans to perform secondary clustering on the CF triplets of the leaf nodes.

[0094] Step 3.6: The result of the silhouette coefficient calculation is as follows Fig.10 As shown in the figure, the optimal number of clusters is determined to be 9, and 9 types of typical flight behaviors are obtained, such as Fig.11 shown.

[0095] Step 4: Calculate the online flight behavior feature value of the aircraft based on the aircraft flight intention model, and use the adaptive BIRCH clustering algorithm to achieve real-time identification of aircraft behavior.

[0096] Step 4.1: Add new flight feature data points in real time, obtain the next key point based on the aircraft flight intention model, use this point as the last point of the turning angle, and calculate the aircraft's online flight behavior feature value.

[0097] Step 4.2: Through the dynamic adjustment of the clustering feature tree, the adaptive BIRCH clustering algorithm is used to further train the model, realize incremental clustering, and achieve real-time recognition of aircraft behavior, as shown in Table 1.

[0098] Table 1

[0099]

[0100] The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are included in the scope of the claims of the present invention to be approved.

Claims

1. A method for identifying flight behavior based on aircraft intention in autonomous operation mode, characterized in that: The following steps are involved: Step 1: Construct an aircraft flight behavior feature system from the temporal and spatial levels, and process the ADS-B data of the airport OD pair; Step 2: Construct a variable flight intention model of the aircraft in normal and abnormal scenarios to obtain the next key intention point; Step 3: Use the BIRCH clustering algorithm to cluster the flight behaviors based on the aircraft's spatiotemporal characteristics and classify the typical flight behaviors of aircraft; Step 4: Calculate the aircraft's online flight behavior feature values ​​based on the aircraft's variable flight intention model and use the adaptive BIRCH clustering algorithm to achieve real-time identification of aircraft behavior.

2. The method for identifying flight behavior based on aircraft intention in an autonomous operation mode according to claim 1, characterized in that: Step 1 specifically includes: Step 1.1: Construct an aircraft flight behavior feature system from the temporal and spatial levels, and select turning angle, current speed, and yaw distance as the aircraft flight temporal and spatial features. The turning angle and yaw distance are defined as follows: ① Turning angle: the spatial angle formed by an aircraft flying through two consecutive flight segments; Where: —Angle between consecutive flight segments; —Previous segment position vector; - the position vector of the next flight segment; ② Deviation distance: the distance at which the aircraft can quickly cut into the planned route; Where: d o is the projection distance of the aircraft’s entry segment, α is the angle between the aircraft’s heading and the entry segment, and p o represents aircraft, p M Indicates the intersection of the extended aircraft heading line and the intercepted flight segment; Step 1.2: Obtain flight plan information, perform equal-interval processing on the ADS-B timestamps of the airport OD pairs, and use Kalman filtering to smooth the track point features; Step 1.3: In order to eliminate the impact of different dimensions between features, Z-score standardized features are used to speed up the convergence speed. The formula is as follows: Among them, x norm is the feature normalization value, x int is the initial value of the feature, is the characteristic mean, σ is the characteristic standard deviation x int .

3. The method for identifying flight behavior based on aircraft intention in autonomous operation mode according to claim 2, characterized in that: Step 2 specifically includes: Step 2.1: Use the binary tree search algorithm to obtain the waypoint W2 closest to the current aircraft and the two waypoints before and after it, W1 and W3; Step 2.2: Establish the aircraft flight intention model in the conventional flight scenario, and obtain the next key point of the aircraft through the angle θ; first determine the inside and outside of the flight segment W1-W2-W3, and calculate the cross product of the two groups of vectors to determine the inside and outside of the current aircraft, as follows: Step 2.3: Determine the flight segment and the intended point; when the aircraft is outside the flight segment, the next key point is calculated as follows: When the aircraft is inside the segment, the next key point is calculated as follows: Among them, L1 is the vector formed by the aircraft and the nearest waypoint, L2 is the vector formed by the aircraft and the next waypoint of the nearest waypoint, τ1 and τ2 are the constraint angles for determining the next key point; Step 2.4: Establish an aircraft flight intention model under abnormal flight scenarios, establish a new flight plan based on the trajectory planning triggered by the abnormal movement, and update the next key point as the trajectory planning point.

4. The method for identifying flight behavior based on aircraft intention in autonomous operation mode according to claim 3, characterized in that: Step 3 specifically includes: Step 3.1: Obtain n m-dimensional ADS-B flight data samples X = {X1, X2, …, X n }, construct its clustering feature tree; Step 3.2: On the cluster feature tree, construct the cluster feature CF triple (N, LS, SS) Where: X i is the feature vector of the i-th sample in CF, x im is the mth feature of the i-th sample, and N represents the sample size; Step 3.3: There is no node on the initial clustering feature tree. Read the aircraft samples one by one and add them to the CF triples. Step 3.4: When the maximum capacity limit of the leaf node is exceeded or the branching factor is exceeded, the node combination and splitting operations will be performed; Step 3.5: After all samples are read, the cluster feature tree completes the preliminary clustering, and then the K-means algorithm is used to perform further secondary clustering on the triplets in these preliminary clusters; Step 3.6: Use the silhouette coefficient to determine the optimal number of clusters, calculated as follows: Where a(i) is the mean distance between sample i and other sample points in the cluster, and b(i) is the mean distance between sample i and each sample point in the nearest cluster.

5. The method for identifying flight behavior based on aircraft intention according to claim 1, characterized in that: Step 4 specifically includes: Step 4.1: Add new flight data points in real time, obtain the next key point based on the aircraft flight intention model, use this point as the last point of the turning angle, and calculate the aircraft's online flight behavior characteristic value; Step 4.2: Input the calculated feature values ​​into the clustering feature tree for dynamic adjustment, adaptive BIRCH clustering algorithm, further train the model, realize incremental clustering, and achieve real-time recognition of aircraft behavior.

Images