Multi-UAV Real-Time Tracking Method, Device, Electronic Device and Medium

Abstract

The present invention provides a multi-UAV real-time tracking method, device, electronic device and medium. The method includes: receiving UAV signals at the current moment, and using a pre-trained UAV identity discrimination model to identify the UAV signals at the current moment to obtain a UAV signal identification result; for the identified UAV signals, clustering the arrival times of the UAV signals received by all TDOA stations at the current moment to obtain the positioning result of each identified UAV at the current moment; obtaining the flight trajectory of each UAV according to the UAV signal identification result and the continuous positioning result of the TDOA station; correcting the UAV signal identification result and the positioning result at the previous moment according to the flight trajectory of each UAV; re-training the UAV identity discrimination model according to the corrected UAV signal identification result and updating the flight trajectory of each UAV according to the corrected positioning result.

Description

Technical Field

[0001] The present invention relates to the field of communication technologies, and in particular, to a method, device, electronic device and medium for real-time tracking of multiple unmanned aerial vehicles (UAVs). Background Art

[0002] The real-time tracking of low-altitude UAVs in cities is an important issue for maintaining public safety, urban order and the safe operation of important facilities. At present, there are many technical problems in the simultaneous tracking of multiple UAVs. The most important one is how to match numerous radio signals with positioning results and flight trajectories.

[0003] For cooperative UAVs, that is, UAVs actively send their own Radio Identity (RID) information, it is relatively easy to bind the positioning result with the identity. However, for non-cooperative UAVs, the detection and positioning system needs to automatically distinguish the attribution of electromagnetic signals. However, the attribution of electromagnetic signals is severely affected by the urban environment and electromagnetic background noise, and cannot be determined very accurately. At the same time, the positioning result is also not very accurate due to factors such as multi-path propagation and noise interference of UAV signals in the urban environment, and even occasionally there will be a large deviation. Therefore, it causes difficulties in real-time tracking of multiple non-cooperative UAVs, and there will be inconsistencies between signal analysis and positioning results and the actual multi-UAV flight trajectories. Summary of the Invention

[0004] The present invention provides a method, device, electronic device and medium for real-time tracking of multiple UAVs to solve the problem that the signal recognition result and positioning result of UAVs are inconsistent with the actual flight trajectories of UAVs.

[0005] According to a first aspect of the present invention, there is provided a method for real-time tracking of multiple UAVs, including:

[0006] Receiving UAV signals at the current moment, and using a pre-trained UAV identity discrimination model to identify the UAV signals at the current moment to obtain a UAV signal recognition result;

[0007] For the identified UAV signals, clustering the arrival times of the UAV signals received by all Time Difference of Arrival (TDOA) stations at the current moment to obtain the positioning result of each identified UAV at the current moment;

[0008] Obtaining the flight trajectory of each UAV according to the UAV signal recognition result and the continuous positioning results of the TDOA stations;

[0009] According to the flight trajectory of each UAV, correcting the UAV signal recognition result and positioning result at the previous moment;

[0010] Retrain the UAV identity discrimination model according to the corrected UAV signal recognition result and update the flight trajectory of each UAV according to the corrected positioning result.

[0011] Optionally, receive the UAV signal at the current moment, and use the pre-trained UAV identity discrimination model to recognize the UAV signal at the current moment to obtain the UAV signal recognition result, specifically including:

[0012] Receive the UAV signal at the current moment; wherein the UAV signal at the current moment is represented in the form of IQ data;

[0013] Cluster the UAV signal at the current moment through a clustering algorithm to obtain multiple signal clusters;

[0014] Take the signals in each signal cluster as inputs respectively, and use the pre-trained UAV identity discrimination model to recognize the UAV signal to obtain the UAV signal recognition result.

[0015] Optionally, clustering the UAV signal at the current moment through a clustering algorithm to obtain multiple signal clusters, including:

[0016] Convert the UAV signal at the current moment from the IQ data form to the constellation diagram form; according to the similarity of the constellation diagram, cluster the UAV signal at the current moment through a clustering algorithm.

[0017] Optionally, taking the signals in each signal cluster as inputs respectively, and using the pre-trained UAV identity discrimination model to recognize the UAV signal, specifically including:

[0018] Select the signals with the top 30% signal-to-noise ratio in each signal cluster as inputs.

[0019] Optionally, for the identified UAV signals, cluster the arrival times of the UAV signals received by all TDOA sites at the current moment to obtain the positioning result of each identified UAV at the current moment, and further include: if the number of positioning results is greater than 2, then take the center of the multiple positioning results as the current positioning result.

[0020] Optionally, obtain the flight trajectory of each UAV according to the UAV signal recognition result and the continuous positioning result of the TDOA site, specifically including:

[0021] According to all the positioning results, initially determine the UAV signals that can form a continuous and smooth trajectory as coming from the same UAV, and filter out the first positioning results to be merged according to the UAV signal recognition result;

[0022] According to the UAV signal recognition result, generate flight trajectories for the positioning results from the same UAV respectively, and screen out the positioning results to be merged secondly according to whether the positioning results deviate significantly from the main flight trajectory;

[0023] Merge the positioning results to be merged firstly and the positioning results to be merged secondly according to the signal recognition result and / or the trajectory feature.

[0024] Optionally, according to all the positioning results, initially determine that the UAV signals that can form a continuous and smooth trajectory are from the same UAV, and screen out the positioning results to be merged firstly according to the UAV signal recognition result, specifically including:

[0025] According to the UAV signal recognition result in the trajectory, judge whether the proportion of the positioning results with inconsistent signal recognition results is less than 10%; the UAV signals corresponding to the positioning results with inconsistent signal recognition results are not from the same UAV as the UAV signals corresponding to other positioning results in this trajectory;

[0026] If so, determine that the positioning results with inconsistent signal recognition results belong to signal recognition abnormal data, and correct the signal recognition result according to the signal recognition abnormal data;

[0027] If not, determine the positioning results with inconsistent signal recognition results as the positioning results to be merged firstly.

[0028] Optionally, according to the UAV signal recognition result, generate flight trajectories for the positioning results from the same UAV respectively, and screen out the positioning results to be merged secondly according to whether the positioning results deviate significantly from the main flight trajectory, including:

[0029] Judge whether the proportion of the number of positioning results that deviate significantly from the main flight trajectory is less than 10%;

[0030] If so, delete the positioning results that deviate from the main flight trajectory;

[0031] If not, determine the positioning results that deviate from the main flight trajectory as the positioning results to be merged secondly.

[0032] According to the second aspect of the present invention, there is provided a multi-UAV real-time tracking device, including:

[0033] A signal recognition module, configured to receive UAV signals at the current moment, and use a pre-trained UAV identity discrimination model to identify the UAV signals, so as to obtain UAV signal recognition results;

[0034] A positioning result determination module, which is used to cluster the arrival times of the UAV signals received by all TDOA stations at the current moment for the identified UAV signals, so as to obtain the positioning results of each identified UAV at the current moment;

[0035] A flight trajectory determination module, which is used to obtain the flight trajectory of each UAV according to the UAV signal recognition result and the continuous positioning results of the TDOA stations;

[0036] A signal recognition result and positioning result correction module, which is used to correct the UAV signal recognition result and positioning result at the previous moment according to the flight trajectory of each UAV;

[0037] A trajectory update module re-trains the UAV identity discrimination model according to the corrected UAV signal recognition result and updates the flight trajectory of each UAV according to the corrected positioning result.

[0038] According to the third aspect of the present invention, there is provided an electronic device, including a processor and a memory; the memory stores a program executable by the processor; wherein, when the processor executes the program, it implements the multi-UAV real-time tracking method provided in the first aspect of the present invention.

[0039] According to the fourth aspect of the present invention, there is provided a machine-readable storage medium, on which a program is stored, and when the program is executed by a processor, it implements the multi-UAV real-time tracking method provided in the first aspect of the present invention.

[0040] The multi-UAV real-time tracking method provided by the present invention performs real-time tracking of multiple low-altitude UAVs in the city based on the method of identity discrimination and positioning analysis. First, the pre-trained UAV identity discrimination model is used to identify the UAV signals at the current moment, and the UAV signal recognition results are obtained. Then, clustering is performed on the arrival times of the UAV signals received by all TDOA stations at the current moment to obtain the positioning results of each identified UAV at the current moment and multiple consecutive moments thereafter; then, the flight trajectory of each UAV is obtained according to the identified UAV signals and the continuous positioning results of the TDOA stations, and the UAV signal recognition results and positioning results at the previous moment are corrected. Finally, the flight trajectory of each UAV is updated.

[0041] The present invention performs joint optimization according to the recognition result of the UAV signal and the real-time positioning result of the UAV signal, respectively obtains the flight trajectory of each UAV, and combines the analysis of historical positioning information and signal recognition according to the current positioning result or signal recognition result to mutually exclude abnormal results, realizes accurate positioning of multiple UAVs and accurate recognition of UAV signals, and further obtains a more accurate real-time tracking trajectory result of UAV flight, and solves the problem that the signal recognition result and positioning result of the UAV are inconsistent with the actual flight trajectory of the UAV. Brief Description of the Drawings

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0043] Figure 1 is a schematic flowchart of a multi-UAV real-time tracking method provided in an exemplary embodiment of the present invention;

[0044] Figure 2 is a schematic flowchart of a UAV signal recognition method provided in an exemplary embodiment of the present invention;

[0045] Figure 3 is a schematic flowchart of a UAV signal recognition method provided in another exemplary embodiment of the present invention;

[0046] Figure 4 is a schematic flowchart of a UAV flight trajectory generation method provided in an exemplary embodiment of the present invention;

[0047] Figure 5 is a schematic flowchart of a UAV flight trajectory generation method provided in another exemplary embodiment of the present invention;

[0048] Figure 6 is a schematic flowchart of a UAV flight trajectory generation method provided in yet another exemplary embodiment of the present invention;

[0049] Figure 7 is a schematic block diagram of a multi-UAV real-time tracking device provided in an exemplary embodiment of the present invention;

[0050] Figure 8 is a schematic diagram of the structure of an electronic device provided in an exemplary embodiment of the present invention. Detailed Embodiments

[0051] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0052] In the specification, claims and above-mentioned drawings of the present invention, the terms "first", "second", "third", "fourth", etc. (if any) are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0053] The technical solution of the present invention will be described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments.

[0054] Please refer to Figure 1 , a multi-UAV real-time tracking method provided in an exemplary embodiment of the present invention includes:

[0055] S1: Receive the UAV signal at the current moment, and use the pre-trained UAV identity discrimination model to identify the UAV signal at the current moment to obtain a UAV signal identification result;

[0056] S2: For the identified UAV signal, cluster the arrival times of the UAV signals received by all TDOA stations at the current moment to obtain the positioning result of each identified UAV at the current moment;

[0057] S3: Obtain the flight trajectory of each UAV according to the identified UAV signal and the continuous positioning results of the TDOA stations;

[0058] S4: According to the flight trajectory of each UAV, correct the UAV signal identification result and positioning result at the previous moment;

[0059] S5: Retrain the UAV identity discrimination model according to the corrected UAV signal identification result and update the flight trajectory of each UAV according to the corrected positioning result.

[0060] The multi-UAV real-time tracking method provided by the present invention performs real-time tracking on multiple UAVs based on the method of identity discrimination and positioning analysis. Steps S1 to S5 are executed at each moment during the flight of the UAVs. After updating the current flight trajectory of the UAVs each time, the UAV identity discrimination model is retrained according to the corrected UAV signal recognition result, so as to obtain a more accurate signal recognition result when identifying the UAV signal at the next moment, and thus more accurately realize the real-time tracking of multiple UAVs.

[0061] Please refer to Figure 2 , where step S1 specifically includes:

[0062] S11: Receive the UAV signal at the current moment; the UAV signal at the current moment is represented in the form of IQ data; the IQ signal is the in-phase quadrature signal, I is in-phase (please supplement Chinese), Q is quadrature (please supplement Chinese), and the phase of Q is 90 degrees different from that of I. The UAV signal at the current moment is in the form of IQ data after quadrature modulation.

[0063] S12: Cluster the UAV signal at the current moment through a clustering algorithm to obtain multiple signal clusters.

[0064] Please refer to Figure 3 , where step S12 includes:

[0065] S121: Convert the UAV signal at the current moment from the form of IQ data to the form of a constellation diagram;

[0066] S122: Cluster the UAV signal at the current moment through a clustering algorithm according to the similarity of the constellation diagram to obtain multiple signal clusters.

[0067] The constellation diagram is a way to represent digital signals on the complex plane in the field of digital communication, so as to intuitively represent signals and the relationships between signals. Converting the UAV signal at the current moment from the form of IQ data to the form of a constellation diagram means presenting the vectors of in-phase (I) and quadrature (Q) in the quadrature amplitude modulation signal in the form of coordinates on a two-dimensional coordinate. For example, in a coordinate, the abscissa is the I axis and the ordinate is the Q axis. The projection of the vector signal on the I axis is the I component, and the projection on the Q axis is the Q component. In this way, any combination of the amplitude of any I and the amplitude of any Q will map a corresponding constellation point on the coordinate diagram. Suppose each constellation point represents a mapping composed of n-bit in-phase (I) signal data and n-bit quadrature (Q) signal data, and there are n×n possible combination states of the I component and the Q component mapped to the constellation diagram.

[0068] Cluster using a clustering algorithm based on the coordinate position of the UAV signal in the constellation diagram at the current moment. Clustering means dividing a data set into different classes or clusters according to a specific criterion (such as distance), so that the similarity of data objects within the same cluster is as large as possible, and at the same time, the difference between data objects not in the same cluster is also as large as possible. That is, after clustering, data of the same class is gathered together as much as possible, and data of different classes is separated as much as possible.

[0069] In one real-time method, the clustering algorithm used is the k-means algorithm, that is, the k-means algorithm. The k-means algorithm is a clustering algorithm based on partitioning, using distance as the criterion for measuring the similarity between data objects. That is, the smaller the distance between data objects, the higher their similarity, and the more likely they are to be in the same cluster. The k in the k-means algorithm represents the number of clusters, and means represents the mean of data objects within the cluster.

[0070] S13: Use the signals in each signal cluster as inputs respectively, and utilize a pre-trained UAV identity discrimination model to identify the UAV signals to obtain UAV signal identification results.

[0071] The UAV identity discrimination model adopts a deep learning algorithm, such as using deep learning algorithms such as convolutional neural networks or long short-term memory networks (LSTM). First, pre-train the identity discrimination model according to the signal feature database, and then input the current signal into the identity discrimination model to obtain the UAV signal identification result.

[0072] In one implementation, from each signal cluster, select the signals with the top 30% signal-to-noise ratio as inputs. Determine whether the signals in the cluster belong to a pre-trained UAV based on the relevant results obtained by the UAV identity discrimination model.

[0073] The signal identification result obtained through step S1, the identified UAV signal represents the identification of the UAV. If this signal has a strong correlation with the signal features in the pre-trained signal database, it can be determined that this signal comes from a known UAV.

[0074] In one implementation of the present invention, in step S2, for the identified UAV signals, cluster the arrival times of all TDOA stations at the current moment, and according to the clustering results, delete the UAV signal data that significantly deviates from the clustering center to obtain the UAV signal data with valid arrival times. The TDOA positioning calculation server calculates the relative arrival time difference according to the valid arrival time data of each TDOA station corresponding to the signal of a certain UAV to obtain the TDOA positioning result of the UAV corresponding to this UAV signal at this moment. In one implementation, if the number of the positioning results is greater than 2, then use the center of the multiple positioning results as the current positioning result.

[0075] Step S3 obtains the flight trajectory of each UAV according to the UAV signal recognition result and the continuous positioning results of the TDOA stations. The continuous positioning results of the TDOA stations refer to all the positioning results obtained by using the TDOA stations for multiple positionings within a period of time starting from the current moment. The positioning result at each moment can be obtained in the manner of Step S2. Please refer to Figure 4 , in one implementation, Step S3 specifically includes:

[0076] S31: According to all the positioning results, initially determine the UAV signals that can form a continuous and smooth trajectory as coming from the same UAV, and screen out the first positioning results to be merged according to the UAV signal recognition result.

[0077] Please refer to Figure 5 , in one implementation, Step S31 specifically includes:

[0078] S311: According to the UAV signal recognition result in the trajectory, judge whether the proportion of the positioning results with inconsistent signal recognition results is less than 10%; the UAV signals corresponding to the positioning results with inconsistent signal recognition results are not from the same UAV as the UAV signals corresponding to other positioning results in this trajectory;

[0079] S312: If so, determine that the positioning results with inconsistent signal recognition results belong to signal recognition abnormal data, and correct the signal recognition result according to the signal recognition abnormal data;

[0080] S313: If not, determine the positioning results with inconsistent signal recognition results as the first positioning results to be merged.

[0081] Step S31: Based on all the TDOA positioning results, the UAV signals that can form a continuous and smooth trajectory are initially determined to come from the same UAV. Among all the positioning results in the formed trajectory, the signal recognition results corresponding to the UAV signals may not all belong to the same UAV. Therefore, it is necessary to screen out the positioning results that need to be re-merged according to the signal recognition results. In one way, it is judged through step S11 the proportion of the positioning results with inconsistent signal recognition results: If the proportion is less than 10%, then it can be considered that these signal recognition results belong to the abnormal signal recognition data, and the positioning results corresponding to these abnormal data are still determined to be the same UAV, and the signal recognition results are corrected according to the abnormal signal recognition data; If it is higher than 10%, it is considered that these signals with inconsistent signal recognition results come from other UAVs. Therefore, the positioning results corresponding to these signals are determined to be the first positioning results to be re-merged, and the first positioning results to be re-merged are removed from this trajectory in this step and re-merged later for the synthesis of other trajectories.

[0082] S32: According to the UAV signal recognition results, the positioning results from the same UAV are respectively generated into flight trajectories, and the second positioning results to be re-merged are screened out according to whether the positioning results significantly deviate from the main flight trajectory.

[0083] Please refer to Figure 6 , in one implementation manner, step S32: specifically includes:

[0084] S321: Judge whether the proportion of the number of positioning results that significantly deviate from the main flight trajectory is less than 10%;

[0085] S322: If so, delete the positioning results that deviate from the main flight trajectory;

[0086] S323: If not, determine the positioning results that deviate from the main flight trajectory as the second positioning results to be re-merged.

[0087] In step S32, according to the UAV signal recognition result, the positioning results corresponding to the signals belonging to the same aircraft are initially determined to come from the same UAV, and the positioning results from the same UAV are respectively formed into trajectories. There will be some positioning points deviating from the main trajectory in the trajectories generated based on these positioning results. If the proportion of the positioning points significantly deviating from the main trajectory in all the positioning result data is less than 10%, it is determined that the positioning result is occasionally abnormal, and the positioning results deviating from the main flight trajectory are deleted; if the proportion is higher than 10%, it cannot be determined that the signals corresponding to these positioning results belong to this UAV. Then, the positioning results deviating from the main flight trajectory are determined as the second positioning results to be merged, and the second positioning results to be merged are removed from this trajectory in this step and re-merged later for use in other trajectory syntheses.

[0088] S33: Merge the first positioning results to be merged and the second positioning results to be merged according to the signal recognition result and / or the trajectory feature.

[0089] After obtaining the flight trajectories of each UAV through steps S31 and S32 and screening out the first positioning results to be merged and the second positioning results to be merged, step S33 will re-merge these positioning results to be merged. All the positioning results to be merged are merged according to the confidence level: if the signal recognition results corresponding to more than 90% of the positioning results to be merged are consistent, these positioning results are merged according to the signal recognition result; if more than 95% of the positioning results to be merged can generate continuous and smooth trajectories, they are merged according to the flight trajectory.

[0090] After obtaining the flight trajectories of each UAV through steps S1 - S3, step S4 corrects the UAV signal recognition results and positioning results at the previous moment according to the flight trajectories of each UAV. In one implementation, it specifically includes: according to the UAV signal recognition result, if the UAV signal recognition result corresponding to a certain positioning result at the previous moment is inconsistent with the UAV identity corresponding to the determined trajectory, this positioning result is re-assigned to the UAV trajectory that matches its signal recognition result. If the UAV signal recognition result corresponding to this positioning result is inconsistent with the UAV identities corresponding to all the trajectories, this positioning result is deleted; if a certain positioning result deviates from the corresponding trajectory by more than 50 meters, it is determined that this positioning result is abnormal, and this positioning result is deleted from the dataset for merging in the current trajectory.

[0091] In step S3, the flight trajectories of each drone are obtained by combining the recognition results of the drone signals and the real-time positioning results of the TDOA stations. During the process of trajectory merging, data results with incorrect signal recognition or abnormal positioning are re-grouped or deleted. In step S4, the historical data at previous times is checked according to the newly synthesized flight trajectories, and the drone signal recognition results and positioning results at previous times are corrected. On the one hand, more accurate drone flight trajectories are obtained, solving the problem that the signal recognition results and positioning results of the drones are inconsistent with the actual flight trajectories of the drones. On the other hand, after the flight trajectories of each drone are generated, the signal recognition results are updated using the data with abnormal signal recognition, so as to re-train the drone identity discrimination model according to the corrected signal recognition result data later, in order to obtain more accurate signal recognition results during the process of drone signal recognition at subsequent times.

[0092] In step S5, the drone identity discrimination model is re-trained according to the corrected drone signal recognition results, and the flight trajectory of each drone is updated according to the corrected positioning results. Re-training the drone identity discrimination model using the corrected drone signal recognition results improves the accuracy of the next signal recognition result. At the same time, updating the flight trajectory of each drone using the corrected positioning results makes the flight trajectory of each drone more accurate.

[0093] Please refer to Figure 7 , according to the second aspect of the present invention, there is provided a multi-drone real-time tracking device 100, which is characterized by including:

[0094] A signal recognition module 101, configured to receive the drone signals at the current moment, and use a pre-trained drone identity discrimination model to recognize the drone signals, so as to obtain drone signal recognition results;

[0095] A positioning result determination module 102, configured to cluster the arrival times of the drone signals received by all TDOA stations at the current moment for the recognized drone signals, so as to obtain the positioning results of each recognized drone at the current moment;

[0096] A flight trajectory determination module 103, configured to obtain the flight trajectories of each drone according to the drone signal recognition results and the continuous positioning results of the TDOA stations;

[0097] A signal recognition result and positioning result correction module 104, configured to correct the drone signal recognition results and positioning results at previous times according to the flight trajectories of each drone;

[0098] A trajectory update module 105, re-trains the drone identity discrimination model according to the corrected drone signal recognition results, and updates the flight trajectory of each drone according to the corrected positioning results.

[0099] Please refer to Figure 8 , which provides an electronic device 40, including:

[0100] A processor 41; and

[0101] A memory 42 for storing executable instructions of the processor;

[0102] Wherein, the processor 41 is configured to execute the methods involved above by executing the executable instructions.

[0103] The processor 41 can communicate with the memory 42 through a bus 43.

[0104] The embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the methods involved above are implemented.

[0105] Those of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions. The foregoing program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps including the above method embodiments; and the foregoing storage medium includes: various media such as ROM, RAM, magnetic disk or optical disc that can store program codes.

[0106] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A multi-UAV real-time tracking method, characterized in that, Including: Receiving a drone signal at the current moment, and using a pre-trained drone identity discrimination model to identify the drone signal at the current moment to obtain a drone signal identification result; For the identified drone signal, clustering the arrival times of the drone signals received by all TDOA stations at the current moment to obtain the positioning result of each identified drone at the current moment; Obtaining the flight trajectory of each drone according to the drone signal identification result and the continuous positioning result of the TDOA station, including: according to all the positioning results, initially determining the drone signals that can form a continuous and smooth trajectory as coming from the same drone, and screening out the first positioning results to be merged according to the drone signal identification result; according to the drone signal identification result, generating flight trajectories for the positioning results from the same drone respectively, and screening out the second positioning results to be merged according to whether the positioning results deviate significantly from the main flight trajectory; merging the first positioning results to be merged and the second positioning results to be merged according to the signal identification result and / or the characteristics of the trajectory; According to the flight trajectory of each drone, correcting the drone signal identification result and the positioning result at the previous moment; Retraining the drone identity discrimination model according to the corrected drone signal identification result and updating the flight trajectory of each drone according to the corrected positioning result.

2. The multi-UAV real-time tracking method according to claim 1, characterized in that, Receiving a drone signal at the current moment, and using a pre-trained drone identity discrimination model to identify the drone signal at the current moment to obtain a drone signal identification result, specifically including: Receiving a drone signal at the current moment; wherein the drone signal at the current moment is represented in the form of IQ data; Clustering the drone signals at the current moment through a clustering algorithm to obtain a plurality of signal clusters; Using each signal in each signal cluster as an input, and using a pre-trained drone identity discrimination model to identify the drone signal to obtain a drone signal identification result.

3. The multi-UAV real-time tracking method according to claim 2, wherein Clustering the drone signals at the current moment through a clustering algorithm to obtain a plurality of signal clusters, including: Converting the drone signals at the current moment from the IQ data form to a constellation diagram form; clustering the drone signals at the current moment through a clustering algorithm according to the similarity of the constellation diagram.

4. The multi-UAV real-time tracking method according to claim 2, wherein Using each signal in each signal cluster as an input, and using a pre-trained drone identity discrimination model to identify the drone signal, specifically including: Selecting the signals with the top 30% signal-to-noise ratio in each signal cluster as inputs.

5. The multi-UAV real-time tracking method according to claim 1, characterized in that After clustering the arrival times of the drone signals received by all TDOA stations at the current moment for the identified drone signal to obtain the positioning result of each identified drone at the current moment, it further includes: if the number of the positioning results is greater than 2, taking the center of the multiple positioning results as the current positioning result.

6. The multi-UAV real-time tracking method according to claim 1, wherein, Based on all the positioning results, the UAV signals that can form a continuous and smooth trajectory are initially determined to come from the same UAV, and according to the UAV signal recognition results, the first positioning results to be merged are screened out, specifically including: According to the UAV signal recognition results in the trajectory, judge whether the proportion of positioning results with inconsistent signal recognition results is less than 10%; the UAV signals corresponding to the positioning results with inconsistent signal recognition results and the UAV signals corresponding to other positioning results in this trajectory do not come from the same UAV; If so, determine that the positioning results with inconsistent signal recognition results belong to the signal recognition abnormal data, and correct the signal recognition results according to the signal recognition abnormal data; If not, determine the positioning results with inconsistent signal recognition results as the first positioning results to be merged.

7. The multi-UAV real-time tracking method according to claim 1 or 6, characterized in that According to the UAV signal recognition results, generate flight trajectories for the positioning results from the same UAV respectively, and according to whether the positioning results significantly deviate from the main flight trajectory, screen out the second positioning results to be merged, including: Judge whether the proportion of the number of positioning results significantly deviating from the main flight trajectory is less than 10%; If so, delete the positioning results deviating from the main flight trajectory; If not, determine the positioning results deviating from the main flight trajectory as the second positioning results to be merged.

8. A multi-UAV real-time tracking device, characterized in that, Including: A signal recognition module, configured to receive the UAV signal at the current moment, and use a pre-trained UAV identity discrimination model to recognize the UAV signal, so as to obtain the UAV signal recognition result; A positioning result determination module, for the recognized UAV signal, cluster the arrival times of the UAV signals received by all TDOA stations at the current moment to obtain the positioning result of each recognized UAV at the current moment; A flight trajectory determination module, obtaining the flight trajectory of each UAV according to the UAV signal recognition result and the continuous positioning results of the TDOA stations, including: based on all the positioning results, initially determining that the UAV signals that can form a continuous and smooth trajectory come from the same UAV, and screening out the first positioning results to be merged according to the UAV signal recognition results; generating flight trajectories for the positioning results from the same UAV respectively according to the UAV signal recognition results, and screening out the second positioning results to be merged according to whether the positioning results significantly deviate from the main flight trajectory; merging the first positioning results to be merged and the second positioning results to be merged according to the signal recognition results and / or the characteristics of the trajectory; A signal recognition result and positioning result correction module, configured to correct the UAV signal recognition results and positioning results at the previous moment according to the flight trajectory of each UAV; A trajectory update module, re-training the UAV identity discrimination model according to the corrected UAV signal recognition results and updating the flight trajectory of each UAV according to the corrected positioning results.

9. An electronic device, characterized in that, Including a processor and a memory; the memory stores a program executable by the processor; wherein, when the processor executes the program, the multi-UAV real-time tracking method described in any one of claims 1-7 is implemented.

10. A machine-readable storage medium, characterized in that, A program is stored thereon, and when the program is executed by a processor, the multi-UAV real-time tracking method described in any one of claims 1-7 is implemented.

Images