Unmanned aerial vehicle abnormal behavior recognition method based on low-altitude situation

Abstract

The application relates to the technical field of unmanned aerial vehicles and discloses an unmanned aerial vehicle abnormal behavior identification method based on low-altitude situation, which comprises the following steps: acquiring low-altitude multi-source situation data of a target region, wherein the low-altitude multi-source situation data comprises multi-band radar plot data, broadcast automatic dependent surveillance identity broadcast data and special network report approval records; determining radar physical tracks, identity tracks and planned report flight paths based on the multi-band radar plot data, the broadcast automatic dependent surveillance identity broadcast data and the special network report approval records; determining a first correlation confidence of the radar physical tracks and the identity tracks, a second correlation confidence of the radar physical tracks and the planned report flight paths and a third correlation confidence of the identity tracks and the planned report flight paths; determining a target cross-source matching state based on the first correlation confidence, the second correlation confidence and the third correlation confidence; and performing abnormal behavior identification on a target object based on the target cross-source matching state to obtain an abnormal behavior identification result.

Description

Technical Field

[0001] This application relates to the field of unmanned aerial vehicle (UAV) technology, specifically to a method for identifying abnormal UAV behavior based on low-altitude situational awareness. Background Technology

[0002] With the rapid development of the low-altitude economy, drones are increasingly used in civilian and commercial fields, but they also face problems such as disorderly flight and illegal operation, posing challenges to public safety and social management.

[0003] In related technologies, low-altitude airspace control mainly relies on manual patrols to identify drones. However, manual patrols have limited coverage and are prone to errors, resulting in low accuracy and inefficiency. Summary of the Invention

[0004] In view of this, this application provides a method for identifying abnormal behavior of UAVs based on low-altitude situational awareness, in order to solve the problems of limited coverage of manual patrols and the ease with which manual patrols may make mistakes, resulting in low accuracy and poor efficiency in identification.

[0005] Firstly, this application provides a method for identifying abnormal UAV behavior based on low-altitude situational awareness, comprising: acquiring low-altitude multi-source situational awareness data of a target area, the low-altitude multi-source situational awareness data including multi-band radar spot data, automatic dependent surveillance identity broadcast data, and private network reporting and approval records; determining, based on the multi-band radar spot data, automatic dependent surveillance identity broadcast data, and private network reporting and approval records, the radar physical track corresponding to the multi-band radar spot data, the identity track corresponding to the automatic dependent surveillance identity broadcast data, and the planned reporting route corresponding to the private network reporting and approval records; determining a first association confidence level between the radar physical track and the identity track, a second association confidence level between the radar physical track and the planned reporting route, and a third association confidence level between the identity track and the planned reporting route; determining the target cross-source matching state based on the first association confidence level, the second association confidence level, and the third association confidence level; and identifying abnormal behavior of the target based on the target cross-source matching state to obtain the abnormal behavior identification result.

[0006] The UAV abnormal behavior identification method based on low-altitude situational awareness provided in this embodiment collects multi-source low-altitude situational awareness data of the target area, including multi-band radar point data, automatic correlation surveillance identity broadcast data, and private network reporting and approval records. Based on these three types of data, radar physical tracks, identity tracks, and planned reporting routes are generated respectively. Then, the first correlation confidence level between the radar physical track and the identity track, the second correlation confidence level between the radar physical track and the planned reporting route, and the third correlation confidence level between the identity track and the planned reporting route are calculated sequentially. The cross-source matching status of the target is determined based on these three types of correlation confidence levels. Combined with the target's cross-source matching status, the method automatically completes the identification of abnormal target behavior. The entire process requires no manual inspection, eliminating the limitations of limited coverage of manual inspections. Furthermore, by quantifying the matching relationship based on the multi-dimensional correlation confidence levels of the three types of tracks, it replaces subjective human judgment, avoiding the errors that easily occur during manual inspections. This simultaneously improves the coverage, accuracy, and efficiency of UAV identification.

[0007] In one possible implementation, based on multi-band radar spot data, automatic dependent surveillance (ADS) identity broadcast data, and dedicated network reporting and approval records, the radar physical track corresponding to the multi-band radar spot data, the identity track corresponding to the ADS identity broadcast data, and the planned reporting route corresponding to the dedicated network reporting and approval records are determined. This includes: performing spatiotemporal registration processing on the multi-band radar spot data to obtain effective discrete point data; performing time-series stitching processing on the effective discrete point data to form the continuous target motion trajectory corresponding to the effective discrete point data; and performing calculation processing on the continuous motion trajectory to obtain a first target dataset. The first target dataset includes calculated position, flight speed, radar cross-section, and heading change. The system generates at least two of the following: rate, flight acceleration, and physical characteristics; based on the first target dataset, a radar physical track is generated; the automatic dependent surveillance broadcast (ADS) identity broadcast data is subjected to a first parsing process to obtain a second target dataset; wherein the second target dataset includes at least one of the following: identity number, time-series coordinates, motion parameters, declared aircraft type, message sequence number key identifier, and motion information; based on the second target dataset, an identity track is generated; the dedicated network reporting and approval records are subjected to a second parsing process to obtain a third target dataset; wherein the third target dataset includes at least one of the following: approved route, take-off and landing point, flight time window, permissible altitude range, and declared aircraft type constraints; based on the third target dataset, a planned reporting route is generated.

[0008] The UAV abnormal behavior recognition method based on low-altitude situational awareness provided in this embodiment obtains a first target dataset and generates radar physical tracks by performing spatiotemporal registration, time-series stitching, and trajectory calculation on multi-band radar point data. It obtains a second target dataset and generates identity tracks by performing a first parsing on automatic dependent surveillance identity broadcast data. It obtains a third target dataset and generates planned reporting routes by performing a second parsing on private network reporting and approval records. Based on the three types of low-altitude monitoring data, it automatically generates three types of standardized tracks. The entire process does not require manual inspection to collect and organize airspace target information. It overcomes the shortcomings of manual inspection, which has limited coverage and can only observe locally. Furthermore, it solves the problem of errors that are easily caused by relying on subjective visual records during manual inspection by uniformly extracting and calculating multi-dimensional objective data such as location, identity number, and approved routes through the data processing flow.

[0009] In one possible implementation, determining the first association confidence level between the radar physical track and the identity track, the second association confidence level between the radar physical track and the planned reported route, and the third association confidence level between the identity track and the planned reported route includes: determining the position residual, velocity residual, heading residual, and altitude residual between the radar physical track and the identity track based on the radar physical track and the identity track; determining the first association confidence level based on the position residual, velocity residual, heading residual, altitude residual, and target error scale; and determining the target based on the radar physical track and the planned reported route. The multi-dimensional deviation items corresponding to the object are determined; among them, the multi-dimensional deviation items include at least two of the following: lateral deviation distance, deviation between actual take-off and landing point and reported take-off and landing point, deviation between flight time period and approval time window, and matching degree between target aircraft type and declared aircraft type; the multi-dimensional deviation items are weighted and fused to determine the second association confidence level; based on the identity track and the planned reported route, the identity number matching relationship, flight time overlap and activity area overlap of the target object are determined; based on the identity number matching relationship, flight time overlap and activity area overlap, the third association confidence level is determined.

[0010] The UAV abnormal behavior recognition method based on low-altitude situational awareness provided in this embodiment calculates the first correlation confidence score by combining the position residual, velocity residual, heading residual, and altitude residual of the radar physical track and the identity track with the target error scale. Then, it obtains the second correlation confidence score by weighted fusion of multi-dimensional deviation terms between the radar physical track and the planned reported flight path. Finally, it calculates the third correlation confidence score based on the identity number matching relationship, flight time overlap, and activity area overlap between the identity track and the planned reported flight path. The entire process relies on multiple types of objective quantitative differences and matching indicators to complete numerical judgments, eliminating the need for manual inspections and subjective judgments based on visual inspection and experience. This solves the shortcomings of limited coverage in manual inspections that cannot simultaneously compare all dimensions of the three types of tracks, and also addresses the problem of errors easily arising from subjective judgments in manual inspections. In one possible implementation, the target cross-source matching status is determined based on a first association confidence level, a second association confidence level, and a third association confidence level, including: determining a first determination result based on the first association confidence level and a first association threshold; determining a second determination result based on the second association confidence level and the second association threshold; determining a third determination result based on the third association confidence level and the third association threshold; and determining the target cross-source matching status based on the first determination result, the second determination result, the third determination result, and auxiliary determination conditions.

[0011] The UAV abnormal behavior recognition method based on low-altitude situational awareness provided in this embodiment obtains a first judgment result and a second judgment result by comparing the first and second association confidence scores with their corresponding association thresholds, respectively. A third judgment result is obtained by comparing the third association confidence score with the third association threshold. The target cross-source matching status is then comprehensively determined by combining the three types of judgment results and auxiliary judgment conditions. By combining hierarchical threshold comparison with multi-dimensional results for joint analysis, the evaluation criteria for the matching relationship of radar physical tracks, identity tracks, and planned reporting routes are unified. This avoids the one-sidedness caused by single-dimensional comparison, fully integrates the matching quantitative information of the three types of tracks, distinguishes different track matching combinations, reduces the bias caused by single confidence score judgment, and improves the completeness and reliability of the target cross-source matching status classification.

[0012] In one possible implementation, the auxiliary determination conditions include whether the same identity track matches multiple radar tracks exceeding the spatial threshold and whether a single continuous radar track matches multiple sets of different identity tracks within a preset time period. Based on the first determination result, the second determination result, the third determination result, and the auxiliary determination conditions, the target cross-source matching status is determined, including: if the same identity track matches multiple radar tracks exceeding the spatial threshold, the target cross-source matching status is determined to be an identity duplication status; if a single continuous radar track matches multiple sets of different identity tracks within a preset time period, the target cross-source matching status is determined to be an identity mutation status; if the same identity track does not match multiple radar tracks exceeding the spatial threshold, and a single continuous radar track does not match multiple sets of different identity tracks within a preset time period... If the first determination result indicates that the radar physical track and the identity track match, and the third determination result indicates that the identity track matches the planned and reported route, the target's cross-source matching status is determined to be a stable and legal association status. If the same identity track does not match multiple radar tracks exceeding the spatial threshold, a single continuous radar track does not match multiple sets of different identity tracks within a preset time period, and the second determination result indicates that the radar physical track does not match the planned and reported route, the target's cross-source matching status is determined to be a reporting mismatch status. If the same identity track does not match multiple radar tracks exceeding the spatial threshold, a single continuous radar track does not match multiple sets of different identity tracks within a preset time period, and the first determination result indicates that the radar physical track and the identity track do not match, the target's cross-source matching status is determined to be an identity conflict status.

[0013] The UAV abnormal behavior recognition method based on low-altitude situational awareness provided in this embodiment distinguishes the target cross-source matching status by layering two auxiliary judgment conditions on the basis of the first judgment result, the second judgment result, and the third judgment result: whether the same identity track matches multiple radar tracks exceeding the spatial threshold, and whether a single continuous radar track matches multiple sets of tracks with different identities within a preset time. The method sets judgment logic for five scenarios: identity duplication, identity mutation, stable and legitimate association, reporting mismatch, and identity conflict. It can distinguish between track matching deviation and special abnormal situations such as identity forgery. It makes up for the shortcoming that it cannot identify multi-track cross-masking behavior by comparing only the confidence threshold of three types of association. It refines the granularity of the target cross-source matching status classification, so that various track coupling contradiction scenarios can be classified into states, thereby improving the detail and scenario adaptability of the target cross-source matching status classification.

[0014] In one possible implementation, based on the target's cross-source matching status, abnormal behavior identification of the target is performed to obtain the abnormal behavior identification result, including: if the target's cross-source matching status is an identity conflict status, continuously extracting the radar cross section, flight altitude, and flight speed time series data within the radar physical track; if the radar cross section within a preset duration is not greater than a preset cross section threshold, the flight altitude is not greater than a preset altitude threshold, and the flight speed is not greater than a first speed threshold, the abnormal behavior identification result is determined to be the result of the target flying illegally.

[0015] The UAV abnormal behavior identification method based on low-altitude situational awareness provided in this embodiment triggers a time-series feature extraction process based on the target cross-source matching state of identity conflict. It continuously extracts time-series data of radar cross section, flight altitude, and flight speed corresponding to radar physical tracks. Then, it combines preset duration constraints to synchronously verify three quantitative indicators: radar cross section threshold, flight altitude threshold, and first speed threshold, to comprehensively determine the result of unauthorized flight. The time-series continuous data constraint avoids misjudgment caused by instantaneous data fluctuations. It combines multi-dimensional fixed quantitative indicators of radar physical tracks to lock the characteristics of small low-altitude illegal aircraft. It links track matching conflict state with target physical flight time-series features for verification, realizing hierarchical and progressive accurate identification of unauthorized flight behavior. It completes the judgment only by relying on standardized quantitative conditions, eliminating the identification error caused by single-dimensional instantaneous data, and improving the rigor and accuracy of unauthorized flight abnormal behavior identification.

[0016] In one possible implementation, based on the target's cross-source matching status, abnormal behavior identification of the target is performed to obtain the abnormal behavior identification result, including: if the target's cross-source matching status is a reporting mismatch status, extracting the radar cross-section, flight altitude, flight speed time-series features of the radar physical track, as well as the equipment registration number corresponding to the identity track; if the equipment corresponding to the identity track has no valid flight reporting record, and the radar physical track indicates that it is continuously in the low-altitude control area, and the radar cross-section, flight altitude, and flight speed all meet the first low-speed condition, then it is determined that the target has abnormal behavior of unlicensed reporting.

[0017] The UAV abnormal behavior identification method based on low-altitude situational awareness provided in this embodiment uses the cross-source matching status of the reported mismatch as a prerequisite for judgment. It simultaneously extracts the radar cross-section, flight altitude, and flight speed temporal features of the radar physical track and the equipment registration number corresponding to the identity track. It sequentially verifies the validity of the reporting record corresponding to the equipment registration number, the airspace location corresponding to the radar physical track, and whether the radar cross-section, flight altitude, and flight speed meet the first low-speed condition. Through multi-dimensional condition linkage verification of track matching status, equipment reporting file, low-altitude airspace location, and low-speed physical characteristics of the aircraft, it distinguishes between two types of violations: simple route deviation and unlicensed reporting without approval. This avoids category confusion caused by single-dimensional judgment and improves the differentiation accuracy and reliability of identifying unlicensed reporting-type abnormal behavior.

[0018] In one possible implementation, based on the target's cross-source matching status, abnormal behavior identification of the target is performed to obtain the abnormal behavior identification result, including: if the target's cross-source matching status is identity copying status, extracting the radar cross section, flight altitude, flight speed, and the equipment registration number corresponding to the identity track of the radar physical track; if the equipment registration number indicates that there is no stable and valid reporting record, and the radar cross section, flight altitude, and flight speed continuously meet the second low speed condition, then it is determined that the target has abnormal behavior of switching identity camouflage.

[0019] The UAV abnormal behavior identification method based on low-altitude situational awareness provided in this embodiment uses the cross-source matching state of identity duplication as a pre-triggered condition. It extracts the radar cross section, flight altitude, flight speed, and the equipment registration number corresponding to the identity track from the radar physical track. It sequentially verifies the stability and validity of the equipment registration number's reporting record and whether the radar cross section, flight altitude, and flight speed continuously meet the second low-speed condition. Through multi-level joint verification of the matching anomaly of multiple targets sharing the same identity on the track, the equipment reporting file, and the low-speed physical characteristics of the aircraft, it distinguishes the phenomenon of simple multiple track sharing of identity from identity spoofing behavior with low-altitude violation characteristics. It refines the judgment boundary between identity duplication phenomenon and real identity spoofing anomaly, avoids misjudgment caused by directly judging based on track matching state, and improves the accuracy and distinguishability of identity spoofing anomaly identification.

[0020] In one possible implementation, based on the target's cross-source matching state, abnormal behavior identification of the target is performed to obtain the abnormal behavior identification result, including: if the target's cross-source matching state is an identity change state, extracting the radar cross section, flight altitude, flight speed, and multiple sets of identity numbers of the radar physical track; if the multiple sets of identity numbers indicate that there are no stable and valid reporting records, and the radar cross section, flight altitude, and flight speed continuously meet the third low speed condition, then it is determined that the target has abnormal behavior of switching identity camouflage.

[0021] The UAV abnormal behavior recognition method based on low-altitude situational awareness provided in this embodiment uses identity change status as a pre-screening condition, extracts the radar cross-section, flight altitude, flight speed and multiple identity numbers of the radar physical track, and then verifies the stability and validity of the reporting records corresponding to the multiple identity numbers. At the same time, it verifies that the radar cross-section, flight altitude and flight speed continuously meet the third low-speed condition before qualitatively identifying identity spoofing abnormal behavior. Through the triple condition superposition verification of cross-source track anomaly, multi-identity file verification and low-altitude low-speed physical characteristics of the aircraft, it can filter out the temporary identity switching illusion caused by brief signal jumps, and only judge targets with complete violation characteristics as spoofing behavior. It refines the judgment boundary between identity change phenomenon and real identity spoofing violation, reduces the probability of false recognition based on multi-layer quantitative constraints, and improves the rigor of judging identity spoofing abnormal behavior.

[0022] Secondly, this application provides a device for identifying abnormal UAV behavior based on low-altitude situational awareness. The device includes: an acquisition module for acquiring low-altitude multi-source situational awareness data of a target area, including multi-band radar spot data, automatic dependent surveillance (ADS) identity broadcast data, and dedicated network reporting and approval records; a first determination module for determining, based on the multi-band radar spot data, ADS identity broadcast data, and dedicated network reporting and approval records, the radar physical track corresponding to the multi-band radar spot data, the identity track corresponding to the ADS identity broadcast data, and the planned reporting route corresponding to the dedicated network reporting and approval records; a second determination module for determining a first association confidence level between the radar physical track and the identity track, a second association confidence level between the radar physical track and the planned reporting route, and a third association confidence level between the identity track and the planned reporting route; a third determination module for determining the target cross-source matching state based on the first association confidence level, the second association confidence level, and the third association confidence level; and an identification module for identifying abnormal behavior of the target based on the target cross-source matching state, thereby obtaining an abnormal behavior identification result.

[0023] Thirdly, this application provides an electronic device, including: a memory and a processor, which are communicatively connected to each other. The memory stores computer instructions, and the processor executes the computer instructions to perform the above-described method for identifying abnormal UAV behavior based on low-altitude situational awareness, or any of its corresponding embodiments.

[0024] Fourthly, this application provides a computer-readable storage medium storing computer instructions for causing a computer to execute the method for identifying abnormal UAV behavior based on low-altitude situation described in the first aspect or any corresponding embodiment.

[0025] Fifthly, this application provides a computer program product, including computer instructions for causing a computer to execute the method for identifying abnormal UAV behavior based on low-altitude situation described in the first aspect or any corresponding embodiment. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the specific embodiments of this application or the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0027] Figure 1 This is a flowchart illustrating a method for identifying abnormal UAV behavior based on low-altitude situational awareness, according to an embodiment of this application. Figure 2 This is a flowchart illustrating another method for identifying abnormal drone behavior based on low-altitude situational awareness, according to an embodiment of this application. Figure 3 This is a flowchart illustrating another method for identifying abnormal drone behavior based on low-altitude situational awareness, according to an embodiment of this application. Figure 4 This is a flowchart illustrating another method for identifying abnormal drone behavior based on low-altitude situational awareness, according to an embodiment of this application. Figure 5 This is a flowchart illustrating another method for identifying abnormal drone behavior based on low-altitude situational awareness, according to an embodiment of this application. Figure 6 This is a schematic diagram of the hardware structure of a computer device according to an embodiment of this application. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0029] According to an embodiment of this application, an embodiment of a method for identifying abnormal behavior of unmanned aerial vehicles based on low-altitude situational awareness is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0030] This embodiment provides a method for identifying abnormal behavior of unmanned aerial vehicles based on low-altitude situational awareness, which can be applied to computer devices such as computers and servers. Figure 1 This is a flowchart illustrating the method for identifying abnormal UAV behavior based on low-altitude situational awareness according to an embodiment of this application, as shown below. Figure 1 As shown, the method for identifying abnormal UAV behavior based on low-altitude situational awareness includes: Step S101: Obtain low-altitude multi-source situational data of the target area. The low-altitude multi-source situational data includes multi-band radar spot data, automatic dependent surveillance identity broadcast data, and private network reporting and approval records.

[0031] Low-altitude multi-source situational awareness data for the target area can indicate heterogeneous raw sensing and approval data sets covering the monitored airspace, which may include multi-band radar spot data, broadcast automatic dependent surveillance identity broadcast data, private network reporting and approval records, etc.

[0032] Multi-band radar track data can indicate the original detection data of discrete points generated by multi-band radar echoes such as X, K, and S. Specifically, it can include coordinates, echo intensity, etc. After processing, RCS, velocity, and micro-Doppler can be calculated to generate real radar physical tracks.

[0033] Low-altitude multi-source situational awareness data for a target area can be used to indicate the low-altitude multi-source situational awareness data for a target object. The target object can be understood as a low-altitude flying vehicle such as a drone flying within the low-altitude airspace. The low-altitude range can be 0–1000 meters.

[0034] Automatic Dependent Surveillance (ADS) broadcast identity data can instruct drones to actively broadcast identity message data streams, storing IDs, declared drone models, location, speed, etc.

[0035] The dedicated network reporting and approval records can indicate the official flight approval ledger stored on the low-altitude air traffic control dedicated network, which may include routes, take-off and landing points, time windows, aircraft type restrictions, etc.

[0036] Specifically, low-altitude multi-source situational data of the target area can be obtained through radar detection, drone identity broadcasting, and government reporting.

[0037] In one possible implementation, unified timing and coordinate benchmarks can be used to aggregate low-altitude multi-source situational data of the target area.

[0038] As an example, a multi-band radar receiver, an ADS-B signal receiving unit, and a fiber optic interface for the management and control network are deployed at the edge base station for urban low-altitude control. The equipment is equipped with a unified Beidou timing module, which continuously and in parallel captures radar points, identity messages, and incremental records of reporting with a sampling period of 100ms. The raw data is locally cached for 10 seconds and then pushed to the local edge processor, while simultaneously recording data quality indicators such as radar signal-to-noise ratio and message packet loss rate.

[0039] Step S102: Based on multi-band radar spot data, automatic dependent surveillance identity broadcast data, and private network reporting and approval records, determine the radar physical track corresponding to the multi-band radar spot data, the identity track corresponding to the automatic dependent surveillance identity broadcast data, and the planned reporting route corresponding to the private network reporting and approval records.

[0040] Radar physical tracks can be understood as the real continuous motion trajectory of a target obtained by spatiotemporal registration, time-series stitching, and filtering of multi-band radar points.

[0041] Identity track can be understood as the trajectory formed by parsing ADS-B identity messages and concatenating declared coordinates and identity numbers in time.

[0042] The planned flight routes can be generated by parsing the private network reporting and approval records.

[0043] Specifically, for multi-band radar spot data, spatiotemporal registration to unify coordinates, clutter filtering to remove false alarms from buildings, and time-series stitching to generate continuous trajectories are performed. Parameters such as position, velocity, RCS, and micro-Doppler are calculated to form the first target dataset, and the radar physical track is output. For broadcast-type automatic dependent surveillance identity broadcast data, message sequence number verification, timestamp stability screening, and parsing of identity number and declared aircraft type are performed to form the second target dataset, and the publicly disclosed identity track is output. For private network reporting and approval records, approved routes, take-off and landing points, flight time windows, and aircraft type constraints are parsed to generate the third target dataset, and the planned reporting routes are output.

[0044] As an example, multi-band radar points are first uniformly converted to WGS84 geographic coordinates, and invalid points are filtered out using a static building clutter template. Adjacent frame points are sequentially stitched using Kalman filtering, and the RCS, acceleration, and micro-Doppler features of consecutive frames are calculated to generate the first target dataset. After verifying that the ADS-B message sequence number is not duplicated and backing up, the identity and aircraft type parameters are extracted to generate the identity track. The flight path polygon, take-off and landing coordinates, and time interval are extracted from the reporting record to generate the planned reporting flight path.

[0045] As an example, for multi-reflection airspace such as canyons and high-rise buildings, radar traces are spliced ​​using particle filtering to distinguish between the micro-Doppler features of drones and birds, and to eliminate multi-path reflection virtual points; identity tracks are added with message timestamp rollback detection to mark suspected replay identities; and reported routes are additionally bound with electronic fence height constraint parameters.

[0046] Step S103: Determine the first correlation confidence level between radar physical track and identity track, the second correlation confidence level between radar physical track and planned reporting route, and the third correlation confidence level between identity track and planned reporting route.

[0047] The first association confidence level can indicate the degree of matching between the quantified radar physical track and the identity track, and can be calculated by weighting the position residual, velocity residual, altitude residual, and heading residual.

[0048] The second correlation confidence level can indicate the degree of matching between the radar physical track and the planned reported route. It can be obtained by weighting multiple deviation items such as route deviation, take-off and landing deviation, time window deviation, and aircraft type matching.

[0049] The third association confidence level can indicate the degree of matching between the identity track and the planned reported route. It can be calculated by combining identity number matching, flight time overlap, and activity area overlap.

[0050] After determining the radar physical track, the identity track, and the planned reporting route, we can further determine the first correlation confidence level between the radar physical track and the identity track, the second correlation confidence level between the radar physical track and the planned reporting route, and the third correlation confidence level between the identity track and the planned reporting route.

[0051] As an example, between the radar physical track and the identity track, the position residual, velocity residual, heading residual, and altitude residual are calculated time-series, and weighted summation is performed using the radar equipment's preset target error scale to obtain the first association confidence level. Between the radar physical track and the planned reported route, multi-dimensional deviation items such as lateral route deviation, take-off and landing point deviation, time window deviation, and aircraft type matching degree are extracted and fused according to the dynamic credibility allocation weights of radar and reported data to obtain the second association confidence level. Between the identity track and the planned reported route, the identity number registration relationship, flight time overlap interval, and activity airspace overlap area are matched time-series, and normalized calculation is performed to obtain the third association confidence level.

[0052] Step S104: Determine the target cross-source matching status based on the first association confidence, the second association confidence, and the third association confidence.

[0053] The target cross-source matching status can indicate the target ternary matching classification by comparing three types of association confidence thresholds and superimposing identity duplication and identity mutation auxiliary judgment conditions. Among them, the target cross-source matching status can include stable legitimate association, reported mismatch, identity conflict, identity duplication, and identity mutation.

[0054] Specifically, after determining the first association confidence level, the second association confidence level, and the third association confidence level, the target cross-source matching status can be further determined.

[0055] As an example, a constant association threshold is set, and three sets of confidence levels are compared sequentially to obtain a basic judgment. The auxiliary conditions of identity duplication and identity mutation are checked in turn, and the statuses of stable and legal association, reporting mismatch, identity conflict, identity duplication, and identity mutation are output according to a fixed priority.

[0056] As an example, the system dynamically adjusts the associated thresholds based on the security level of the electronic fence. For sensitive core areas, the confidence threshold is lowered; for ordinary suburban airspace, the threshold is raised. The spatial thresholds and preset durations of auxiliary judgment conditions are adjusted synchronously according to the scenario. For instance, electronic fence classification standards are pre-stored. Level 1 sensitive core fences include airport airspace with the highest security level (Level 1), while Level 3 ordinary suburban fences include rural farmland and open suburban areas with the lowest security level (Level 3). Simultaneously, fixed baseline parameters for scenarios without classification are pre-stored, including a first associated baseline threshold of 0.7 for radar and identity matching qualification lines, a spatial baseline threshold of 150 meters for identity duplication auxiliary judgment, and a preset baseline duration of 20 seconds for identity mutation auxiliary judgment. Classification adjustment rules are also set, requiring the confidence threshold for Level 1 sensitive core areas to be lowered by 0.1. 5. The spatial threshold is expanded to 1.5 times the baseline, and the judgment time is shortened to 0.5 times the baseline. For Level 3 ordinary suburban areas, the confidence threshold needs to be increased by 0.15, the spatial threshold reduced to 0.6 times the baseline, and the judgment time extended to 1.8 times the baseline. For Level 1 sensitive core areas, the first correlation dynamic threshold is calculated to be 0.55, the identity replication spatial dynamic threshold is 225 meters, and the identity mutation dynamic duration is 10 seconds. When multi-band radar captures radar physical tracks Tr1 and Tr2 that are 180 meters apart and continuously match the same broadcast automatic correlation surveillance identity for 12 seconds, the threshold is adjusted accordingly. Under the condition of track Ti, the first association confidence scores of Tr1 and Ti (0.62) and Tr2 and Ti (0.58) both exceed the dynamic threshold to complete the basic matching. The distance between the two radar physical tracks exceeds the spatial threshold, and the matching time exceeds the dynamic time of identity mutation. Based on this, the target's cross-source matching state is determined to be an identity duplication state. Early identification of disguised drones is achieved by lowering the confidence threshold and shortening the judgment time in sensitive areas. For the ordinary suburban area of ​​suburban farmland (Level 3), the calculated first association dynamic threshold is 0.85, and the identity duplication spatial dynamic threshold is... With a threshold of 90 meters and a dynamic duration of identity mutation of 36 seconds, when the radar detects radar physical tracks TrA and TrB that are 110 meters apart and match the same broadcast automatic dependent surveillance identity track Ti for only 15 seconds, the first association confidence of TrA and Ti is 0.79, which does not reach the dynamic threshold, and the matching duration does not meet the dynamic duration of identity mutation. Therefore, the auxiliary judgment conditions for identity duplication and identity mutation are not triggered. It is only identified as short-term signal clutter interference and no abnormal cross-source state is marked. Invalid alarms are filtered according to the rules of raising the confidence line, narrowing the spatial threshold, and lengthening the judgment duration in urban and suburban areas.

[0057] Step S105: Based on the target cross-source matching status, perform abnormal behavior identification on the target object to obtain the abnormal behavior identification result.

[0058] The results of abnormal behavior identification can indicate the process of judging illegal behaviors such as unauthorized flight, unlicensed reporting, and identity spoofing, by taking the target cross-source matching status as a prerequisite and extracting radar time-series physical features and equipment registration files, combined with various low-speed thresholds and duration constraints.

[0059] Specifically, after determining the cross-source matching status of the target, abnormal behavior identification of the target can be further performed to obtain the abnormal behavior identification results.

[0060] As an example, the target cross-source matching status is used as the trigger condition for differentiated identification. Different matching statuses retrieve corresponding feature combinations and threshold constraints for processing. Specifically, in the identity conflict status, radar physical track radar cross section, flight altitude, and flight speed time series data are extracted. If the duration and low speed threshold are met, illegal flight is determined. In the reporting mismatch status, radar time series features and equipment registration number are extracted. If there is no valid reporting and the flight is continuously at low altitude and low speed, it is determined to be unlicensed reporting. In the identity duplication or identity mutation status, radar physical parameters and multiple identity numbers are extracted. If there is no stable reporting for multiple identities and the flight is continuously at low speed, identity masquerading is determined.

[0061] In one feasible scenario, the main urban area of ​​a provincial capital city is covered by multiple types of no-fly electronic fences, including government office areas, hospitals, and transportation hubs. 32 multi-band radars and ADS-B signal receiving base stations are deployed throughout the area, and a city-level low-altitude control network is established, with thousands of drone flights reported for daily operation.

[0062] The base stations on the edge of the main urban area collect real-time streaming data of multi-band radar spot data and broadcast automatic correlation surveillance identity broadcast data. The management network adds new network reporting and approval records every 5 minutes. The unified Beidou time synchronization completes data aggregation. Simultaneously collects radar blockage and message packet loss data for dynamic credibility calculation. Kalman filtering is used to process radar tracks, filtering out building clutter to generate radar physical tracks; ADS-B message sequence numbers are verified to remove replay data and generate identity tracks; reporting ledgers are parsed and bound to no-fly zones to generate planned reporting routes; weights are dynamically allocated based on the radar's real-time signal-to-noise ratio to calculate the first, second, and third association confidence levels between radar physical tracks and identity tracks, retaining complete residual and deviation evidence; the three sets of confidence levels are compared with the adaptive association threshold after the core area is downgraded, and two auxiliary conditions are simultaneously verified: whether the same identity matches multiple radar tracks and whether a single radar has multiple identities in a short period of time, outputting five types of target cross-source matching status; the matching status is divided into identification logic: for targets with conflicting identities, RCS, altitude, and speed are verified to determine unauthorized flights; for targets with mismatched reporting, the equipment reporting files are checked to identify unlicensed reporting; for targets with duplicated or mutated identities, multiple identity filing records are verified to identify identity spoofing; the probability trajectory prediction module is called simultaneously to push early warnings to abnormal targets that are predicted to enter the no-fly zone.

[0063] The UAV abnormal behavior identification method based on low-altitude situational awareness provided in this embodiment collects multi-source low-altitude situational awareness data of the target area, including multi-band radar point data, automatic correlation surveillance identity broadcast data, and private network reporting and approval records. Based on these three types of data, radar physical tracks, identity tracks, and planned reporting routes are generated respectively. Then, the first correlation confidence level between the radar physical track and the identity track, the second correlation confidence level between the radar physical track and the planned reporting route, and the third correlation confidence level between the identity track and the planned reporting route are calculated sequentially. The cross-source matching status of the target is determined based on these three types of correlation confidence levels. Combined with the target's cross-source matching status, the method automatically completes the identification of abnormal target behavior. The entire process requires no manual inspection, eliminating the limitations of limited coverage of manual inspections. Furthermore, by quantifying the matching relationship based on the multi-dimensional correlation confidence levels of the three types of tracks, it replaces subjective human judgment, avoiding the errors that easily occur during manual inspections. This simultaneously improves the coverage, accuracy, and efficiency of UAV identification.

[0064] In one possible implementation, step S102 above, which determines the radar physical track corresponding to the multi-band radar spot data, the identity track corresponding to the automatic dependent surveillance broadcast (ADS) identity broadcast data, and the planned flight path corresponding to the private network filing and approval records based on multi-band radar spot data, ADS identity broadcast data, and private network filing and approval records, may include the following steps: Step a1: Perform spatiotemporal registration processing on the multi-band radar spot data to obtain effective discrete point data.

[0065] Effective discrete point data can indicate the true target discreteness after multi-band radar spot traces have undergone spatiotemporal registration and clutter filtering to remove interference residues such as buildings and clouds.

[0066] Specifically, after determining the multi-band radar spot data, effective discrete point data can be further determined.

[0067] As an example, the spatiotemporal registration process unifies the BeiDou time stamp of the global radar and converts it to the unified WGS84 geographic coordinate system to eliminate point offset errors caused by clock differences and coordinate reference differences among multiple radar devices. After the registration is completed, clutter filtering is performed simultaneously to filter static clutter from mountains, tall buildings, and clouds, as well as false alarms from multipath reflections. Invalid detection points with a signal-to-noise ratio lower than the preset threshold are eliminated to obtain effective discrete point data of the real target points.

[0068] As an example, for vehicle-mounted and mobile portable multi-band radar, a dynamic rolling coordinate correction algorithm is used to complete the spatiotemporal registration process, and the coordinate offset is corrected according to the vehicle positioning. After registration, an adaptive signal-to-noise ratio threshold is used to dynamically filter out short-term dynamic clutter (birds, floating balloons), and to distinguish between low-speed natural interference and UAV location outputs effective discrete point data.

[0069] Step a2 involves performing time-series stitching on the effective discrete point data to form the target continuous motion trajectory corresponding to the effective discrete point data.

[0070] The continuous motion trajectory of the target can indicate a seamless and temporally coherent sequence of target points obtained by splicing the time sequence, which can be used to characterize the complete flight process of the aircraft.

[0071] Specifically, after determining the effective discrete point data, the target continuous motion trajectory corresponding to the effective discrete point data can be further determined.

[0072] As an example, Kalman filtering is used as the time-series stitching algorithm to correlate the effective discrete point data with the points in the previous and next frames. The missing points in the short time are filled in by predicting the motion model, and the smooth and uninterrupted continuous motion trajectory of the target is output.

[0073] As an example, for complex airspace with high-rise buildings, canyons, and multiple reflections, interactive multi-model particle filtering is used to complete the time-series stitching process, simultaneously distinguishing three types of moving target points: uniform speed, turning, and acceleration, eliminating false discrete points generated by multi-path reflections, and generating continuous target motion trajectories.

[0074] Step a3: Solve the continuous motion trajectory to obtain the first target dataset; wherein the first target dataset includes at least two of the following: calculated position, flight speed, radar cross section, rate of change of heading, flight acceleration, and physical characteristics.

[0075] The first target dataset can indicate a set of structured parameters calculated from a continuous motion trajectory. The first target dataset may include two or more parameters such as calculated position, flight speed, radar cross section, rate of change of heading, flight acceleration, and physical characteristics.

[0076] The calculated location can indicate the three-dimensional geographic coordinates of the target measured by the radar.

[0077] Flight speed can indicate the target's movement rate calculated based on the time difference between adjacent points.

[0078] Radar cross section can be inferred from radar echo intensity and is used to distinguish the size categories of targets such as drones, birds, and balloons.

[0079] The rate of change of heading indicates the extent of change in the heading of a target per unit time.

[0080] Physical characteristics can include micro-Doppler spectrum, echo multi-band ratio, etc., to distinguish rotorcraft, fixed-wing UAVs from natural interference objects, etc.

[0081] Specifically, after determining the continuous motion trajectory, the first target dataset can be further determined.

[0082] As an example, the coordinates of each frame are extracted as the solution position; the flight speed is solved by the displacement or time difference between adjacent frames; the radar cross section is obtained by combining the radar echo energy; a polynomial fit is performed on the continuous trajectory, and the rate of change of heading is calculated by the time difference of the heading angle; the flight acceleration is obtained by the time difference of the flight speed; the echo spectrum is extracted to generate micro-Doppler-like physical features; and then the above parameters are packaged and structured and stored as the first target dataset.

[0083] Step a4: Generate radar physical tracks based on the first target dataset.

[0084] Specifically, after determining the first target dataset, the radar physical track can be further determined.

[0085] As an example, the calculated position, flight speed, radar cross section, rate of change of heading, flight acceleration, and physical features of the first target dataset are reused, and a complete time-series radar physical track is generated by binding timestamps frame by frame.

[0086] Step a5: Perform a first parsing process on the broadcast-type automatic dependent surveillance identity broadcast data to obtain a second target dataset; wherein, the second target dataset includes at least one of the following: identity number, time-series coordinates, motion parameters, declared aircraft type, message sequence number key identifier, and motion information.

[0087] The second target dataset can indicate the set of information obtained by parsing the identity message. The second target dataset may include two or more items such as identity number, time coordinates, motion parameters, declared aircraft type, message serial number key identifier, and motion information.

[0088] The identification number can indicate the unique registration identifier of the drone equipment.

[0089] Time-series coordinates can indicate the three-dimensional spatial position of the drone as declared in the identity message.

[0090] Motion parameters can indicate the self-reported speed, altitude, and heading data within the identity message.

[0091] The declared aircraft type can indicate the aircraft model that the drone operator must register.

[0092] The message sequence number is a key identifier that indicates the sequential numbering of messages and can be used to identify replay and forged messages.

[0093] Specifically, after determining the broadcast identity data of the automatic dependent surveillance broadcast, a second target dataset can be further determined.

[0094] As an example, the first parsing process is performed on the broadcast-type automatic dependent surveillance identity broadcast data of drone broadcasts. First, the continuity of key identifiers of message sequence number and whether the timestamp has been rolled back are checked to filter out identity replay and duplicate forged messages. Then, the identity number, time coordinate, motion parameters, declared aircraft type and motion information fields in the message are extracted in a structured manner and packaged to generate the second target dataset.

[0095] Step a6: Generate identity tracks based on the second target dataset.

[0096] Specifically, after determining the second target dataset, identity tracks can be further determined.

[0097] As an example, the output time-series second target dataset is grouped and concatenated by identity number, and sorted by message timestamp. The self-reported time coordinates and motion parameters of each frame are spliced ​​together to form an identity track.

[0098] Step a7 involves performing a second parsing process on the private network reporting and approval records to obtain a third target dataset; wherein the third target dataset includes at least one of the following: approved flight routes, take-off and landing points, flight time windows, permitted altitude ranges, and constraints on the declared aircraft type.

[0099] The third target dataset can instruct the parsing of private network reporting and approval records to generate a set of compliance constraints, which can approve at least two or more constraints, such as flight routes, take-off and landing points, flight time windows, permitted altitude ranges, and declared aircraft types.

[0100] Approved flight routes can indicate the permitted flight path polygons approved by the control authorities.

[0101] The takeoff and landing points can indicate the approved geographical coordinates for takeoff and landing.

[0102] Flight time windows can indicate the range of start and end times during which flight operations are permitted.

[0103] The permitted altitude range can indicate the maximum and minimum altitudes for flight.

[0104] The declared aircraft type constraint can indicate the type of aircraft that can be approved for a license.

[0105] Specifically, after determining the private network filing and approval records, a third target dataset can be further determined.

[0106] As an example, the private network registration and approval records stored in the low-altitude private network are retrieved, and the second parsing process is performed to structurally disassemble the approval constraint clauses. The polygon coordinates of the approved flight route, take-off and landing points, start and end times of the flight time window, upper and lower limits of the allowable altitude range, and aircraft type constraints are extracted and packaged into the third target dataset.

[0107] Step a8: Based on the third target dataset, generate the planned flight routes for reporting.

[0108] Specifically, after determining the third target dataset, the planned flight routes can be further determined.

[0109] As an example, based on the third target dataset, the approved routes, take-off and landing points, flight time windows, permitted altitude ranges, and aircraft type constraints are all standardized and encapsulated in terms of time and space to generate planned reporting routes with complete compliance constraints.

[0110] The UAV abnormal behavior recognition method based on low-altitude situational awareness provided in this embodiment obtains a first target dataset and generates radar physical tracks by performing spatiotemporal registration, time-series stitching, and trajectory calculation on multi-band radar point data. It also obtains a second target dataset and generates identity tracks by performing a first parsing on automatic dependent surveillance identity broadcast data, and obtains a third target dataset and generates planned reporting routes by performing a second parsing on private network reporting and approval records. Based on the three types of low-altitude monitoring data, it automatically generates three types of standardized tracks. The entire process does not require manual inspection to collect and organize airspace target information. It overcomes the shortcomings of manual inspection, which has limited coverage and can only observe locally. Furthermore, it solves the problem of errors that are easily caused by relying on subjective visual records during manual inspection by uniformly extracting and calculating multi-dimensional objective data such as location, identity number, and approved routes through the data processing flow.

[0111] In one possible implementation, determining the first association confidence level between the radar physical track and the identity track, the second association confidence level between the radar physical track and the planned reported route, and the third association confidence level between the identity track and the planned reported route in step S203 above may include the following steps: Step b1: Based on the radar physical track and the identity track, determine the position residual, velocity residual, heading residual, and altitude residual between the radar physical track and the identity track.

[0112] Position residuals can indicate the spatial difference between the radar-calculated position and the self-reported three-dimensional coordinates at the same moment.

[0113] The velocity residual indicates the difference between the radar-measured flight speed and the speed reported by the identity message at the same moment.

[0114] The heading residual can indicate the angle difference between the heading angle calculated by the radar and the heading angle reported by the identity at the same time.

[0115] Altitude residuals can indicate the difference between the radar-measured altitude and the altitude reported by the identity message at the same moment.

[0116] Specifically, after determining the radar physical track and the identity track, the position residual, velocity residual, heading residual, and altitude residual between the radar physical track and the identity track can be further determined.

[0117] As an example, following the time alignment rules, the radar physical track and the identity track are unified to the same timestamp. The radar's measured position, flight speed, heading angle, and altitude are extracted frame by frame at the same time, as well as the identity track's self-reported time coordinates, motion parameters, and altitude data. The difference is calculated using the corresponding parameters, and four types of deviations are output: position residual, velocity residual, heading residual, and altitude residual.

[0118] Step b2: Determine the first correlation confidence level based on the position residual, velocity residual, heading residual, altitude residual, and target error scale.

[0119] The target error scale can indicate the inherent measurement error calibration constant of radar and ADS-B equipment.

[0120] Specifically, after determining the position residual, velocity residual, heading residual, and altitude residual, the first correlation confidence level can be determined by combining the target error scale.

[0121] As an example, the four types of residuals are dimensionlessly normalized using a preset target error scale to eliminate differences in distance, speed, and angle units; the normalized residuals are converted into individual matching scores using an exponential decay function; and the scores for position, speed, heading, and altitude are linearly weighted and summed to obtain the first association confidence level.

[0122] Step b3: Based on the radar physical track and the planned reported route, determine the multi-dimensional deviation items corresponding to the target; among them, the multi-dimensional deviation items include at least two of the following: lateral deviation distance, deviation between the actual take-off and landing point and the reported take-off and landing point, deviation between the flight time period and the approval time window, and the matching degree between the target aircraft type and the declared aircraft type.

[0123] Multidimensional deviation terms can indicate a set of quantitative deviation indicators between radar-measured flight data and official approval constraints.

[0124] Lateral deviation distance indicates the vertical straight-line distance from the target's real-time coordinates to the centerline of the approved flight path.

[0125] The deviation between the actual take-off and landing point and the reported take-off and landing point can indicate the spatial distance between the aircraft's actual take-off and landing coordinates and the approved take-off and landing coordinates.

[0126] The deviation between the flight time period and the approval time window can indicate the time difference between the actual start and end times of the target flight and the approved time interval.

[0127] The matching degree between the target aircraft type and the declared aircraft type can indicate the score of the radar physical characteristics inferred from the aircraft type and the reported and registered aircraft type.

[0128] Specifically, after determining the radar physical track and the planned reporting route, the multi-dimensional deviation items corresponding to the target can be further determined.

[0129] As an example, a full-domain comparison is performed between the complete time-series data of the radar physical track and the approval constraints of the planned and reported flight routes. The lateral deviation distance from the target coordinates to the approved flight route is calculated at each time step. Key take-off and landing coordinates of the aircraft are extracted, and the deviation between the actual take-off and landing points and the reported take-off and landing points is calculated. The complete flight start and end times are statistically analyzed, and the deviation between the flight period and the approval time window is calculated. Based on the radar cross-section and physical characteristics within the radar physical track, the target aircraft type is inferred, and a score is assigned to the reported and declared aircraft type to obtain the matching degree between the target aircraft type and the declared aircraft type. Any two or more of the above can form a multi-dimensional deviation item.

[0130] Step b4: Perform weighted fusion processing on the multi-dimensional bias terms to determine the second association confidence level.

[0131] Specifically, after determining the multidimensional bias terms, the second association confidence level can be further determined.

[0132] As an example, we first perform reverse normalization of each type of deviation item from 0 to 1, then dynamically allocate fusion weights based on the timeliness of the reported data update and the reliability of radar detection, and finally normalize the overall output to output the second correlation confidence level in the 0-1 interval after weighted summation.

[0133] As an example, the update time of the private network filing and approval record is read. If the filing information expires, the weight of aircraft type matching and take-off and landing deviation is reduced. When the radar detection is stable, the weight of lateral deviation of the route is increased. The second association confidence is generated by dynamically weighting and fusing multi-dimensional deviation items.

[0134] Step b5: Based on the identity track and the planned reported route, determine the matching relationship of the identity number corresponding to the target, the overlap of flight time, and the overlap of activity area.

[0135] The identity number matching relationship is a binary matching identifier that indicates whether the device ID carried by the identity track matches the ID registered in the private network filing ledger.

[0136] Flight time overlap indicates the percentage of time that overlaps between the actual flight time of an identified flight track and the reported permitted time window.

[0137] The overlap of activity areas can indicate the percentage of overlap between the identity track flight airspace polygon and the reported and approved operation airspace polygon.

[0138] Specifically, after determining the identity track and the planned reported route, it is possible to further determine the matching relationship of the identity number corresponding to the target, the degree of overlap of flight time, and the degree of overlap of activity area.

[0139] As an example, the identity track carrying the identity number is compared with the registration ID bound to the planned flight route to output the binary identity number matching relationship; the complete flight duration of the identity track and the allowed flight time interval are statistically analyzed, and the overlap ratio is calculated to obtain the flight time overlap; spatial overlay operation is performed on the activity airspace polygon of the identity track and the approved operation polygon to calculate the overlap area ratio to obtain the activity area overlap.

[0140] Step b6: Determine the third association confidence level based on the identity number matching relationship, flight time overlap, and activity area overlap.

[0141] Specifically, after determining the identity number matching relationship, flight time overlap, and activity area overlap, the third association confidence level can be further determined.

[0142] As an example, a basic weight is assigned to the matching relationship of binary identity numbers, and then the overlap of flight time and overlap of activity area are weighted and superimposed. The three are then normalized to the range of 0 to 1 to generate a third association confidence score.

[0143] The UAV abnormal behavior recognition method based on low-altitude situational awareness provided in this embodiment calculates the first correlation confidence score by combining the position residual, velocity residual, heading residual, and altitude residual of radar physical tracks and identity tracks with the target error scale. Then, it obtains the second correlation confidence score by weighted fusion of multi-dimensional deviation terms between radar physical tracks and planned reported routes. Finally, it calculates the third correlation confidence score based on the identity number matching relationship between identity tracks and planned reported routes, flight time overlap, and activity area overlap. The entire process relies on multiple types of objective quantitative differences and matching indicators to complete numerical judgment, eliminating the need for manual inspection and relying on visual and subjective experience to judge the target matching situation. This solves the shortcomings of limited coverage of manual inspection and inability to simultaneously compare all dimensions of the three types of tracks, as well as the problem of errors that are easily generated by subjective judgment in manual inspection.

[0144] In one possible implementation, determining the target cross-source matching status based on the first association confidence level, the second association confidence level, and the third association confidence level in step S204 above may include the following steps: Step c1: Determine the first judgment result based on the first association confidence level and the first association threshold.

[0145] The first association threshold can indicate a pre-set critical value used to delineate the boundary between radar and identity track matching or non-matching. Specifically, it can be dynamically adjusted according to the electronic fence security level.

[0146] Specifically, after determining the first association confidence level and the first association threshold, the first judgment result can be further determined.

[0147] As an example, the first association confidence level is compared with the first association threshold. If the first association confidence level is greater than or equal to the first association threshold, a first judgment result of matching is output, indicating that the drone's broadcast identity and the actual flight measured by radar are basically consistent. If the first association confidence level is less than the first association threshold, a first judgment result of mismatch is output, indicating that the spatiotemporal and motion deviations between the two are too large, and there is a risk of false identity or identity conflict.

[0148] Step c2: Determine the second judgment result based on the second association confidence level and the second association threshold.

[0149] The second correlation threshold can indicate the critical value that distinguishes whether a real flight meets the reporting requirements.

[0150] Specifically, after determining the second association confidence level and the second association threshold, a second judgment result can be further determined.

[0151] As an example, the second association confidence level is compared with the second association threshold: if the second association confidence level is greater than or equal to the second association threshold, the second judgment result of the report matching is output, which indicates that the drone route, take-off and landing, model and time all meet the approval requirements; if the second association confidence level is less than the second association threshold, the second judgment result of the report mismatch is output, which indicates that the actual flight deviates from the approval boundary.

[0152] Step c3: Determine the third judgment result based on the third association confidence level and the third association threshold.

[0153] The third association threshold can indicate the critical value that distinguishes whether the identity number, airspace, and time are consistent with the reported information.

[0154] Specifically, after determining the third association confidence level and the third association threshold, the third judgment result can be further determined.

[0155] As an example, the third association confidence level is compared with the third association threshold: if the third association confidence level is greater than or equal to the third association threshold, the third judgment result of identity reporting matching is output, which indicates that the current identity number has a valid approval record and the space and time are matched; if the third association confidence level is less than the third association threshold, the third judgment result of identity reporting mismatch is output, which indicates that the identity has no corresponding compliant reporting or the scope of operation is seriously inconsistent.

[0156] Step c4: Based on the first judgment result, the second judgment result, the third judgment result, and the auxiliary judgment conditions, determine the target cross-source matching status.

[0157] Specifically, after determining the first judgment result, the second judgment result, the third judgment result, and the auxiliary judgment conditions, the target cross-source matching status can be further determined.

[0158] As an example, first determine whether there is a case of identity duplication where the same identity track matches multiple (two or more) radar tracks exceeding the spatial threshold, and whether a single continuous radar track matches multiple sets of different identity tracks within a preset time period. If any of the auxiliary conditions are met, the target cross-source matching status of identity duplication or identity mutation is directly output. If neither of the two auxiliary judgment conditions is met, the first judgment result, the second judgment result, and the third judgment result are combined for comprehensive distinction. When the first judgment result and the third judgment result are both matches, it is judged as a stable and legal association status. When the second judgment result is a mismatch, it is judged as a reporting mismatch status. When the first judgment result is a mismatch, it is judged as an identity conflict status.

[0159] Specifically, the auxiliary determination conditions include whether the same identity track matches multiple radar tracks exceeding the spatial threshold and whether a single continuous radar track matches multiple sets of different identity tracks within a preset time period; wherein, step c4 above includes: Step c41: If the same identity track matches multiple radar tracks that exceed the spatial threshold, then the target cross-source matching state is determined to be an identity copying state.

[0160] The spatial threshold can indicate a pre-set spatial determination threshold.

[0161] Identity duplication status can instruct multiple drones to share the same legitimate identity message for disguised flight.

[0162] Specifically, it iterates through all the pairing relationships between identity tracks and radar physical tracks. For a single track with the same identity, it checks whether it is simultaneously bound to two or more independent radar physical tracks whose spatial distance exceeds the spatial threshold. If the condition is met, the target's cross-source matching status is directly determined to be an identity copying status.

[0163] For example, a spatial threshold of 150 meters is set, and a single-frame scanning cycle of 1 second is used to search for identity-radar pairing relationships frame by frame. As long as the same identity track matches two radar physical tracks with a distance of more than 150 meters in a single frame, it is instantly marked as identity copying status.

[0164] Step c42: If a single continuous radar track matches multiple sets of tracks with different identities within a preset time period, the target cross-source matching status is determined to be an identity change status.

[0165] The preset duration can indicate a continuous timer window used to detect frequent identity switching.

[0166] The identity mutation state can instruct a single drone to frequently switch between different identity messages during flight to evade supervision.

[0167] Specifically, a timing window is opened for each continuous radar track, and the identity track numbers that are paired successively within the preset time window are counted. If two or more sets of completely different identity numbers appear in the window, the target cross-source matching status is determined to be an identity mutation status.

[0168] For example, a fixed preset duration of 20 seconds is used to count the identity number on each radar track in a sliding 20-second window. If two or more different IDs appear in the window, a sudden identity change is directly determined.

[0169] Step c43: If the same identity track does not match multiple radar tracks that exceed the spatial threshold, a single continuous radar track does not match multiple sets of different identity tracks within a preset time period, the first determination result indicates that the radar physical track and the identity track match, and the third determination result indicates that the identity track matches the planned reported route, the target cross-source matching status is determined to be a stable and legal association status.

[0170] A stable and legitimate association status can indicate that the information of radar-measured flight, drone identity disclosure, and official reporting and approval are consistent.

[0171] Specifically, after excluding high-risk spoofing such as identity duplication and identity mutation, the determination of a compliant and non-abnormal target status requires the simultaneous fulfillment of three pre-constraints: no multiple radars sharing the same identity, no single machine switching multiple identities in a short period of time, the first determination result being a match between radar and identity, and the third determination result being a match between identity and reporting. Then, the target's cross-source matching status can be determined to be a stable and legitimate association status.

[0172] Step c44: If the same identity track does not match multiple radar tracks that exceed the spatial threshold, a single continuous radar track does not match multiple sets of tracks with different identities within a preset time period, and the second determination result indicates that the radar physical track does not match the planned reported route, the target cross-source matching status is determined to be a reported mismatch status.

[0173] The "mismatch in reporting" status can indicate that the drone's identity is valid and matches the reporting, but the actual flight route, time, and take-off and landing points deviate from the official approval constraints.

[0174] Specifically, if the second determination result shows that the actual flight measured by the radar does not match the officially planned flight route; in this state, the drone's identity is real and valid, and the identity can correspond to the reported file, but the actual flight deviates from the approved flight route, time, take-off and landing, and altitude constraints, then the target cross-source matching status is determined to be a reporting mismatch status.

[0175] Step c45: If the same identity track does not match multiple radar tracks that exceed the spatial threshold, a single continuous radar track does not match multiple sets of different identity tracks within a preset time, and the first determination result indicates that the radar physical track and identity track do not match, the target cross-source matching status is determined to be an identity conflict status.

[0176] An identity conflict state can indicate that the identity information broadcast by a drone is seriously inconsistent with the actual flight time and space and motion characteristics measured by radar.

[0177] Specifically, if the first determination result shows that the radar physical track and the identity track fail to match, it indicates that there is a discrepancy between the information such as the position, speed, and altitude broadcast by the UAV and the objective flight data measured by the radar, and there is a risk of illegal flight such as broadcasting false messages and disabling the real identity. In this case, the target cross-source matching status is determined to be an identity conflict status.

[0178] The UAV abnormal behavior recognition method based on low-altitude situational awareness provided in this embodiment obtains a first judgment result and a second judgment result by comparing the first and second association confidence scores with their corresponding association thresholds, respectively. A third judgment result is obtained by comparing the third association confidence score with the third association threshold. The target cross-source matching status is then comprehensively determined by combining the three types of judgment results and auxiliary judgment conditions. By combining hierarchical threshold comparison with multi-dimensional results for joint analysis, the evaluation criteria for the matching relationship of radar physical tracks, identity tracks, and planned reporting routes are unified. This avoids the one-sidedness caused by single-dimensional comparison, fully integrates the matching quantitative information of the three types of tracks, distinguishes different track matching combinations, reduces the bias caused by single confidence score judgment, and improves the completeness and reliability of the target cross-source matching status classification.

[0179] By layering the target cross-source matching status based on the first, second, and third judgment results, and superimposing two auxiliary judgment conditions—whether the same identity track matches multiple radar tracks exceeding the spatial threshold, and whether a single continuous radar track matches multiple sets of tracks with different identities within a preset time period—the system differentiates the target cross-source matching status in a hierarchical manner. Judgment logic is set for five scenarios: identity duplication, identity mutation, stable and legitimate association, reporting mismatch, and identity conflict. This distinguishes between track matching deviations and special anomalies such as identity forgery, addressing the shortcoming of failing to identify multi-track cross-masking behavior based solely on comparisons of the confidence thresholds of the three association types. The system refines the granularity of target cross-source matching status classification, allowing various track coupling and contradictory scenarios to be categorized, thus improving the detail and scenario adaptability of target cross-source matching status division.

[0180] like Figure 2 As shown, in one possible implementation, step S206 above, which involves identifying abnormal behavior of the target object based on the target cross-source matching status, to obtain the abnormal behavior identification result, may include the following steps: Step S2061: If the target cross-source matching status is an identity conflict status, continuously extract the radar cross section, flight altitude and flight speed time series data within the radar physical track.

[0181] Radar cross section can indicate the quantitative physical parameters obtained by radar echo inversion, and can be used to distinguish targets such as small drones, birds, and balloons.

[0182] Flight altitude can indicate the vertical height of a target above the ground as measured by radar.

[0183] Flight speed time-series data can indicate a frame-by-frame flight speed sequence stored with consecutive timestamps, which can be used to determine the characteristics of a target maintaining low-speed flight for a long period of time.

[0184] Specifically, if the target cross-source matching status is in an identity conflict state, then the radar cross section, flight altitude, and flight speed time series data can be extracted from the radar physical track.

[0185] As an example, the system monitors the cross-source matching status of targets. Once an identity conflict is identified, the system automatically initiates continuous time-series extraction logic. It continuously reads the radar cross-section, flight altitude, and flight speed frame by frame from the complete radar physical track, and packages them into continuous time-series data for joint verification.

[0186] Step S2062: If the radar cross section within the preset duration is not greater than the preset cross section threshold, the flight altitude is not greater than the preset altitude threshold, and the flight speed is not greater than the first speed threshold, the abnormal behavior identification result is determined to be the result of the target object flying illegally.

[0187] The preset scattering cross-sectional area threshold can indicate the critical echo parameter that distinguishes between small drones, large aircraft, and naturally floating objects.

[0188] The first speed threshold can indicate the critical speed for low-speed determination.

[0189] The presence of unauthorized flights of a target object can indicate the illegal flight behavior of small, low-altitude drones without valid and credible identities or proper registration.

[0190] Specifically, if the target's radar cross-section is consistently less than or equal to a preset cross-section threshold, its flight altitude is consistently less than or equal to a preset altitude threshold, and its flight speed is consistently less than or equal to a first speed threshold, then it can be determined that there is a small, low-altitude, illegal drone flight without a valid and credible identity and without proper reporting.

[0191] For example, preset globally unified parameters: preset duration 30 seconds, preset scattering cross-sectional area threshold 0.1㎡, preset height threshold 100 meters, and first velocity threshold 15m / s; the time-series data is completely traversed through a 30-second window, and only when all three thresholds are met in every frame within the window is it determined to be an unauthorized flight.

[0192] The UAV abnormal behavior identification method based on low-altitude situational awareness provided in this embodiment triggers a time-series feature extraction process based on the target cross-source matching state of identity conflict. It continuously extracts time-series data of radar cross section, flight altitude, and flight speed corresponding to radar physical tracks. Then, it combines preset duration constraints to synchronously verify three quantitative indicators: radar cross section threshold, flight altitude threshold, and first speed threshold, to comprehensively determine the result of unauthorized flight. The time-series continuous data constraint avoids misjudgment caused by instantaneous data fluctuations. It combines multi-dimensional fixed quantitative indicators of radar physical tracks to lock the characteristics of small low-altitude illegal aircraft. It links track matching conflict state with target physical flight time-series features for verification, realizing hierarchical and progressive accurate identification of unauthorized flight behavior. It completes the judgment only by relying on standardized quantitative conditions, eliminating the identification error caused by single-dimensional instantaneous data, and improving the rigor and accuracy of unauthorized flight abnormal behavior identification.

[0193] like Figure 3 As shown, in one possible implementation, step S206 above, which involves identifying abnormal behavior of the target object based on the target cross-source matching status, to obtain the abnormal behavior identification result, may include the following steps: Step S3061: If the target cross-source matching status is a reported mismatch status, extract the radar cross section, flight altitude, flight speed time sequence features of the radar physical track, and the device registration number corresponding to the identity track.

[0194] The device registration number corresponding to the identity track can indicate the unique registered device ID of the UAV bound in the automatic dependent surveillance broadcast message, which can be used to retrieve all flight reporting and approval ledgers within the low-altitude airspace control network.

[0195] Specifically, the system monitors the cross-source matching status of targets. Once a mismatch is detected, a two-dimensional data synchronization extraction operation is triggered. On one hand, it continuously extracts the temporal characteristics of radar cross-section, flight altitude, and flight speed from radar physical tracks to preserve objective temporal evidence of the target's continuous low-altitude, low-speed flight. On the other hand, it analyzes the registration number of the current identity track reading device as a unique index for retrieving the low-altitude private network reporting ledger.

[0196] Step S3062: If the device corresponding to the identity track has no valid flight reporting record, and the radar physical track indication is continuously in the low-altitude control area, and the radar cross-section, flight altitude, and flight speed all meet the first low-speed condition, then it is determined that the target object has abnormal behavior of unlicensed reporting.

[0197] The first low-speed condition can indicate the upper limit of radar cross-section, upper limit of flight altitude, and upper limit of flight speed.

[0198] Specifically, by searching the ledger through the equipment registration number, it is confirmed that the current equipment does not have a valid flight reporting record that is suitable for the current flight time and space. The radar physical track time series data proves that the target has been staying in the low-altitude control area for a long time, and the radar cross-section, flight altitude and flight speed throughout the entire process all meet the first low speed condition. Therefore, it is determined that the target has abnormal behavior of unlicensed reporting.

[0199] For example, if fixed first low-speed condition parameters are set: radar cross-section ≤ 0.1㎡, flight altitude ≤ 100m, and flight speed ≤ 15m / s, and all of these conditions are met for 30 consecutive seconds and there is no matching reporting record in the local ledger for the equipment registration number, it is determined that the registration is unlicensed.

[0200] The UAV abnormal behavior identification method based on low-altitude situational awareness provided in this embodiment uses the cross-source matching status of the reported mismatch as a prerequisite for judgment. It simultaneously extracts the radar cross-section, flight altitude, and flight speed temporal features of the radar physical track and the equipment registration number corresponding to the identity track. It sequentially verifies the validity of the reporting record corresponding to the equipment registration number, the airspace location corresponding to the radar physical track, and whether the radar cross-section, flight altitude, and flight speed meet the first low-speed condition. Through multi-dimensional condition linkage verification of track matching status, equipment reporting file, low-altitude airspace location, and low-speed physical characteristics of the aircraft, it distinguishes between two types of violations: simple route deviation and unlicensed reporting without approval. This avoids category confusion caused by single-dimensional judgment and improves the differentiation accuracy and reliability of identifying unlicensed reporting-type abnormal behavior.

[0201] like Figure 4 As shown, in one possible implementation, step S206 above, which involves identifying abnormal behavior of the target object based on the target cross-source matching status, to obtain the abnormal behavior identification result, may include the following steps: Step S4061: If the target cross-source matching status is identity copying status, extract the radar cross section, flight altitude, flight speed, and device registration number corresponding to the identity track of the radar physical track.

[0202] Specifically, the system continuously monitors the cross-source matching status of targets. Once it is determined that the current target matching relationship is in the identity duplication state, the synchronous data extraction logic is immediately initiated: on the one hand, it continuously extracts three types of time-series physical parameters, namely radar cross section, flight altitude, and flight speed, frame by frame from multiple radar physical tracks that match the identity, so as to retain objective evidence of synchronous low-altitude activities of multiple targets; on the other hand, it analyzes the identity track corresponding to the shared identity and reads the bound device registration number.

[0203] Step S4062: If the device registration number indicates that there is no stable and valid reporting record, and the radar cross section, flight altitude, and flight speed continuously meet the second low speed condition, then it is determined that the target has abnormal behavior of switching identities and camouflaging.

[0204] The second low-speed condition can indicate the judgment criteria set for identity cloning and camouflage targets, simultaneously constraining the upper limits of radar cross-section, flight altitude, and flight speed. For example, radar cross-section ≤ 0.08㎡, flight altitude ≤ 80m, and flight speed ≤ 12m / s.

[0205] Switching identities to masquerade as abnormal behavior can instruct multiple aircraft to share the same legitimate equipment registration number to evade airspace supervision, and can result in long-term low-altitude, low-speed flight without valid approval.

[0206] Specifically, by searching the dedicated network filing and approval records through the equipment registration number, it is confirmed that the equipment does not have a stable and valid filing record that is compatible with the current airspace and time period, and the radar cross-section, flight altitude, and flight speed corresponding to multiple radar physical tracks meet the second low speed condition within a continuous timing window, then it is determined that the target has abnormal behavior of switching identities and camouflaging.

[0207] The UAV abnormal behavior identification method based on low-altitude situational awareness provided in this embodiment uses the cross-source matching state of identity duplication as a pre-triggered condition. It extracts the radar cross section, flight altitude, flight speed, and the equipment registration number corresponding to the identity track from the radar physical track. It sequentially verifies the stability and validity of the equipment registration number's reporting record and whether the radar cross section, flight altitude, and flight speed continuously meet the second low-speed condition. Through multi-level joint verification of the matching anomaly of multiple targets sharing the same identity on the track, the equipment reporting file, and the low-speed physical characteristics of the aircraft, it distinguishes the phenomenon of simple multiple track sharing of identity from identity spoofing behavior with low-altitude violation characteristics. It refines the judgment boundary between identity duplication phenomenon and real identity spoofing anomaly, avoids misjudgment caused by directly judging based on track matching state, and improves the accuracy and distinguishability of identity spoofing anomaly identification.

[0208] like Figure 5 As shown, in one possible implementation, step S206 above, which involves identifying abnormal behavior of the target object based on the target cross-source matching status, to obtain the abnormal behavior identification result, may include the following steps: Step S5061: If the target cross-source matching status is a sudden identity change status, extract the radar cross section, flight altitude, flight speed and multiple sets of identity numbers of the radar physical track.

[0209] Multiple sets of identification numbers can indicate the set of unique device identifiers corresponding to multiple identification tracks matched sequentially within a preset time window for a single radar physical track.

[0210] Specifically, the system monitors the cross-source matching status of the target. Once a sudden identity change is detected, two types of continuous data extraction logic are initiated simultaneously: First, the system continuously extracts the radar cross section, flight altitude, and flight speed time-series physical parameters frame by frame from the single continuous radar physical track, preserving objective measured evidence of the target's low-altitude, low-speed flight throughout the entire process; Second, the system summarizes all identity tracks matched sequentially within the timing window and extracts all identity numbers to form multiple sets of identity numbers.

[0211] Step S5062: If multiple sets of identity number indications do not have stable and valid reporting records, and the radar cross section, flight altitude, and flight speed continuously meet the third low speed condition, then it is determined that the target has abnormal behavior of switching identities and disguise.

[0212] The third low-speed condition can indicate the upper limit of radar cross-section, upper limit of flight altitude, and upper limit of flight speed. For example, radar cross-section ≤ 0.12 m², flight altitude ≤ 120 m, and flight speed ≤ 18 m / s.

[0213] The abnormal behavior of switching identities can instruct a single drone to circumvent supervision by repeatedly changing multiple sets of identity numbers during flight, and all identities are not reported stably and effectively, and the drone maintains low-altitude and low-speed flight.

[0214] Specifically, if multiple sets of identity numbers indicate that there are no stable and valid reporting records, and the radar cross section, flight altitude, and flight speed continuously meet the third low-speed condition, it proves that the target implementing the identity switching is a small low-altitude low-speed UAV, and that the private network reporting and approval records corresponding to each set of numbers of the UAV do not have stable and valid reporting records that are compatible with the current flight airspace and flight time period. In this case, it is determined that the target has abnormal behavior of switching identity and disguise.

[0215] The UAV abnormal behavior recognition method based on low-altitude situational awareness provided in this embodiment uses identity change status as a pre-screening condition, extracts the radar cross-section, flight altitude, flight speed and multiple identity numbers of the radar physical track, and then verifies the stability and validity of the reporting records corresponding to the multiple identity numbers. At the same time, it verifies that the radar cross-section, flight altitude and flight speed continuously meet the third low-speed condition before qualitatively identifying identity spoofing abnormal behavior. Through the triple condition superposition verification of cross-source track anomaly, multi-identity file verification and low-altitude low-speed physical characteristics of the aircraft, it can filter out the temporary identity switching illusion caused by brief signal jumps, and only judge targets with complete violation characteristics as spoofing behavior. It refines the judgment boundary between identity change phenomenon and real identity spoofing violation, reduces the probability of false recognition based on multi-layer quantitative constraints, and improves the rigor of judging identity spoofing abnormal behavior.

[0216] This embodiment also provides a UAV abnormal behavior recognition device based on low-altitude situational awareness. This device is used to implement the above embodiments and preferred embodiments, and details already described will not be repeated. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0217] This embodiment provides a device for identifying abnormal behavior of unmanned aerial vehicles (UAVs) based on low-altitude situational awareness. The device includes: an acquisition module for acquiring low-altitude multi-source situational awareness data of a target area, including multi-band radar spot data, automatic dependent surveillance (ADS) identity broadcast data, and private network reporting and approval records; a first determination module for determining, based on the multi-band radar spot data, ADS identity broadcast data, and private network reporting and approval records, the radar physical track corresponding to the multi-band radar spot data, the identity track corresponding to the ADS identity broadcast data, and the planned reporting route corresponding to the private network reporting and approval records; a second determination module for determining a first correlation confidence level between the radar physical track and the identity track, a second correlation confidence level between the radar physical track and the planned reporting route, and a third correlation confidence level between the identity track and the planned reporting route; a third determination module for determining the target cross-source matching status based on the first, second, and third correlation confidence levels; and an identification module for identifying abnormal behavior of the target based on the target cross-source matching status, obtaining an abnormal behavior identification result.

[0218] Further functional descriptions of the above modules and units are the same as those in the corresponding embodiments described above, and will not be repeated here.

[0219] In this embodiment, the UAV abnormal behavior recognition device based on low-altitude situation is presented in the form of a functional unit. Here, a functional unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that execute one or more software or fixed programs, and / or other devices that can provide the above functions.

[0220] This application also provides a computer device having the above-described device for identifying abnormal drone behavior based on low-altitude situation.

[0221] Please see Figure 6 , Figure 6 This is a schematic diagram of the structure of a computer device provided in an optional embodiment of this application, such as... Figure 6As shown, the computer device includes one or more processors 10, memory 20, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components communicate with each other via different buses and can be mounted on a common motherboard or otherwise installed as needed. The processors can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as display devices coupled to the interfaces). In some alternative implementations, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple computer devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system). Figure 6 Take a processor 10 as an example.

[0222] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GDA), or any combination thereof.

[0223] The memory 20 stores instructions executable by at least one processor 10 to cause at least one processor 10 to perform the method shown in the above embodiments.

[0224] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the computer device. Furthermore, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0225] The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0226] The computer device also includes an input device 30 and an output device 40. The processor 10, memory 20, input device 30, and output device 40 can be connected via a bus or other means. Figure 6 Taking the example of a connection between China and Israel via a bus.

[0227] Input device 30 can receive input numerical or character information, and generate key signal inputs related to user settings and function control of the computer device, such as a touchscreen, keypad, mouse, trackpad, touchpad, joystick, one or more mouse buttons, trackball, joystick, etc. Output device 40 may include display devices, auxiliary lighting devices (e.g., LEDs), and haptic feedback devices (e.g., vibration motors). The aforementioned display devices include, but are not limited to, liquid crystal displays, light-emitting diodes, displays, and plasma displays. In some alternative embodiments, the display device may be a touchscreen.

[0228] The computer device also includes a communication interface for communicating with other devices or communication networks.

[0229] This application also provides a computer-readable storage medium. The methods described in this application can be implemented in hardware or firmware, or implemented as recordable on a storage medium, or implemented as computer code downloaded over a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and subsequently stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the methods shown in the above embodiments are implemented.

[0230] A portion of this application can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to this application through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0231] Although embodiments of this application have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of this application, and all such modifications and variations fall within the scope defined by the appended claims.

Claims

1. A method for identifying abnormal behavior of unmanned aerial vehicles (UAVs) based on low-altitude situational awareness, characterized in that, The method includes: Acquire low-altitude multi-source situational data of the target area, including multi-band radar spot data, automatic dependent surveillance identity broadcast data, and private network reporting and approval records. Based on multi-band radar spot data, automatic dependent surveillance identity broadcast data, and private network reporting and approval records, the radar physical track corresponding to the multi-band radar spot data, the identity track corresponding to the automatic dependent surveillance identity broadcast data, and the planned reporting route corresponding to the private network reporting and approval records are determined. Determine the first correlation confidence level between radar physical track and identity track, the second correlation confidence level between radar physical track and planned reported route, and the third correlation confidence level between identity track and planned reported route; Based on the first association confidence level, the second association confidence level, and the third association confidence level, the target cross-source matching status is determined; Based on the target cross-source matching status, abnormal behavior identification is performed on the target object to obtain the abnormal behavior identification result; Based on the first association confidence level, the second association confidence level, and the third association confidence level, the target cross-source matching status is determined, including: Based on the first association confidence level and the first association threshold, a first determination result is determined; Based on the second association confidence level and the second association threshold, a second determination result is determined; Based on the third association confidence level and the third association threshold, a third determination result is determined; Based on the first determination result, the second determination result, the third determination result, and the auxiliary determination conditions, the target cross-source matching status is determined; The auxiliary determination conditions include whether the same identity track matches multiple radar tracks exceeding the spatial threshold and whether a single continuous radar track matches multiple sets of different identity tracks within a preset time period. The determination of the target cross-source matching status based on the first determination result, the second determination result, the third determination result, and the auxiliary determination conditions includes: If the same identity track matches multiple radar tracks that exceed the spatial threshold, the target cross-source matching status is determined to be identity duplication status; If a single continuous radar track matches multiple tracks with different identities within a preset time period, the target's cross-source matching status is determined to be an identity change status. If the same identity track does not match multiple radar tracks that exceed the spatial threshold, a single continuous radar track does not match multiple sets of different identity tracks within a preset time period, the first determination result indicates that the radar physical track and the identity track match, and the third determination result indicates that the identity track matches the planned reported route, the target cross-source matching status is determined to be a stable and legal association status. If the same identity track does not match multiple radar tracks exceeding the spatial threshold, a single continuous radar track does not match multiple sets of different identity tracks within a preset time period, or the second determination result indicates that the radar physical track does not match the planned reported route, the target cross-source matching status is determined to be a reported mismatch status. If the same identity track does not match multiple radar tracks exceeding the spatial threshold, a single continuous radar track does not match multiple sets of different identity tracks within a preset time period, and the first determination result indicates that the radar physical track and identity track do not match, the target cross-source matching status is determined to be an identity conflict status.

2. The method for identifying abnormal UAV behavior based on low-altitude situational awareness according to claim 1, characterized in that, Based on multi-band radar spot data, automatic dependent surveillance (ADS) identity broadcast data, and dedicated network reporting and approval records, the radar physical tracks corresponding to the multi-band radar spot data, the identity tracks corresponding to the ADS identity broadcast data, and the planned reporting routes corresponding to the dedicated network reporting and approval records are determined, including: Spatiotemporal registration processing is performed on the multi-band radar spot data to obtain effective discrete point data; The effective discrete point data is subjected to time-series splicing processing to form the target continuous motion trajectory corresponding to the effective discrete point data; The continuous motion trajectory is processed to obtain a first target dataset; wherein, the first target dataset includes at least two of the following: calculated position, flight speed, radar cross section, rate of change of heading, flight acceleration, and physical characteristics; Based on the first target dataset, generate radar physical tracks; The broadcast-type automatic dependent surveillance identity broadcast data is subjected to a first parsing process to obtain a second target dataset; wherein, the second target dataset includes at least one of the following: identity number, time-series coordinates, motion parameters, declared aircraft model, message sequence number key identifier, and motion information; Based on the second target dataset, generate identity tracks; The private network filing and approval records are subjected to a second parsing process to obtain a third target dataset; wherein, the third target dataset includes at least one of the following: approved flight routes, take-off and landing points, flight time windows, permissible altitude ranges, and aircraft type constraints. Based on the third target dataset, the planned flight route is generated.

3. The method for identifying abnormal UAV behavior based on low-altitude situational awareness according to claim 1, characterized in that, Determine the first correlation confidence level between radar physical tracks and identity tracks, the second correlation confidence level between radar physical tracks and planned reported routes, and the third correlation confidence level between identity tracks and planned reported routes, including: Based on the radar physical track and the identity track, determine the position residual, velocity residual, heading residual and altitude residual between the radar physical track and the identity track; The first association confidence level is determined based on the position residual, the velocity residual, the heading residual, the altitude residual, and the target error scale; Based on the radar physical track and the planned reported route, determine the multi-dimensional deviation items corresponding to the target; wherein, the multi-dimensional deviation items include at least two of the following: lateral deviation distance, deviation between actual take-off and landing point and reported take-off and landing point, deviation between flight time period and approval time window, and matching degree between target aircraft type and declared aircraft type; The multi-dimensional deviation items are weighted and fused to determine the second association confidence level; Based on the identity track and the planned reported route, determine the identity number matching relationship, flight time overlap and activity area overlap of the target; Based on the identity number matching relationship, flight time overlap, and activity area overlap, a third association confidence level is determined.

4. The method for identifying abnormal UAV behavior based on low-altitude situational awareness according to claim 1, characterized in that, The step of identifying abnormal behavior of the target object based on the target cross-source matching state and obtaining the abnormal behavior identification result includes: If the target cross-source matching status is an identity conflict status, continuously extract the radar cross section, flight altitude and flight speed time series data within the radar physical track; If the radar cross section is not greater than the preset cross section threshold, the flight altitude is not greater than the preset altitude threshold, and the flight speed is not greater than the first speed threshold within the preset duration, the abnormal behavior identification result is determined to be that the target object is flying illegally.

5. The method for identifying abnormal UAV behavior based on low-altitude situational awareness according to claim 1, characterized in that, The step of identifying abnormal behavior of the target object based on the target cross-source matching state and obtaining the abnormal behavior identification result includes: If the target's cross-source matching status is a reported mismatch status, extract the radar cross section, flight altitude, flight speed time sequence features of the radar physical track, as well as the device registration number corresponding to the identity track; If the device corresponding to the identity track has no valid flight reporting record, and the radar physical track indication is continuously in the low-altitude control area, and the radar cross-section, flight altitude, and flight speed all meet the first low-speed condition, then the target object is determined to have abnormal behavior of unlicensed reporting.

6. The method for identifying abnormal UAV behavior based on low-altitude situational awareness according to claim 1, characterized in that, The step of identifying abnormal behavior of the target object based on the target cross-source matching state and obtaining the abnormal behavior identification result includes: If the target's cross-source matching status is identity copying, extract the radar cross section, flight altitude, flight speed, and the device registration number corresponding to the identity track of the radar physical track; If the equipment registration number indicates that there are no stable and valid reporting records, and the radar cross section, flight altitude, and flight speed continuously meet the second low speed condition, then it is determined that the target has abnormal behavior of switching identities and camouflaging.

7. The method for identifying abnormal UAV behavior based on low-altitude situational awareness according to claim 1, characterized in that, The step of identifying abnormal behavior of the target object based on the target cross-source matching state and obtaining the abnormal behavior identification result includes: If the target's cross-source matching status is in an identity change state, extract the radar cross section, flight altitude, flight speed, and multiple sets of identity numbers of the radar physical track; If multiple sets of identification numbers indicate that there are no stable and valid reporting records, and the radar cross section, flight altitude, and flight speed continuously meet the third low-speed condition, then it is determined that the target is exhibiting abnormal behavior such as switching identities and camouflaging.

8. An electronic device, characterized in that, include: A processor and a memory connected to the processor; wherein the memory stores instructions executable by the processor, the instructions being executed by the processor to cause the processor to perform the UAV abnormal behavior recognition method based on low-altitude situation as described in any one of claims 1-7.

Images