Low-altitude unmanned aerial vehicle trajectory tracking and monitoring method based on 5G-A communication and inductance integrated base station

Abstract

The invention discloses a low-altitude unmanned aerial vehicle trajectory tracking and monitoring method based on a 5G-A communication sensing integrated base station, and relates to the technical field of low-altitude traffic management and communication sensing fusion, and the method comprises the steps: firstly collecting multi-source data such as a communication sensing fusion signal, environment interference and unmanned aerial vehicle attributes, and carrying out the alignment and packaging of a unified timestamp and a coordinate system into a synchronous data frame; then, deep fusion and anti-interference processing are carried out on the data frames, noise is filtered out, and pure fusion data is generated; and furthermore, real-time track calculation and motion trend prediction are carried out on pure data by utilizing multi-base-station cooperative calculation and prediction. Based on this, through a multi-target feature recognition and clustering separation mechanism, independent individual trajectories are accurately stripped from a complex mixed data stream, and compliance verification and anomaly judgment are performed on the trajectories in combination with an airspace rule base. In this way, the problems of signal interference and multi-target aliasing in a complex environment can be effectively solved, and therefore high-precision global tracking of the low-altitude unmanned aerial vehicle and real-time monitoring of abnormal behaviors are achieved.

Description

Technical Field

[0001] This application relates to the field of low-altitude traffic management and communication sensing fusion technology, and more specifically, to a method for tracking and monitoring the trajectory of low-altitude unmanned aerial vehicles based on a 5G-A integrated sensing base station. Background Technology

[0002] With the booming development of the low-altitude economy, drones are increasingly being used in urban logistics, environmental monitoring, and emergency rescue, making low-altitude airspace an important space for economic activities. Drone flight trajectory data, encompassing key information such as location coordinates and flight speed, is a core element of airspace management and has irreplaceable value in ensuring low-altitude safety, preventing illegal intrusion, and optimizing airspace resource allocation. Currently, drone monitoring technology based on cellular networks has become a research hotspot. Existing technical solutions mostly employ multi-base station collaboration, primarily determining drone coordinates by measuring 4G base station signal strength and combining collaborative data from multiple base stations.

[0003] Specifically, existing technologies primarily rely on the signal characteristics of communication links for calculations. While utilizing existing base station resources, they suffer from significant technical shortcomings when facing complex urban low-altitude environments. First, such solutions are essentially still communication-based positioning, lacking proactive sensing capabilities. When drones are silent or communication signals are blocked, the system cannot effectively locate them based on signal strength, leading to monitoring failure. Second, regarding anti-interference and data fusion, existing technologies often physically and logically separate communication and sensing functions. Base stations are only responsible for data transmission and cannot accurately align and deeply fuse radar sensing data with communication data at the microsecond-level timestamp, creating data silos where "communication is possible but location is unknown, and sensing is possible but identity is difficult to establish." Furthermore, existing target recognition mechanisms typically employ feature-weighted similarity algorithms, ignoring the dynamic impact of complex environments (such as rainfall or strong electromagnetic interference) on feature stability, resulting in insufficient recognition accuracy and robustness, making it difficult to meet the requirements of low-altitude surveillance. Finally, in multi-target scenarios, existing technologies lack refined clustering and separation mechanisms for individual signal characteristics. When multiple drones fly simultaneously, signal aliasing and trajectory misjudgment are very likely to occur, making it difficult to meet the regulatory requirements for meter-level high-precision tracking and real-time identification of abnormal behavior.

[0004] Therefore, there is an urgent need for an optimized method for tracking and monitoring the trajectory of low-altitude unmanned aerial vehicles (UAVs). Summary of the Invention

[0005] This application is made in order to solve the above-mentioned technical problems.

[0006] According to one aspect of this application, a method for tracking and monitoring the trajectory of a low-altitude unmanned aerial vehicle (UAV) based on a 5G-A integrated sensing base station is provided, comprising: Data alignment, coordinate unification, and packaging are performed on the synthetic fusion signal, environmental interference dataset, and UAV basic attribute data to obtain synchronized raw data frames; The synchronized raw data frames are subjected to synesthetic data fusion and anti-interference processing to obtain clean fused data; Trajectory calculation and prediction are performed on clean fused data provided by multiple base stations to obtain real-time trajectory and predicted trajectory; Based on a clean fused data stream, multi-target identification and trajectory separation are performed on real-time trajectories to obtain independent individual trajectories; Based on predicted trajectories and a spatial rule base, abnormal trajectories of independent individuals are identified to obtain abnormal events.

[0007] Compared with existing technologies, this application provides a low-altitude UAV trajectory tracking and monitoring method based on a 5G-A integrated sensing base station. First, it collects multi-source data including sensing fusion signals, environmental interference, and UAV attributes, which are then encapsulated into synchronized data frames after being aligned with a unified timestamp and coordinate system. Subsequently, the data frames undergo deep fusion and anti-interference processing to filter out noise and generate clean fused data. Then, using multi-base station collaborative calculation and prediction, real-time trajectory calculation and motion trend prediction are performed on the clean data. Based on this, through multi-target feature recognition and clustering separation mechanisms, independent individual trajectories are accurately extracted from the complex mixed data stream, and the trajectories are validated for compliance and anomaly determination using a spatial rule base. This effectively overcomes the challenges of signal interference and multi-target aliasing in complex environments, thereby achieving high-precision full-domain tracking and real-time monitoring of abnormal behavior of low-altitude UAVs. Attached Figure Description

[0008] The above and other objects, features, and advantages of this application will become more apparent from the more detailed description of the embodiments of this application in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the embodiments of this application to explain this application and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0009] Figure 1 This is a flowchart of a low-altitude UAV trajectory tracking and monitoring method based on a 5G-A integrated sensing base station, according to an embodiment of this application.

[0010] Figure 2 This is a data flow diagram of a low-altitude UAV trajectory tracking and monitoring method based on a 5G-A integrated sensing base station according to an embodiment of this application.

[0011] Figure 3 This is a flowchart of sub-step S2 of the low-altitude UAV trajectory tracking and monitoring method based on a 5G-A integrated sensing base station according to an embodiment of this application.

[0012] Figure 4 This is a flowchart of sub-step S3 of the low-altitude UAV trajectory tracking and monitoring method based on a 5G-A integrated sensing base station according to an embodiment of this application.

[0013] Figure 5 This is a flowchart of sub-step S4 of the low-altitude UAV trajectory tracking and monitoring method based on a 5G-A integrated sensing base station according to an embodiment of this application.

[0014] Figure 6 This is a flowchart of sub-step S42 of the low-altitude UAV trajectory tracking and monitoring method based on a 5G-A integrated sensing base station according to an embodiment of this application. Detailed Implementation

[0015] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0016] To address the problems mentioned above, this application proposes a method for tracking and monitoring the trajectory of low-altitude unmanned aerial vehicles (UAVs) based on a 5G-A integrated sensing base station. Figure 1 This is a flowchart of a low-altitude UAV trajectory tracking and monitoring method based on a 5G-A integrated sensing base station, according to an embodiment of this application. Figure 2 This is a data flow diagram of a low-altitude UAV trajectory tracking and monitoring method based on a 5G-A integrated sensing base station, according to an embodiment of this application. Figure 1 and Figure 2 As shown, the low-altitude UAV trajectory tracking and monitoring method based on a 5G-A integrated sensing base station includes the following steps: S1, performing data alignment, coordinate unification, and packaging on the sensing fusion signal, environmental interference dataset, and UAV basic attribute data to obtain synchronized original data frames; S2, performing sensing data fusion and anti-interference processing on the synchronized original data frames to obtain clean fusion data; S3, performing trajectory calculation and prediction on the clean fusion data provided by multiple base stations to obtain real-time trajectories and predicted trajectories; S4, performing multi-target identification and trajectory separation on the real-time trajectory based on the clean fusion data stream to obtain independent individual trajectories; S5, performing abnormal trajectory identification on the independent individual trajectories based on the predicted trajectory and airspace rule base to obtain abnormal events.

[0017] In the aforementioned low-altitude UAV trajectory tracking and monitoring method based on a 5G-A integrated sensing base station, step S1 involves aligning, unifying coordinates, and packaging the sensing fusion signal, environmental interference dataset, and UAV basic attribute data to obtain synchronized raw data frames. It should be understood that due to the different sources of the sensing fusion signal, environmental interference data, and UAV basic attribute data, there are issues with inconsistent timestamps and different spatial coordinate systems, leading to ineffective correlation in subsequent data processing. Therefore, this application aligns the sensing fusion signal, environmental interference dataset, and UAV basic attribute data in terms of time dimension, unifies spatial coordinates, and structures them for packaging, thereby constructing a standardized raw data foundation. This eliminates the heterogeneity of multi-source data, ensuring that subsequent fusion processing, trajectory calculation, and other steps can be carried out based on a consistent data benchmark, providing a prerequisite guarantee for the accuracy of the entire monitoring process.

[0018] Specifically, in one possible embodiment, step S1 is implemented as follows: First, multi-source data acquisition is performed. The 5G-A integrated sensing base station deployed at key urban nodes activates its various functional modules: the millimeter-wave radar module transmits detection signals and receives echo signals reflected by the UAV, capturing core sensing information such as distance, angle, and speed; the Massive MIMO communication module captures the UAV's uplink 5G-A communication signals in real time, demodulating and extracting data such as device identification and flight status; the environmental perception sensor simultaneously collects environmental interference data such as electromagnetic interference intensity, temperature and humidity, visibility, and terrain occlusion parameters. Simultaneously, basic attribute data such as registration and filing information associated with the device identification, aircraft parameters, maximum flight speed, and endurance are retrieved from the UAV management system database through a standardized interface. Next, the high-precision GPS synchronization clock source built into the base station edge computing node adds a unified format timestamp to the collected sensing fusion signals and environmental interference data, ensuring that the time base of all data streams is consistent, with accuracy controlled within 1ms. Subsequently, coordinate unification was performed, converting all spatial location-related data, including UAV and base station locations, from the sensor fusion signal to the WGS-84 geodetic coordinate system, completely eliminating coordinate system differences between different acquisition modules. Finally, data packaging was completed. Following a pre-defined structured data specification, the time-aligned and coordinate-unified sensor fusion signal, environmental interference data, and associated UAV basic attribute data were integrated and encapsulated to form a raw data frame containing data source identifiers, timestamps, spatial coordinates, and various attribute information. This ensures data integrity, consistency, and traceability, providing standardized, high-quality input data for subsequent sensor data fusion and anti-interference processing.

[0019] In the aforementioned method for tracking and monitoring the trajectory of low-altitude UAVs based on a 5G-A integrated sensing base station, step S2 involves performing sensing data fusion and anti-interference processing on the synchronized original data frames to obtain clean fused data. It should be understood that because the synchronized original data frames contain redundant information such as environmental noise and electromagnetic interference, and the sensing and communication signals are not effectively correlated, the data quality is insufficient to support high-precision trajectory monitoring. Therefore, this application further performs deep correlation of sensing data and targeted filtering of environmental interference on the original data frames to obtain fused data that is both complete and reliable. This breaks down the separation between sensing and communication data, reduces the impact of complex environments on the data, and provides high-quality data support for subsequent accurate trajectory calculation and prediction.

[0020] In particular, in one specific embodiment, Figure 3 This is a flowchart of sub-step S2 of the low-altitude UAV trajectory tracking and monitoring method based on a 5G-A integrated sensing base station according to an embodiment of this application. Figure 3 As shown, step S2 includes: S21, performing cross-domain data association matching on the synchronized original data frame to obtain associated data objects; S22, performing fusion filtering and state estimation on the associated data objects to obtain clean fused data.

[0021] Specifically, step S21 involves performing cross-domain data association matching on the synchronized original data frames to obtain the associated data objects. It should be understood that, because the core sensing information such as distance, angle, and speed collected by the millimeter-wave radar module in the synchronized original data frames is separated from the device identification and flight status data captured and demodulated by the Massive MIMO communication module, lacking a clear correspondence, it is impossible to accurately attribute multi-dimensional data to specific UAV individuals, affecting the targeted nature of subsequent processing. Therefore, this application further conducts precise cross-domain data association and matching based on the spatiotemporal characteristics in the data frames to establish a unique mapping relationship between signal features and individual UAVs. This clarifies the subject of each set of data, eliminates data confusion in multi-target scenarios, provides a precise data association foundation for subsequent anti-interference processing, state estimation, and multi-target separation, and ensures the targeted nature and effectiveness of the entire monitoring process.

[0022] Specifically, in one possible embodiment, step S21 is implemented as follows: First, the spatial position, velocity information, and corresponding timestamps collected and calculated by the millimeter-wave radar module are extracted from the original data frame, along with key data such as device identification and flight status obtained by the Massive MIMO communication module demodulation. Next, using the high-precision timestamp assigned by the base station edge computing node as a benchmark, position-velocity information and device identification-flight status information within the same time window are selected to ensure data consistency in the time dimension. Finally, based on the correlation of spatial positions, a matching verification is performed, binding position-velocity information and device identification-flight status information belonging to the same UAV to form structured associated data objects. Data that fails to match is marked as targets to be identified, and its feature information is fully preserved, ensuring the comprehensiveness and accuracy of the association matching.

[0023] Specifically, step S22 involves performing fusion filtering and state estimation on the associated data objects to obtain clean fused data. It is worth noting that the clean fused data includes identity identifiers, optimal state estimates, and associated attributes. The optimal state estimates include estimated position, estimated velocity, and covariance matrix, while the associated attributes include aircraft type information, registration information, and data source identifiers. It should be understood that although the associated data objects establish a correspondence between identity and data, they still contain abnormal measurement values ​​caused by meteorological interference, electromagnetic noise, etc., and lack a comprehensive and accurate characterization of the UAV's motion state, making it impossible to directly support high-precision trajectory calculation. Therefore, this application further performs fusion filtering on the associated data objects to remove noise, performs state estimation to obtain complete motion parameters, and integrates identity identifiers and associated attributes to form high-quality clean fused data. This results in data with unique identity, accurate state, and complete attributes, effectively improving data reliability and providing a solid data foundation for subsequent multi-base station collaborative positioning, multi-target trajectory separation, and abnormal behavior identification, ensuring the high-precision operation of the entire monitoring system.

[0024] Specifically, in one possible embodiment, step S22 is implemented as follows: First, to address complex environmental interference, a dual processing strategy is employed to preprocess the data: a narrow beam is generated by focusing the UAV's direction using Massive MIMO beamforming technology to suppress sidelobe interference signals. Second, using the identity identifier in the associated data object as the core index, the corresponding original measurement data such as position and velocity are extracted. A Kalman filter-particle filter fusion algorithm is then used to process these data layer by layer, filtering out abnormal data caused by meteorological noise and terrain occlusion, thus correcting positioning errors. Specifically, this algorithm primarily targets complex environmental interference, using the identity identifier in the associated data object as the core index to extract the corresponding original measurement data such as position and velocity, and then processes these data layer by layer. Its core function is to filter out abnormal data caused by meteorological noise (such as rain and fog attenuation) and terrain occlusion (such as reflections from tall buildings), thereby correcting positioning errors. Based on the filtered effective data, the estimated position and velocity of the UAV are calculated using a state estimation model, while simultaneously calculating the covariance matrix to quantify the reliability of the state estimation. Finally, the identity identifier, the above-mentioned optimal state estimation results, and the model information, registration information, and data source identifier extracted from the associated data objects are structurally integrated to form complete, clean, and fused data, ensuring that the information in each dimension of the data is interconnected and accurate.

[0025] In the aforementioned low-altitude UAV trajectory tracking and monitoring method based on a 5G-A integrated sensing base station, step S3 involves calculating and predicting the trajectory of clean fusion data provided by multiple base stations to obtain real-time and predicted trajectories. It should be understood that clean fusion data from a single base station has limited coverage and can only reflect the UAV's motion state from a single perspective, failing to meet the requirements for high-precision tracking across the entire area and early risk avoidance. Furthermore, the high-speed movement of UAVs demands extremely high real-time trajectory updates. Therefore, this application further integrates clean fusion data from multiple base stations to perform real-time trajectory calculation and short-term motion trend prediction, thereby constructing complete and forward-looking trajectory information. This overcomes the monitoring limitations of a single base station, enabling accurate depiction of the UAV's trajectory and early prediction of its future path, providing a sufficient time window for subsequent anomaly identification and emergency response, and ensuring the timeliness and effectiveness of regulatory response.

[0026] In particular, in one specific embodiment, Figure 4 This is a flowchart of sub-step S3 of the low-altitude UAV trajectory tracking and monitoring method based on a 5G-A integrated sensing base station according to an embodiment of this application. Figure 4As shown, step S3 includes: S31, setting the clean fusion data from this base station as local clean fusion data and setting the clean fusion data from at least two neighboring cooperative base stations as cooperative multi-source positioning data; S32, performing multi-station cooperative positioning calculation on the local clean fusion data and cooperative multi-source positioning data to obtain the current location point; S33, updating the real-time trajectory status of the current location point based on the historical trajectory point sequence to obtain the real-time trajectory; S34, performing time-series trajectory prediction on the real-time trajectory to obtain the predicted trajectory.

[0027] Specifically, in step S31, the clean fusion data from this base station is designated as local clean fusion data, and the clean fusion data from at least two neighboring cooperative base stations is designated as cooperative multi-source positioning data. It should be understood that since the clean fusion data provided by multiple base stations comes from different sources, failure to clearly distinguish them can lead to confusion in data attribution during subsequent cooperative calculations, failing to fully leverage the complementary advantages of different base station data and affecting positioning accuracy. Therefore, this application further classifies and identifies the clean fusion data provided by multiple base stations, clearly defining local data and cooperative data from at least two neighboring base stations, thereby establishing a clear foundation for data collaboration. This enables the accurate retrieval of data from different sources during subsequent positioning calculations, fully utilizing the real-time nature of local data and the spatial complementarity of neighboring base station data, providing orderly and controllable data support for multi-station cooperative positioning, and ensuring the accuracy and reliability of the positioning results.

[0028] Specifically, in one possible embodiment, step S31 is implemented as follows: First, using the base station identification information inherent in the data, the clean fused data collected and processed by this base station is marked as local clean fused data to ensure rapid retrieval of local data. Then, based on a preset base station cooperative networking scheme, at least two neighboring base stations with overlapping coverage areas and stable communication links with this base station are selected. Subsequently, using a standardized data transmission protocol, the clean fused data sent by these neighboring base stations is received and uniformly marked as collaborative multi-source positioning data. Finally, the two types of data are stored in different data cache areas, along with a collection timestamp and base station location information, to facilitate rapid association and retrieval during subsequent collaborative calculations.

[0029] Specifically, step S32 involves performing multi-station collaborative positioning calculations on the local clean fusion data and the collaborative multi-source positioning data to obtain the current location. It should be understood that the positioning error of clean fusion data from a single base station is relatively large due to factors such as measurement angle and environmental interference, making it difficult to meet the monitoring requirements for meter-level positioning of UAVs. However, multi-base station data possesses spatial complementarity and can mutually correct measurement deviations. Therefore, this application further integrates local and collaborative multi-source positioning data, conducting joint calculations through a multi-station collaborative positioning model to improve the positioning accuracy of the UAV's current location. This fully leverages the spatial collaborative advantages of multiple base stations, effectively offsetting the measurement errors of a single base station, improving positioning accuracy to the meter level, providing a precise location benchmark for subsequent real-time trajectory generation and anomaly detection, and ensuring the core data quality of the entire monitoring process.

[0030] Specifically, in one possible embodiment, step S32 is implemented as follows: First, the angle of arrival (AHA) parameter and signal strength measured by the local base station are extracted from the local clean fusion data, and the time difference of arrival (TDOA) parameter and signal strength of the neighboring cooperative base stations relative to the local base station are extracted from the cooperative multi-source positioning data. Specifically, a local Cartesian coordinate system needs to be constructed, with the local base station coordinates set as the origin coordinates of 0 meters, 0 meters, and 30 meters, and the coordinates of the two cooperative base stations set as 500 meters, 0 meters, and 30 meters, and 250 meters, 433 meters, and 30 meters, respectively. The system reads that the azimuth angle of the UAV measured by the local base station is 45 degrees and the pitch angle is 60 degrees, and simultaneously reads that the time difference of arrival between the first cooperative base station and the local base station is -0.5 microseconds, and the time difference of arrival between the second cooperative base station and the local base station is 0.2 microseconds. Then, the above time differences are converted into distance difference constraints, and the distance difference from the UAV to the first cooperative base station and the local base station is calculated to be -150 meters, and the distance difference to the second cooperative base station and the local base station is 60 meters. To eliminate errors from a single data source and solve for 3D coordinates, the system constructs an overdetermined set of observation equations including ray constraints on the angle of arrival and hyperboloid constraints on the time difference of arrival, and solves it using weighted least squares. During this process, signal strength indices are used for weighting; for example, the weighting coefficient for a signal strength of -70 dB / mW received by the local base station is set to 0.9, and the weighting coefficient for a signal strength of -75 dB / mW received by the cooperating base station is set to 0.8, thereby reducing the impact of low-quality signals on the positioning results. Finally, the system uses a Gauss-Newton iterative algorithm to numerically solve the above weighted observation equations. The system uses a rough distance of 200 meters estimated by signal strength combined with angle information as the initial iteration value. After three iterations, the iteration stops when the magnitude of the position update is less than the set convergence threshold of 0.1 meters. The final output is the solved 3D coordinate values ​​of 100.2 meters, 99.8 meters, and 150.1 meters as the current position of the UAV, thus achieving meter-level high-precision positioning with an error control within three meters.

[0031] Specifically, step S33 involves updating the real-time trajectory status of the current location point based on the historical trajectory point sequence to obtain the real-time trajectory. It should be understood that a single current location point only reflects the spatial position of the drone at a specific moment and cannot demonstrate the continuity and trend of its trajectory. However, monitoring requires understanding the complete movement path of the drone to determine its flight intentions. Therefore, this application further integrates and updates the trajectory status of newly added current location points by combining the stored historical trajectory point sequence, thereby generating a continuous and complete real-time trajectory. This connects discrete location points into trajectory information with temporal sequence and motion logic, clearly presenting the drone's flight path, speed changes, and heading adjustments. This provides complete motion data support for subsequent multi-target separation and abnormal behavior identification, ensuring the comprehensiveness of monitoring judgments.

[0032] Specifically, in one possible embodiment, step S33 is implemented as follows: First, the stored historical trajectory point sequence is retrieved from the data cache. This sequence contains information such as position, velocity, and timestamp at previous times. Then, the newly calculated current position point is concatenated with the historical trajectory point sequence in timestamp order. Subsequently, based on the spatial distance and time interval between the current position point and adjacent historical position points, the instantaneous velocity and heading angle of the UAV are calculated. Finally, the updated position, velocity, heading angle, and corresponding timestamps are integrated into a new trajectory status node, replacing the outdated historical data to form an updated real-time trajectory, ensuring the timeliness and continuity of the trajectory information.

[0033] Specifically, step S34 involves performing time-series trajectory prediction on the real-time trajectory to obtain a predicted trajectory. It should be understood that due to the high speed and maneuverability of drones, relying solely on real-time trajectories for monitoring can easily lead to response delays and an inability to promptly avoid sudden flight risks. Furthermore, determining trajectory anomalies requires early detection of deviation trends. Therefore, this application further analyzes historical motion data from the real-time trajectory and uses a time-series prediction model to mine motion patterns, thereby predicting the drone's flight path over a future period. This allows for the early acquisition of potential drone flight trajectories, timely detection of trajectory deviation risks, and provides regulatory authorities with sufficient emergency response time, effectively improving response efficiency to sudden violations and trajectory changes, and ensuring the safety and controllability of low-altitude airspace.

[0034] Specifically, in one possible embodiment, step S34 is implemented as follows: First, historical trajectory data from the recent period is extracted from the real-time trajectory, such as historical 10-second trajectory data, including feature information such as position, speed, and heading angle at each moment, and standardized to eliminate the influence of data dimensions. Then, the processed historical trajectory data is input into a pre-trained LSTM time-series prediction model, which learns historical motion patterns to capture the UAV's flight mode. Subsequently, based on the input historical features, the model outputs the predicted position within the next 3 seconds (prediction error ≤ 0.5 meters). Finally, these predicted positions are concatenated in chronological order to form a complete predicted trajectory, and the prediction confidence level is labeled to detect trajectory deviation risks in advance, providing a clear reference benchmark for subsequent anomaly detection.

[0035] In the aforementioned low-altitude UAV trajectory tracking and monitoring method based on a 5G-A integrated sensing base station, step S4 involves performing multi-target identification and trajectory separation on the real-time trajectory based on a clean fused data stream to obtain independent individual trajectories. It should be understood that when multiple UAVs fly simultaneously in low-altitude scenarios, real-time trajectories are easily confused due to signal aliasing. Existing technologies lack effective individual differentiation mechanisms, resulting in the inability to accurately track the flight path of each UAV. Therefore, this application further utilizes the unique features in the clean fused data stream, combined with multi-target identification algorithms and trajectory separation strategies, to achieve precise binding between individuals and trajectories. This overcomes the trajectory confusion bottleneck in multi-target scenarios, establishing an independent trajectory file for each UAV (including unregistered models), ensuring that regulatory authorities can clearly grasp the flight intentions of individual targets, and providing accurate individual data support for subsequent anomaly detection and emergency response.

[0036] In particular, in one specific embodiment, Figure 5 This is a flowchart of sub-step S4 of the low-altitude UAV trajectory tracking and monitoring method based on a 5G-A integrated sensing base station according to an embodiment of this application. Figure 5 As shown, step S4 includes: S41, extracting features from each frame of clean fused data in the clean fused data stream to obtain a set of individual signal feature vectors, wherein the individual signal feature vectors include communication features, radar features, and kinematic features; S42, based on the UAV individual feature library, performing target recognition on each individual signal feature vector in the set of individual signal feature vectors to obtain identified target data and a set of unidentified target feature vectors; S43, performing trajectory clustering and association assignment on the set of unidentified target feature vectors and the mixed trajectory points of unidentified targets corresponding to the set of unidentified target feature vectors to obtain independent individual trajectories.

[0037] Specifically, in step S41, features are extracted from each frame of clean fused data in the clean fused data stream to obtain a set of individual signal feature vectors. These individual signal feature vectors include communication features, radar features, and kinematic features. It should be understood that single-dimensional data features are insufficient to uniquely identify individual UAVs and are easily distorted by environmental interference, failing to support accurate differentiation in multi-target scenarios. Therefore, this application further extracts exclusive features from the three dimensions of communication, radar, and kinematics in the clean fused data, and constructs a set of individual signal feature vectors after standardization. This integrates the complementary advantages of features from different dimensions, forming unique and stable individual feature identifiers, effectively resisting the impact of environmental interference on feature recognition, providing a highly recognizable feature foundation for subsequent target recognition and trajectory separation, and ensuring the accuracy of multi-target differentiation.

[0038] Specifically, in one possible embodiment, step S41 is implemented as follows: For each frame of clean fused data, communication features such as signal modulation mode and channel quality indication are first extracted from the 5G-A communication signal through decoding. Next, radar features such as the mean and variance of the radar cross-section are calculated through radar echo signal analysis. Combined with the changes in adjacent positions in the real-time trajectory, kinematic features such as flight speed variation and turning angular velocity are calculated. Subsequently, the three types of features are normalized to eliminate dimensional differences, and then concatenated and integrated according to a preset structure to form standardized individual signal feature vectors, ultimately constructing a complete feature vector set.

[0039] Specifically, in step S42, based on the UAV individual feature library, target identification is performed on each individual signal feature vector in the individual signal feature vector set to obtain identified target data and an unidentified target feature vector set. It should be understood that since the individual signal feature vector set contains feature information of both registered and unregistered UAVs, failure to distinguish between them would lead to the inability to confirm the identity of legitimate targets and affect the efficiency of subsequent processing of unregistered targets. Therefore, this application further uses a pre-set UAV individual feature library to match and identify feature vectors by calculating the similarity between the real-time captured signal features and the known target features in the feature library, thereby classifying known and unknown targets. This allows for the rapid confirmation of the legitimate identity of registered UAVs, establishing a correspondence between identity and trajectory, and accurately separating the feature vectors of unregistered targets, providing a clear data foundation for subsequent clustering and separation, and improving the targeting and efficiency of multi-target identification.

[0040] In particular, the aforementioned UAV target recognition mechanisms have a fundamental weakness in calculating the similarity between real-time captured signal features and known target features in the feature library. These mechanisms typically employ standard cosine similarity or Euclidean distance algorithms, which implicitly assume equal weighting of features, treating all dimensions in the feature vector equally. However, this assumption does not hold true in complex and variable low-altitude environments. The discriminative power and stability of UAV signal features are not static but closely related to specific external environments. For example, in rainy weather, the measurement of radar cross-section fluctuates drastically, significantly reducing its reliability as a recognition criterion; similarly, features demodulated from communication signals in areas with complex electromagnetic environments may be distorted. Existing mechanisms lack consideration for this dynamic relationship between features and the environment, failing to adaptively adjust the importance of different feature dimensions according to the real-time scenario. This makes their similarity calculation results highly susceptible to interference from unstable feature dimensions, especially under non-ideal conditions such as severe weather or strong electromagnetic interference, significantly reducing the accuracy and robustness of recognition. Meanwhile, the method of globally comparing real-time feature vectors with the feature database incurs a huge computational burden as the number of targets and the database size increase, making it difficult to meet the high timeliness requirements of low-altitude surveillance. To address the above issues, in a preferred embodiment of this application, a scene-adaptive dynamic feature weighted recognition method is proposed. This method abandons the static model of equal feature weights, dynamically generates weights by evaluating the stability of the environment and the features themselves in real time, and uses these weights to optimize the similarity calculation process. Simultaneously, a hierarchical screening strategy is introduced to improve computational efficiency.

[0041] In particular, in one specific embodiment, Figure 6 This is a flowchart of sub-step S42 of the low-altitude UAV trajectory tracking and monitoring method based on a 5G-A integrated sensing base station according to an embodiment of this application. Figure 6 As shown, step S42 includes: S421, performing real-time environmental perception and feature stability assessment based on environmental interference data and the sequence of individual signal feature vectors of the same target to obtain a dynamic weight vector; S422, performing dynamic weighted similarity calculation and matching on individual signal feature vectors based on the dynamic weight vector and the UAV individual feature library to obtain a weighted similarity score; S423, based on the weighted similarity score, performing candidate set screening and data separation on the individual signal feature vector set based on hierarchical quantization to obtain the identified target data and the unidentified target feature vector set.

[0042] More specifically, step S421 involves real-time environmental perception and feature stability assessment based on environmental interference data and the sequence of individual signal feature vectors of the same target to obtain a dynamic weight vector. It should be understood that to achieve dynamic weighting of features, it is necessary to obtain the basis for determining the weight magnitude. This basis comes from the real-time impact of the external environment and the historical performance of the feature data itself. Specifically, firstly, real-time environmental interference data is collected, and the historical feature stream of the target within the most recent time window is retrieved. By analyzing this data, on the one hand, the influence factor of environmental factors on the stability of each feature dimension is calculated; on the other hand, the inherent stability is assessed by statistically analyzing the dispersion of each feature dimension in the historical data. Subsequently, these two assessment results are fused to generate a dynamic weight vector that reflects the credibility of each feature dimension in the current scene. Its calculation formula is as follows:

[0043] in, This represents the final generated dynamic weight vector, whose dimension is the same as the feature vector. It is a vector of environmental impact factors calculated based on the real-time environment; It is the feature intrinsic stability vector obtained by calculating the reciprocal of the normalized standard deviation of each dimension in the historical feature stream; It is a hyperparameter used to balance the weights of external environmental influences and the intrinsic stability of features; It is the total dimension of the feature vector. That is, it incorporates prior knowledge from the external environment (through...) (reflected) and statistical characteristics derived from the data itself (through) This approach combines [various methods] to ensure that the generated weights possess both rapid responsiveness to environmental changes and robustness based on historical data. In other words, it produces a dynamic weight vector that is no longer fixed but can precisely quantify the credibility of each feature dimension in the current specific scenario, laying the foundation for accurate identification in the future.

[0044] More specifically, in a particular example of this application, a specific numerical calculation embodiment is given to aid in the illustration: balance coefficient =0.6 (focusing on real-time environmental impact), Environmental Impact Factor Vector =[0.8,0.9,0.7], corresponding to communication features, radar features, and kinematic features respectively. A larger value indicates that the feature is less affected by environmental interference, representing the feature's intrinsic stability vector. =[0.75,0.85,0.65], which is derived from statistical analysis of 100 historical frames of data, representing the total feature dimension. =3. First, calculate the sum of the feature intrinsic stability vectors: =0.75+0.85+0.65=2.25, then we obtain the normalized feature stability vector: =[0.75 / 2.25,0.85 / 2.25,0.65 / 2.25]≈[0.333,0.378,0.289]. Next, calculate the weighted components: Environmental Impact Weighted Components. =0.6×[0.8,0.9,0.7]=[0.48,0.54,0.42], characteristic stability weighted component =0.4×[0.333,0.378,0.289]≈[0.133,0.151,0.116]. Finally, the dynamic weight vector is synthesized as follows: =[0.48+0.133,0.54+0.151,0.42+0.116]=[0.613,0.691,0.536]. This shows that in the current scenario, radar features have the highest weight (0.691), followed by communication features (0.613). Kinematic features are significantly affected by environmental interference (0.536). Subsequent similarity calculations will focus on weighting radar features to improve the robustness of target recognition in complex environments.

[0045] More specifically, step S422 involves dynamically weighting and matching the individual signal feature vectors based on the dynamic weight vector and the UAV individual feature library to obtain a weighted similarity score. That is, the dynamic weights generated in the previous step are applied to the similarity calculation, fundamentally changing the traditional feature weighting model. Specifically, during execution, the original individual signal feature vectors are no longer directly calculated. Features in the library Instead of calculating the similarity directly, the two vectors are first weighted using the dynamic weight vector generated in the previous step. This weighting is achieved through element-wise multiplication of the vectors (Hadamard product), and then the cosine similarity is calculated based on the weighted vectors. The formula is as follows:

[0046] in, The weighted similarity score; It is a feature vector of an individual signal captured in real time; It is a stock feature vector in the individual feature library of drones; It represents the Hadamardi (or Hadama) stack; Represents the dot product of vectors; This represents the Euclidean norm of the vector. In other words, through weighting operations, the contribution of feature dimensions deemed highly reliable in the current scenario is amplified in similarity calculation, while the influence of low-reliability feature dimensions is effectively suppressed. This allows the similarity calculation process to intelligently focus on the most critical and reliable feature information, thereby achieving accurate and robust matching of drone identities in complex dynamic scenarios with noise and interference.

[0047] More specifically, in a particular example of this application, a specific numerical calculation embodiment is given to aid in the illustration: the currently extracted individual signal feature vector =[0.7,0.8,0.6] (Communication features: modulation matching degree; Radar features: RCS mean normalized value; Kinematic features: similarity of velocity change patterns), the inventory feature vector in the UAV individual feature library. =[0.65,0.78,0.55] (pre-stored registered drone features), dynamic weight vector =[0.613,0.691,0.536] (using the above calculation results). First, calculate the Hadamard product: , =[0.613×0.65,0.691×0.78,0.536×0.55]≈[0.398,0.539,0.295]. Next, calculate the vector dot product: (0.429×0.398)+(0.553×0.539)+(0.322×0.295)≈0.171+0.298+0.095=0.564. Then calculate the Euclidean norm: = ≈ ≈0.770, = ≈ ≈0.732. Finally, the weighted similarity score was calculated: =0.564 / (0.770×0.732)≈0.564 / 0.564=1.0. Therefore, a similarity score close to 1.0 indicates that the detected drone highly matches the registered drones in the feature database, and the system can accurately determine it as a legitimate target, avoiding misidentification caused by environmental interference.

[0048] More specifically, step S423, based on weighted similarity scores, uses hierarchical quantization to filter and separate candidate sets of individual signal feature vectors to obtain identified target data and unidentified target feature vector sets. This addresses the high computational complexity of global traversal search and improves recognition efficiency. Specifically, the execution process employs a coarse-to-fine strategy. First, an offline vector quantization preprocessing is performed on the massive UAV individual feature database, clustering it into multiple clusters represented by prototype features. When a real-time individual signal feature vector enters the system, it is quickly compared with these computationally inexpensive prototype features, rapidly identifying the few most likely matching candidate clusters. Then, the system uses the computationally intensive dynamic weighted similarity algorithm defined in the previous step for fine-grained matching only within these significantly reduced candidate sets. Finally, the target identity is determined and data separation is performed based on whether the highest weighted similarity score obtained from the fine-grained matching exceeds a preset threshold. This step, by introducing hierarchical quantitative screening, significantly reduces the number of unnecessary fine comparisons, thereby greatly reducing computational latency while ensuring identification accuracy. This enables the entire identification mechanism to meet the stringent timeliness requirements of the low-altitude traffic management field and ensures the system's real-time response capability.

[0049] Through the implementation of the aforementioned technical means, this improved mechanism significantly optimizes traditional target recognition methods, achieving a comprehensive technical objective of enhancing recognition accuracy, robustness, and real-time performance. Specifically, the introduced scene-adaptive dynamic feature weighted similarity calculation method enables the system to intelligently evaluate and utilize the most reliable signal features in specific environments. This effectively overcomes the problem of decreased recognition rate caused by feature instability in complex scenarios such as severe weather and strong electromagnetic interference, thus greatly enhancing recognition accuracy and environmental adaptability. Simultaneously, the candidate set screening strategy based on hierarchical quantization optimizes the large-scale global search into an efficient coarse-screening + fine-screening mode, significantly reducing the computational complexity of the algorithm and ensuring that even in scenarios with multiple targets and large-capacity feature libraries, a second-level rapid response to drone identity verification can be achieved. Ultimately, this mechanism constructs a precise and efficient drone identification system, providing more reliable and timely technical support for low-altitude safety supervision and effectively addressing the safety challenges posed by unauthorized and reckless drone flights.

[0050] Specifically, step S43 involves clustering and associating the unidentified target feature vector set and the mixed trajectory points of the unidentified targets corresponding to the unidentified target feature vector set to obtain independent individual trajectories. It should be understood that because unidentified targets (such as unregistered drones) lack registered identity identifiers, their corresponding mixed trajectory points cannot be directly attributed to specific individuals, leading to a blind spot in the trajectory tracking of such targets. Therefore, this application further employs a density-based clustering algorithm to cluster the unidentified target feature vectors and combines this with the spatial and temporal information of the mixed trajectory points for association allocation. In this way, the feature clustering patterns of different individuals can be mined from the unidentified mixed data, accurately assigning the mixed trajectory points to the corresponding clusters, generating an independent trajectory for each unidentified target, achieving effective tracking of unregistered targets such as unregistered drones, and filling regulatory blind spots.

[0051] Specifically, in one possible embodiment, step S43 is implemented as follows: First, the set of unidentified target feature vectors is used as input. A density-based DBSCAN clustering algorithm is employed to cluster the sensory signal features of multiple drones. Clusters are automatically formed based on feature similarity, with each cluster corresponding to a potential unauthorized drone target. Then, mixed trajectory points corresponding to the unidentified target feature vectors are extracted and matched with the cluster features based on the spatiotemporal correlation of the trajectory points. Subsequently, a unique temporary identifier is assigned to each cluster, and the successfully matched trajectory points are bound to the temporary identifier. Finally, cross-validation using multi-base station positioning data corrects trajectory point attribution errors, integrating them to form a complete and independent individual trajectory for each unidentified target, thus solving the problem of tracking multiple drones simultaneously and achieving accurate tracking of unauthorized drones.

[0052] In the aforementioned low-altitude UAV trajectory tracking and monitoring method based on a 5G-A integrated sensing base station, step S5 involves identifying abnormal trajectories of individual trajectories based on predicted trajectories and an airspace rule base to obtain abnormal events. It should be understood that because individual trajectories only present the actual flight path of the UAV without considering airspace usage regulations and future flight trends, risks such as airspace violations, sudden trajectory changes, or equipment malfunctions cannot be detected in a timely manner, making it difficult to meet real-time monitoring needs. Therefore, this application further utilizes the compliance standards of the airspace rule base and the forward-looking nature of predicted trajectories to conduct multi-dimensional anomaly verification of individual trajectories, thereby comprehensively capturing various violations and abnormal scenarios. This enables a shift from passive recording to proactive early warning, accurately identifying multiple anomalies such as airspace violations, trajectory changes, and equipment disconnection, providing regulatory authorities with complete anomaly information, and ensuring the timeliness and targeted nature of emergency response.

[0053] Specifically, in one possible embodiment, step S5 is implemented as follows: First, compliance parameters such as no-fly zones and restricted flight altitudes are retrieved from the airspace rule base, along with the corresponding predicted trajectory data of the UAV. Then, the real-time position, speed, and heading information of the individual trajectory are compared with the airspace rule base for compliance, and deviation analysis is performed with the predicted trajectory. Subsequently, anomalies are determined according to preset standards. For example, a sudden change in flight speed (increase / decrease ≥ 5 m / s), a sudden change in heading angle (≥ 30° / s), or a deviation from the predicted trajectory ≥ 2 meters is considered a trajectory anomaly. A communication signal interruption exceeding a certain time and the absence of matching radar signals is considered an equipment anomaly. Finally, for confirmed anomalies, key information such as UAV identification, anomaly occurrence time, location, and anomaly type is integrated to generate a structured anomaly event, ensuring the anomaly information is complete and can directly support subsequent emergency response work.

[0054] In summary, the low-altitude UAV trajectory tracking and monitoring method based on a 5G-A integrated sensing base station, as described in this application, is explained. First, it collects multi-source data including sensing fusion signals, environmental interference, and UAV attributes, which are then encapsulated into synchronized data frames after being aligned with a unified timestamp and coordinate system. Subsequently, the data frames undergo deep fusion and anti-interference processing to filter out noise and generate clean fused data. Then, using multi-base station collaborative calculation and prediction, real-time trajectory calculation and motion trend prediction are performed on the clean data. Based on this, through multi-target feature recognition and clustering separation mechanisms, independent individual trajectories are accurately extracted from the complex mixed data stream, and the trajectories are validated for compliance and anomaly determination using a spatial rule base. This effectively overcomes the challenges of signal interference and multi-target aliasing in complex environments, thereby achieving high-precision full-domain tracking and real-time monitoring of abnormal behavior of low-altitude UAVs.

[0055] As described above, the low-altitude UAV trajectory tracking and monitoring system 100 based on a 5G-A integrated sensing base station according to embodiments of this application can be implemented in various wireless terminals, such as servers with a low-altitude UAV trajectory tracking and monitoring algorithm based on a 5G-A integrated sensing base station. In one possible implementation, the low-altitude UAV trajectory tracking and monitoring system 100 based on a 5G-A integrated sensing base station according to embodiments of this application can be integrated into the wireless terminal as a software module and / or hardware module. For example, the low-altitude UAV trajectory tracking and monitoring system 100 based on a 5G-A integrated sensing base station can be a software module in the operating system of the wireless terminal, or it can be an application developed for the wireless terminal; of course, the low-altitude UAV trajectory tracking and monitoring system 100 based on a 5G-A integrated sensing base station can also be one of many hardware modules of the wireless terminal.

[0056] Alternatively, in another example, the low-altitude drone trajectory tracking and monitoring system 100 based on the 5G-A integrated sensing base station and the wireless terminal can also be separate devices, and the low-altitude drone trajectory tracking and monitoring system 100 based on the 5G-A integrated sensing base station can be connected to the wireless terminal via wired and / or wireless networks, and transmit interactive information in accordance with the agreed data format.

Claims

1. A low-altitude unmanned aerial vehicle track tracking and monitoring method based on a 5G-A-sensing integrated base station, characterized in that, The method comprises the following steps: data alignment, coordinate unification and packaging are performed on the common sense fusion signal, the environmental interference data set and the unmanned aerial vehicle basic attribute data to obtain a synchronized original data frame; pure fusion data is obtained by performing common sense data fusion and anti-interference processing on the synchronized original data frame; real-time trajectory and predicted trajectory are obtained by performing trajectory calculation and prediction on the pure fusion data provided by multiple base stations; independent individual trajectories are obtained by performing multi-target identification and trajectory separation on the real-time trajectory based on the pure fusion data stream; abnormal events are obtained by performing abnormal trajectory identification on the independent individual trajectories based on the predicted trajectory and the airspace rule library. 2.The method of claim 1, wherein, The common sense data fusion and anti-interference processing on the synchronized original data frame to obtain pure fusion data comprises: cross-domain data association matching is performed on the synchronized original data frame to obtain associated data objects; fusion filtering and state estimation are performed on the associated data objects to obtain pure fusion data. 3.The method of claim 2, wherein, The pure fusion data comprises identity, optimal state estimation and associated attributes, the optimal state estimation comprises estimated position, estimated speed and covariance matrix, and the associated attributes comprise model information, registration information and data source identification.

4. The low-altitude unmanned aerial vehicle trajectory tracking monitoring method based on the 5G-A common sensing integrated base station according to claim 1, characterized in that, The trajectory calculation and prediction on the pure fusion data provided by multiple base stations to obtain real-time trajectory and predicted trajectory comprises: the pure fusion data from the current base station is set as local pure fusion data, and the pure fusion data from at least two adjacent cooperative base stations is set as cooperative multi-source positioning data; multi-station cooperative positioning solution is performed on the local pure fusion data and the cooperative multi-source positioning data to obtain a current position point; real-time trajectory state updating is performed on the current position point based on a historical trajectory point sequence to obtain a real-time trajectory; time sequence trajectory prediction is performed on the real-time trajectory to obtain a predicted trajectory. 5.The method of claim 1, wherein, The multi-target identification and trajectory separation on the real-time trajectory based on the pure fusion data stream to obtain independent individual trajectories comprises: feature extraction is performed on each frame of pure fusion data in the pure fusion data stream to obtain a set of individual signal feature vectors, the individual signal feature vector comprises communication feature, radar feature and kinematics feature; target identification is performed on each individual signal feature vector in the set of individual signal feature vectors based on an unmanned aerial vehicle individual feature library to obtain identified target data and a set of unidentified target feature vectors; trajectory clustering and association assignment are performed on the set of unidentified target feature vectors and the mixed trajectory points of the unidentified targets corresponding to the set of unidentified target feature vectors to obtain independent individual trajectories. 6.The method of claim 5, wherein the 5G-A based low-altitude UAV trajectory tracking and monitoring method is characterized in that, The target identification performed on each individual signal feature vector in the set of individual signal feature vectors based on the unmanned aerial vehicle individual feature library to obtain identified target data and a set of unidentified target feature vectors comprises: real-time environmental perception and feature stability evaluation are performed based on the environmental interference data and the sequence of individual signal feature vectors of the same target to obtain a dynamic weight vector; dynamic weighted similarity calculation and matching are performed on the individual signal feature vector based on the dynamic weight vector and the unmanned aerial vehicle individual feature library to obtain a weighted similarity score; Based on the weighted similarity score, the individual signal feature vector set is screened based on the hierarchical quantization candidate set and separated from the data to obtain the identified target data and the un-identified target feature vector set.

7. The low-altitude unmanned aerial vehicle trajectory tracking monitoring method based on the 5G-A common sensing integrated base station according to claim 6, characterized in that, Based on the dynamic weight vector and the individual feature library of the unmanned aerial vehicle, the individual signal feature vector is dynamically weighted similarity calculated and matched to obtain a weighted similarity score, including: the individual signal feature vector is dynamically weighted similarity calculated and matched by the following formula, the formula is: wherein, is a weighted similarity score; is an individual signal feature vector; is one of the library feature vectors in the drone individual feature library; denotes a Hadamard product; denotes a vector dot product; denotes the Euclidean norm of a vector, is a dynamic weight vector.

Images