Unmanned aerial vehicle target perception method based on beam reconstruction and environment perception adaptation

Abstract

The application discloses a UAV target sensing method based on beam reconstruction and environment sensing adaptation, and relates to the field of UAV target sensing.The application realizes dynamic adjustment of subarray beam weight and nonlinear optimization of power distribution through a beam reconstruction and environment sensing adaptation technology, so that the UAV can continuously maintain a target receiving signal with a high signal-to-noise ratio under a complex external environment and diversified task state, the beam pointing and gain distribution are optimized, and the beam is pre-offset compensated through target motion trend prediction, so that the target detection probability, tracking accuracy and signal stability are significantly improved, continuous and reliable sensing of a high-maneuvering target is realized, and the energy utilization efficiency is taken into account.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) target perception, specifically a UAV target perception method based on beam reconstruction and adaptive environmental perception. Background Technology

[0002] Current UAV target perception technologies primarily rely on fixed beams, preset beam scanning, or static adjustment strategies based on certain environmental parameters. In fixed-beam schemes, the UAV scans the target area using a fixed beam direction, suitable for scenarios where the target is relatively stationary or the environment changes little. However, when the target is maneuvering or the environment changes drastically, unstable signal reception and reduced detection probability are common problems. Some schemes employ beam scanning strategies, periodically scanning different directions to cover the target area. While this can acquire target information to some extent, the scanning cycle is long, failing to meet the real-time tracking requirements of highly dynamic targets. Furthermore, uneven energy distribution can lead to wasted power in some subarrays.

[0003] In recent years, some studies have begun to introduce methods for adjusting beams using environmental parameters. For example, beam angles can be corrected using flight altitude, meteorological conditions (such as wind speed, air pressure, and humidity), and visibility to adapt to complex external environments. However, these methods typically consider only one or a few environmental parameters, failing to adequately account for the impact of mission status information such as mission priority, number of targets, and mission coverage area on beam control. Furthermore, these methods lack a complete closed-loop adjustment mechanism for subarray weight initialization, power allocation, and target motion prediction. Beam adjustment often relies on historical experience or simple rules, making it impossible to guarantee stable signal-to-noise ratio and tracking accuracy in multi-target, highly maneuverable, and complex environments. Summary of the Invention

[0004] This invention proposes a UAV target perception method based on beam reconstruction and adaptive environmental perception. Through the construction of joint environment-task feature vectors, dynamic reconstruction of subarray weights, nonlinear modeling of power allocation, and prediction of target motion trends, real-time adaptive beam adjustment is achieved. Compared with existing technologies, this invention can maintain beam stability under different environmental conditions and mission states, improve target detection probability and tracking accuracy, while simultaneously optimizing energy utilization and maintaining the signal-to-noise ratio.

[0005] The UAV target perception method based on beam reconstruction and adaptive environmental perception includes the following steps:

[0006] S1. Obtain the external environment parameters and mission status information of the current UAV, and establish a joint feature vector of environment-mission;

[0007] Specifically, an environment-task joint feature vector is established by acquiring the external environmental parameters and mission status information of the UAV. This vector unifies environmental constraints and mission priorities into a high-dimensional vector for subsequent subarray beam weight optimization. Environmental parameters include weather conditions, flight altitude, and visibility, reflecting the external limitations on the UAV's detection capabilities. Mission status information includes mission type, number of targets, priority, and mission coverage area, reflecting the intensity and regional distribution of mission execution requirements. By quantifying and weighting these parameters, the resulting joint feature vector can express the environmental constraints of target detection and tracking, thereby reflecting the priority of mission execution and ensuring that subarray beam adjustment in complex environments can balance target detectability and mission requirements.

[0008] S2. Based on the joint feature vector of environment and task, the subarray beam weights are dynamically reconstructed. The weight vectors are adjusted in real time through a continuously variable weight update algorithm. Combined with the minimum mean square error criterion and constraint optimization method, the environmental features are mapped to the subarray weight vector. The target signal-to-noise ratio is calculated based on the output signal of the new subarray beam formed by the subarray weight vector.

[0009] Specifically, the subarray beam weights are dynamically reconstructed based on the joint environment-mission feature vector. A continuously variable weight update algorithm iteratively adjusts the weight vector, and the environmental features are mapped to the subarray weight vector using the minimum mean square error criterion and constraint optimization methods. The minimum mean square error criterion ensures that the weighted synthesized output signal is closest to the ideal target signal, thereby maximizing the target signal-to-noise ratio (SNR) and suppressing interference noise. Constraint optimization guarantees that the subarray weights meet physical and power limitations. Through this mapping process, the subarray can dynamically adjust its amplitude and phase under different environmental and mission conditions, achieving continuous adjustability of beam direction and gain, thus ensuring stable received signal quality and accurately calculating the target SNR.

[0010] S3. Based on the target signal-to-noise ratio, combined with external environmental parameters and task status information, perform nonlinear modeling of the power allocation function; by constructing power allocation curves, perform differentiated configuration of the output power of the left and right subarrays, and generate differentiated power allocation results for the left and right subarrays;

[0011] Specifically, by combining the target signal-to-noise ratio (SNR) with external environmental parameters and mission status information, a nonlinear model of the power allocation function is performed. The principle is to use the SNR to reflect target detectability, environmental parameters to reflect transmission loss and interference levels, and mission status to reflect mission priority. These factors are used as inputs to the power allocation function, generating a power allocation curve using a weighted exponential function. This curve describes the optimal output power ratio of the left and right subarrays under the current environmental and mission status, ensuring that the energy distribution satisfies both signal reception requirements and energy efficiency.

[0012] S4. Based on the subarray weight vector and the differential power allocation results, construct a dynamic beamforming matrix, and process the target echo signal according to the constructed dynamic beamforming matrix, and calculate the detection probability, tracking accuracy and signal-to-noise ratio preservation.

[0013] Specifically, a dynamic beamforming matrix is ​​constructed using subarray weight vectors and differentiated power allocation results. This matrix is ​​then used to weight and filter the target echo signal, achieving signal enhancement and interference suppression. The dynamic beamforming matrix focuses the received signal towards the target direction while simultaneously suppressing noise and sidelobe interference by weighting and modulating the amplitude and phase of each subarray. This optimizes detection probability, tracking accuracy, and signal-to-noise ratio. This step translates the previous weight optimization and power allocation results into actual signal processing performance.

[0014] S5. Based on the processed target echo signal and the calculated target signal-to-noise ratio, predict the target's position at the next moment using the target motion trend prediction model; based on the predicted position, pre-offset the subarray weight vector of the dynamic beamforming matrix to pre-offset the beam pointing to the target's future position.

[0015] Specifically, the target's position at the next moment is estimated through a target motion trend prediction model. The principle is to use the processed echo signal, the calculated target signal-to-noise ratio, detection probability, tracking accuracy, and signal-to-noise ratio preservation as constraints and weights to construct a prediction model that includes target presence, motion error, and signal quality. Recursive filtering or extended Kalman filtering methods are typically used for state updates and trajectory prediction. Based on the predicted position, it is mapped to the subarray weight vector of the dynamic beamforming matrix, and pre-offset adjustments are made to ensure the beam is pointed in advance to the target's future position, enabling continuous tracking and signal acquisition of highly maneuverable targets and improving target detection stability and tracking accuracy.

[0016] The beneficial effects of the invention are:

[0017] This invention achieves dynamic adjustment of subarray beam weights and nonlinear optimization of power allocation through beam reconstruction and environmental perception adaptive technology. This enables UAVs to maintain a high signal-to-noise ratio target reception signal in complex external environments and diverse mission states, optimizes beam pointing and gain distribution, and achieves beam pre-offset compensation through target motion trend prediction. This significantly improves target detection probability, tracking accuracy and signal stability, enabling continuous and reliable perception of highly maneuverable targets while taking into account energy utilization efficiency. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the UAV target perception method based on beam reconstruction and environmental perception adaptation according to an embodiment of the present invention. Detailed Implementation

[0019] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings, but the scope of protection of the present invention is not limited to the following description.

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention; that is, the described embodiments are only a part of the embodiments of the invention, and not all of them. The components of the embodiments of the invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0021] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention. It should be noted that relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations.

[0022] Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or machine that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or machine. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or machine that includes said element.

[0023] The features and performance of the present invention will be further described in detail below with reference to embodiments.

[0024] Example 1

[0025] Among them, such as Figure 1 A UAV target perception method based on beam reconstruction and adaptive environmental perception includes the following steps:

[0026] S1. Obtain the external environment parameters and mission status information of the current UAV, and establish a joint feature vector of environment-mission;

[0027] S2. Based on the joint feature vector of environment and task, the subarray beam weights are dynamically reconstructed. The weight vectors are adjusted in real time through a continuously variable weight update algorithm. Combined with the minimum mean square error criterion and constraint optimization method, the environmental features are mapped to the subarray weight vector. The target signal-to-noise ratio is calculated based on the output signal of the new subarray beam formed by the subarray weight vector.

[0028] S3. Based on the target signal-to-noise ratio, combined with external environmental parameters and task status information, perform nonlinear modeling of the power allocation function; by constructing power allocation curves, perform differentiated configuration of the output power of the left and right subarrays, and generate differentiated power allocation results for the left and right subarrays;

[0029] S4. Based on the subarray weight vector and the differential power allocation results, construct a dynamic beamforming matrix, and process the target echo signal according to the constructed dynamic beamforming matrix, and calculate the detection probability, tracking accuracy and signal-to-noise ratio preservation.

[0030] S5. Based on the processed target echo signal and the calculated target signal-to-noise ratio, predict the target's position at the next moment using the target motion trend prediction model; based on the predicted position, pre-offset the subarray weight vector of the dynamic beamforming matrix to pre-offset the beam pointing to the target's future position.

[0031] Specifically, the above embodiments, based on beam reconstruction and adaptive environmental perception technology, achieve continuous and variable adjustment of subarray beam weights and environmental constraint mapping, enabling beam synthesis to be dynamically optimized according to real-time external environmental parameters and mission status information. Combined with a nonlinear power allocation function, the output power of the left and right subarrays is configured differently to ensure optimal reception of target signals under different flight altitudes, visibility, and mission coverage conditions. At the same time, a target motion trend prediction model is used to pre-offset compensate the beam pointing, so that the beam points to the future position of the target in advance. This improves the target detection probability and tracking accuracy while maintaining a stable signal-to-noise ratio, enhances the UAV's continuous perception capability of multiple targets in complex and variable environments, optimizes energy utilization efficiency and array thermal balance, and achieves highly reliable and high-precision adaptive target perception.

[0032] Furthermore, the external environmental parameters include weather conditions, flight altitude, and visibility; the mission status information includes the current mission type, number of targets, priority, and mission coverage area.

[0033] Furthermore, in step S1, the specific steps for establishing the joint feature vector of the environment and the task are as follows:

[0034] S101. Normalize the external environment parameters and task status information to obtain the external environment vector and task status feature vector;

[0035] S102. Based on task priority and coverage requirements, assign weights to task status information to obtain a weighted task status feature vector; and based on the evaluation of beam sensitivity by environmental parameters, transform the external environment vector into environmental correction weights.

[0036] S103. Perform feature fusion between the weighted task state feature vector and the environment correction weight to obtain and output the joint feature vector.

[0037] Specifically, in practice, environmental perception sensors on the UAV collect meteorological conditions (such as wind speed, temperature, and humidity), flight altitude, and visibility data. Simultaneously, information such as the current task type, task priority, and task coverage area is obtained from the task management system. Specifically, the current task type is converted into a vector form using one-hot encoding or integer mapping, allowing different task categories to be numerically represented. Task priorities are quantified, for example, high, medium, and low priorities are mapped to normalized values ​​to form one-dimensional features, reflecting task importance in feature fusion. Furthermore, the task coverage area is divided into a grid, with each grid point generating a binary value indicating whether it is within the coverage area. All grid point vectors are concatenated to form a coverage area feature vector. The task type vector, task priority vector, and coverage area vector are weighted and fused according to set weights to obtain a unified task state feature vector. Further, raw parameters from different sources are normalized to eliminate dimensional differences, forming standardized external environment vectors and task state feature vectors. In task feature processing, weights are assigned to various task status information based on the importance of the task and the size of the coverage area, forming a weighted task status feature vector. At the same time, the sensitivity of various parameters in the environmental vector to beam performance is evaluated. For example, high wind speed may affect beam pointing accuracy, and high temperature may limit output power. The evaluation results are mapped to an environmental correction weight vector. The weighted task status feature vector and the environmental correction weight vector are linearly weighted or weighted summed according to a preset fusion rule to obtain a joint feature vector that comprehensively considers task requirements and environmental constraints. This vector is directly used as the input for the dynamic reconstruction of subarray beam weights.

[0038] Furthermore, in step S103, the joint feature vector is specifically constructed through weighted fusion, as follows:

[0039] ;

[0040] Wherein, represents the joint feature vector of environment-task, and The relative weights of environmental features in the joint vector are represented by the following: The weights of task features in the joint vector are represented by the following: The environmental adjustment weights are used to adjust the influence of environmental features on the joint vector. This represents the normalized environmental feature vector, the This represents the weighted feature vector of the task.

[0041] Furthermore, step S2 specifically includes the following sub-steps:

[0042] S201. Decompose the joint feature vector to obtain task features and environment features; and generate an initial submatrix weight vector based on the task features and environment features.

[0043] S202. Using the initial subarray weight vector as the initial value, the subarray weight vector is iteratively adjusted through the weight update algorithm. In each iteration, the joint feature vector of environment and task is used as a constraint condition, and the minimum mean square error is used as the optimization objective to adjust the amplitude and phase distribution of the subarray weights, thus mapping the environmental features to the subarray weight vector.

[0044] S203. Based on the new subarray beam formed by the subarray weight vector, the received signal is weighted and synthesized, and the energy-to-noise power ratio of the synthesized signal is calculated to obtain the target signal-to-noise ratio.

[0045] Specifically, the joint environment-task feature vector is decomposed into task feature sub-vectors and environment feature sub-vectors. The task feature sub-vectors are derived from the UAV mission scheduling system, including the current mission type, number of targets, target priority, and mission coverage area. These are read in real time from the mission scheduling database or flight control system, normalized, and combined with priority allocation weights to form a task weight vector, quantifying the importance of each target in beam allocation. The environment feature sub-vectors are derived from data collected by the UAV's onboard sensors, including meteorological conditions (wind speed, wind direction, temperature, humidity) and flight altitude (from barometric altimeter or GPS). The system uses altitude information and visibility (measured by optical or infrared sensors) and calculates constraint coefficients for beam amplitude and phase using an environmental sensitivity evaluation function to generate an environmental correction weight vector, ensuring beam adjustment remains within physically operable limits. The task weight vector and the environmental correction weight vector are then weighted and fused to obtain an initial subarray weight vector, ensuring the initial beam simultaneously considers high-priority target coverage areas and environmental constraints. Using this initial subarray weight vector as the initial value for iteration, a continuously variable weight update algorithm is employed to progressively adjust the amplitude and phase. Each iteration uses the joint characteristic vector of the environment and task as constraints and the minimum mean square error criterion as the optimization objective to adjust the subarray weight distribution, minimizing the error between the subarray output signal and the ideal beam pointing, ensuring beam stability during task switching and environmental changes. The process involves a smooth transition; generating a new subarray beam using the final subarray weight vector obtained through iteration; weighting and synthesizing the received target echo signal; and calculating the signal energy to noise power ratio to obtain the target signal-to-noise ratio. For example, the subarray weight vector obtained through the dynamic reconstruction algorithm is used as the weighting coefficient, and the transmitted or received signal of each subarray is multiplied by the corresponding weight to form a weighted signal sequence. All weighted subarray signals are superimposed in the time or spatial domain to obtain the synthesized total received signal, which enhances the signal energy in the target direction. The synthesized signal energy is statistically calculated, typically by summing the squares to obtain the signal power. Simultaneously, the received noise is independently measured or estimated to calculate the noise power. The ratio of the synthesized signal power to the noise power is then calculated to obtain the target signal-to-noise ratio.

[0046] In addition, the environmental sensitivity evaluation function is a function used in UAV target perception systems to quantify the impact of environmental parameters on beam performance. It maps environmental factors such as meteorological conditions (e.g., wind speed, humidity, precipitation), flight altitude, and visibility into a numerical index, representing the sensitivity of a subarray beam or the entire array to changes in the external environment under the current conditions. A higher value indicates a more significant impact of the environment on beam performance, requiring stronger adaptive adjustment; a lower value indicates a more robust beam to environmental changes. The construction principle is as follows: based on the impact of each environmental parameter on signal propagation, target detection, and tracking performance, a weighting function or response curve is defined to integrate multi-dimensional environmental information into a single sensitivity index. The calculation method is usually obtained by fitting empirical models, simulation models, or historical observation data. For example, the decrease in target signal-to-noise ratio, beam gain loss, or change in interference gain under different meteorological conditions is mapped to weight values, which are then used as constraints or adjustment factors during feature fusion or beam weight reconstruction.

[0047] Furthermore, in step S202, the minimum mean square error as the optimization objective is specifically expressed as follows:

[0048] ;

[0049] Among them, the The minimum mean squared error loss function is denoted as . The submatrix weight vector represents the joint feature vector of the environment and the task. The function represents the expected value. The target echo signal is represented by the expected value, and the subarray weight vector is represented by the conjugate transpose. Represents the received signal vector, the This represents the weight constraint function.

[0050] Specifically, in beamforming, the subarray weight vector The goal is to maximize the target signal reception while suppressing noise and interference. To quantify this goal, a minimum mean square error cost function is introduced, where the mean square error is expressed as: The Indicates the desired signal, the Represents a constant.

[0051] Furthermore, step S201 specifically includes the following sub-steps:

[0052] S2011. Decompose the joint feature vector into task feature sub-vectors and environment feature sub-vectors; extract target priority information based on the task feature sub-vectors, and construct a task weight vector based on the target priority information; extract environmental information based on the environment feature sub-vectors, and construct an environment-corrected weight vector based on the environmental information.

[0053] S2012. Linearly fuse the task weight vector and the environment correction weight vector to obtain the initial subarray weight vector.

[0054] Specifically, the environment-task joint feature vector is decomposed into task feature sub-vectors and environment feature sub-vectors. The task feature sub-vectors are obtained from the UAV mission scheduling system, including the current mission type, number of targets, target priority, and mission coverage area. They are acquired by reading data from the mission scheduling database or flight control system in real time, and a task weight vector is constructed based on the target priority information to quantify the relative importance of each target in beam allocation. The environment feature sub-vectors are obtained from data collected by UAV onboard sensors, including meteorological conditions (wind speed, wind direction, temperature, humidity), flight altitude (read via barometric altimeter or GPS), and visibility (measured via optical or infrared sensors). An environment-corrected weight vector is generated based on a sensitivity evaluation function to ensure that beam adjustment is constrained by environmental limitations. The task weight vector and the environment-corrected weight vector are linearly fused to obtain the initial subarray weight vector, so that the beam allocation stage takes into account both high-priority target coverage and environmental physical constraints. In step S103, the weighted task feature vector and the environment feature vector are fused to form a joint environment-task feature vector. The purpose is to encode multi-dimensional environmental and task information in a unified manner, providing a complete input basis for subsequent beam weight optimization. In step S2011, the joint feature vector is decomposed into task feature sub-vectors and environment feature sub-vectors. The purpose is to extract task priority and environmental constraint information respectively, so that task orientation and environmental adaptability can be quantified independently, facilitating the separate construction of task weight vectors and environment-corrected weight vectors. In step S2012, the task weight vector and the environment-corrected weight vector are linearly fused. The purpose is to comprehensively map task requirements and environmental constraints into an initial subarray weight vector, thereby providing directly usable initial values ​​for subarray beam dynamic control and realizing the effective conversion of information into control parameters.

[0055] Furthermore, step S3 specifically includes the following sub-steps:

[0056] S301. Combine the output signal-to-noise ratio, meteorological conditions, flight altitude, visibility from the external environment parameters, and mission type, priority, and coverage area from the mission status information into the input set for power allocation modeling;

[0057] S302. Using the input set as the independent variable, construct a power allocation function through a weighted exponential function, and generate a power allocation curve to describe the power allocation relationship between the left and right subarrays under different environments and task states based on the power allocation function;

[0058] S303. Based on the power distribution curve, calculate the output power values ​​of the left and right subarrays respectively to form differentiated power distribution results.

[0059] Specifically, the target signal-to-noise ratio obtained through dynamic adjustment of subarray beam weights, along with external environmental parameters including weather conditions, flight altitude, visibility, and mission status information such as mission type, target priority, and mission coverage area, are integrated to form the input set for power allocation modeling. This set comes from real-time data collected by airborne sensors and mission scheduling data from the flight control system. Using this input set as the independent variable, a power allocation function is established through a weighted exponential function, where the exponential weights are adjusted according to target priority and environmental constraints to quantify the power allocation priority of different subarrays under different conditions. Based on the constructed power allocation function, curves describing the power allocation relationship between the left and right subarrays under various environmental and mission states are generated. The specific output power values ​​of the left and right subarrays are calculated through the curves to achieve differentiated power allocation, allowing subarrays corresponding to high-priority targets to receive higher power, while the power of subarrays severely constrained by environmental conditions is appropriately reduced, ensuring a balance between overall energy consumption optimization and target detection performance.

[0060] Specifically, the power allocation function is expressed as follows:

[0061] ;

[0062] Among them, the Indicates the power allocation ratio, the Indicates the index weight, the This represents the input set, which combines the target signal-to-noise ratio, environmental parameters, and task information.

[0063] Finally, the output is:

[0064] ;

[0065] ;

[0066] Among them, the Indicates the lower limit of the subarray output power, the This indicates the upper limit of the subarray output power, the and This indicates the coverage adjustment amount, used to fine-tune the power of the left / right subarrays based on the mission coverage area or beam direction.

[0067] Furthermore, step S4 specifically includes the following sub-steps:

[0068] S401. Merge the subarray weight vector with the differential power allocation results to construct a set of basic parameters;

[0069] S402. Generate a dynamic beamforming matrix based on the fused set of basic parameters;

[0070] S403. The received target echo signal is weighted, synthesized, and filtered using a beamforming matrix;

[0071] S404. Calculate the detection probability, tracking accuracy, and signal-to-noise ratio retention respectively.

[0072] Specifically, the weight vector of each subarray is fused with the differentiated power allocation results of the left and right subarrays to form a set of basic parameters describing the output characteristics, power distribution, and phase information of the current subarray. The fusion process involves: obtaining the initial weight vector of each subarray, which includes amplitude and phase information; determining the power ratio coefficient of each subarray based on the calculated differentiated power allocation results of the left and right subarrays; multiplying the amplitude portion of the weight vector of each subarray by its corresponding power ratio coefficient to obtain a new amplitude value while maintaining the phase information of the original weight vector; then recombinating the adjusted amplitude with the original phase to form a new complex form of the subarray weight vector; and finally inputting the set of adjusted weight vectors of all subarrays into... The parameters are used as a set of basic parameters; a dynamic beamforming matrix is ​​constructed using this set of basic parameters, and the gain and phase distribution of each subarray in different directions are determined in matrix form to achieve precise control of beam spatial pointing; the received target echo signal is input into the dynamic beamforming matrix, and the echoes of multiple subarrays are superimposed through weighted synthesis, while filtering is performed to suppress noise and interference to obtain an enhanced target signal; based on the synthesized signal, the detection probability is calculated, and the target presence is estimated through the target signal amplitude and noise statistical characteristics; the tracking accuracy is calculated, and the tracking performance is quantified by the target position and velocity estimation error; the signal-to-noise ratio (SNR) preservation is calculated, and the ability of beamforming to preserve signal quality is evaluated by comparing the SNR changes of the signal before and after beamforming.

[0073] Furthermore, in step S404, the calculation process is as follows:

[0074] The detection probability is calculated by combining the target signal amplitude with the noise statistical characteristics.

[0075] Tracking accuracy is calculated based on the target's azimuth and velocity estimation errors;

[0076] The signal-to-noise ratio (SNR) retention is calculated by comparing the rate of change of SNR before and after beamforming.

[0077] Specifically, the above calculation process is expressed as follows:

[0078] ;

[0079] ;

[0080] ;

[0081] Among them, the Indicates the detection probability, the The received target echo signal amplitude is derived from the weighted output signal after the subarray dynamic beamforming. Represents a probability operator, i.e. The probability, the The detection threshold, determined by system design or task requirements, is used to distinguish target signals from noise; Indicating tracking accuracy, the The target azimuth angle estimate is derived from subarray output signal processing and direction estimation algorithms (such as MUSIC, MVDR, or Kalman filtering). Indicates the true azimuth of the target, the This represents the estimated target velocity. Indicates the target's true speed, the Indicating signal-to-noise ratio retention, the This represents the output signal-to-noise ratio after weighting by the subarray dynamic beamforming matrix. This indicates that the target's original signal-to-noise ratio is calculated from the received signal power and the noise power. This represents the expected function.

[0082] Specifically, the main purpose of calculating the subarray weight vector, target signal-to-noise ratio (SNR), and target echo signal is to achieve adaptive optimization and dynamic adjustment of UAV target perception. The calculation of the subarray weight vector is first used to form the subarray beam. By weighting the amplitude and phase of the received signals from different subarrays, the beam can be more precisely focused on the target direction, thereby enhancing the energy of the target echo signal while suppressing interference and noise from non-target directions. This is crucial for maintaining stable detection performance in complex environments. The calculation of the target SNR quantifies the identifiability of the target signal relative to noise under the current beam configuration. It not only reflects the effectiveness of beamforming but also provides a basis for subsequent power allocation and subarray weight adjustment, enabling the system to dynamically optimize output power under different flight altitudes, visibility, and mission conditions, balancing detection performance and energy efficiency. Processing the target echo signal and calculating the detection probability, tracking accuracy, and SNR preservation is to evaluate the overall effect of the current subarray configuration and beam direction on target detection. Detection probability represents the likelihood that the system will correctly identify the target under the current configuration; tracking accuracy reflects the error range of the system's target position estimation at continuous time intervals; and signal-to-noise ratio (SNR) preservation measures the retention of signal quality before and after beamforming. These indicators are interrelated and provide reliable quantitative information for predicting target motion trends. By using these parameters as input, the system can predict the target's position at the next moment and pre-adjust the subarray weight vector based on the prediction results, causing the beam to point towards the target's future position in advance, thus maintaining continuous perception capability even when the target is maneuvering rapidly or the environment changes. The core principle of the entire calculation process lies in using environmental and mission information to guide the dynamic adjustment of subarray weights and beam direction. By providing real-time feedback on target signal quality and motion trends, the system achieves high-precision, low-latency, and adaptive target detection and tracking capabilities for the UAV platform in complex and ever-changing environments.

[0083] Furthermore, in step S402, the dynamic beamforming matrix is ​​represented as follows:

[0084] ;

[0085] Among them, the Represents the dynamic beamforming matrix, the The submatrix weight vector is described in the following text. Indicates the output power of the left subarray, the This indicates the output power of the right subarray.

[0086] Furthermore, step S5 specifically includes the following sub-steps:

[0087] S501. Based on the processed target echo signal, signal-to-noise ratio, and calculated detection probability, tracking accuracy, and signal-to-noise ratio preservation, construct a target motion trend prediction model; wherein, the detection probability is used to constrain the target existence confidence level, the tracking accuracy is used to limit the error range of the predicted trajectory, and the signal-to-noise ratio preservation is used to correct the weight of the echo signal quality on the prediction model.

[0088] S502. Predict the target's position at the next moment by constructing a target motion trend prediction model;

[0089] S503. Map the predicted position to the subarray weight vector corresponding to the dynamic beamforming matrix, and pre-offset the subarray weight vector to pre-offset the beam to the future position of the target.

[0090] Specifically, the processed target echo signal and its corresponding signal-to-noise ratio (SNR), detection probability, tracking accuracy, and SNR preservation are used as inputs to construct a target motion trend prediction model. By establishing a mapping relationship between the target state vector and the time series, the future position of the target is estimated. In the prediction model, the detection probability is used to constrain the target existence confidence level, ensuring that prediction is only performed under high confidence conditions. Tracking accuracy is used to limit the error range of the predicted trajectory. The SNR preservation is used as a weight correction factor to correct the impact of echo signal quality on the prediction. The predicted position of the target at the next moment is calculated through this prediction model. The predicted position is mapped to the subarray weight vector corresponding to the dynamic beamforming matrix. The subarray weight vector is pre-offset adjusted according to the predicted position, so that the beam is offset in advance to point to the future position of the target, realizing synchronous tracking of beam and target motion, and improving detection and tracking accuracy.

[0091] Furthermore, in step S503, the pre-offset adjustment is specifically expressed as follows:

[0092] ;

[0093] ;

[0094] ;

[0095] Among them, the This represents the subarray weight vector after pre-offset, the The submatrix weight vector is described in the following text. Represents the predicted offset vector, the and These represent the target's predicted position estimate at the next time step and the target's position estimate at the current time step, respectively. This represents the target velocity estimate at the current moment. Indicates the time step, the This represents a mapping function used to represent the change in the predicted location of the target. Convert it into the corresponding subarray weight offset vector.

[0096] For example, the core objective of the target motion trend prediction model is to predict the target's position and motion state at the next moment based on the target echo signal received by the UAV and the calculated signal-to-noise ratio, detection probability, tracking accuracy, and signal-to-noise ratio preservation, in order to achieve beam pre-offset and dynamic tracking. This can be understood from the following aspects:

[0097] First, the target motion trend prediction model employs a state-space model to predict target position and velocity in continuous or discrete time. The model takes historical target observation data as input, including the amplitude and phase information of the received target echo signal, the currently estimated target position, velocity, and direction of motion. It also combines signal-to-noise ratio (SNR) and detection probability to quantify the confidence level of target existence, and uses tracking accuracy and SNR preservation to weightedly correct the reliability of the observation data. This allows the model to maintain reasonable predictive ability even with environmental noise or signal attenuation.

[0098] Secondly, the prediction model employs recursive or state-space methods, such as Kalman filtering or extended Kalman filtering. The model uses the target's current position, velocity, and acceleration as state variables to establish a dynamic equation for the motion state, while simultaneously inputting the observed signals as the measurement equation. By minimizing the state estimation error, the model can continuously update the target's position and velocity predictions, thus forming a smooth trajectory prediction.

[0099] Furthermore, the model incorporates task status and environmental constraints during updates. For example, when a task has a high priority or limited coverage area, the model weights the predictions to ensure the beam is directed towards the target of greater interest; when weather conditions or flight altitude change, the model considers potential signal attenuation and beam offset corrections to maintain prediction accuracy.

[0100] Finally, the model output includes the target's estimated position, velocity, and motion trend vector at the next moment. This result is directly used to pre-offset the subarray weight vector of the dynamic beamforming matrix, enabling the beam to point to the target's future position in advance and achieve continuous tracking.

[0101] Example 2

[0102] Furthermore, as a preferred embodiment of the above embodiments, a UAV target perception system based on beam reconstruction and adaptive environmental perception is proposed. This system is implemented based on the UAV target perception method based on beam reconstruction and adaptive environmental perception proposed in the above embodiments, specifically including:

[0103] An environment-mission information acquisition module is used to collect UAV flight environment and mission status information and generate a joint environment-mission feature vector; the environment-mission information acquisition module includes:

[0104] The external environment parameter acquisition unit is used to acquire environmental information such as meteorological conditions, flight altitude, and visibility;

[0105] The task status acquisition unit is used to obtain the current task type, number of targets, priority, and task coverage area;

[0106] The feature fusion unit is used to weight and fuse external environmental parameters and task status information to generate a joint environment-task feature vector, which is then output to the subarray beam weight dynamic reconstruction module.

[0107] The subarray beam weight dynamic reconstruction module is used to generate and update the subarray weight vector in real time based on the environment-task joint feature vector, and to calculate the target signal-to-noise ratio;

[0108] The subarray beam weight dynamic reconstruction module includes:

[0109] The initial weight generation unit is used to decompose the joint feature vector into task features and environment features, generate task weight vector and environment-corrected weight vector respectively, and fuse them into an initial subarray weight vector.

[0110] A continuously variable weight update unit is used to iteratively adjust the subarray weight vector, combining the minimum mean square error criterion and environment-task constraints.

[0111] The target signal-to-noise ratio calculation unit is used to perform weighted synthesis and energy-to-noise ratio calculation on the output signals of the subarray beams to obtain the target signal-to-noise ratio.

[0112] The differentiated power allocation module is used to generate the power allocation results of the left and right subarrays based on the target signal-to-noise ratio, environmental parameters and mission status information;

[0113] The differentiated power allocation module includes:

[0114] The power allocation input building unit is used to combine target signal-to-noise ratio, weather conditions, flight altitude, visibility, mission type, priority, and coverage area into a power allocation modeling input set.

[0115] The power allocation function modeling unit is used to establish a nonlinear power allocation function using a weighted exponential function.

[0116] The power allocation curve generation unit is used to generate curves describing the relationship between the output power of the left and right subarrays based on the power allocation function.

[0117] The output power calculation unit is used to calculate the output power values ​​of the left and right subarrays based on the power allocation curve, generate differentiated power allocation results, and output them to the dynamic beamforming module.

[0118] The dynamic beamforming and signal processing module is used to form a beam matrix based on the subarray weight vector and the differential power allocation result, and to process the received echo signal.

[0119] The dynamic beamforming and signal processing module includes:

[0120] The basic parameter fusion unit is used to fuse the subarray weight vector with the differential power allocation result to form a basic parameter set;

[0121] The dynamic beamforming matrix generation unit is used to generate a dynamic beamforming matrix based on a set of basic parameters.

[0122] The echo signal weighting processing unit is used to perform weighted synthesis and filtering of the received target echo signal using a beam matrix.

[0123] The performance index calculation unit is used to calculate the detection probability, tracking accuracy, and signal-to-noise ratio retention, and is used to evaluate the beamforming effect.

[0124] The target motion trend prediction and beam pre-offset module is used to predict the future position of the target and adjust the beam pointing in advance;

[0125] The target motion trend prediction and beam pre-offset module includes:

[0126] The target motion trend prediction unit is used to establish a target motion prediction model using echo signals, signal-to-noise ratio, and performance indicators, and to estimate the target's position at the next moment.

[0127] The prediction position mapping unit is used to map the predicted position to the subarray weight vector corresponding to the dynamic beamforming matrix;

[0128] The beam pre-offset adjustment unit is used to adjust the subarray weight vector according to the predicted position, so as to pre-offset the beam pointing to the future position of the target.

[0129] Example 3

[0130] Furthermore, in conjunction with the above embodiments, an application scenario for an UAV target perception method based on beam reconstruction and adaptive environmental perception is proposed. This scenario applies the UAV target perception method based on beam reconstruction and adaptive environmental perception, specifically:

[0131] When conducting multi-target patrol missions on the outskirts of the city, the UAV needs to simultaneously track moving vehicles, motorcycles, and low-altitude drones. The patrol area is approximately 30 square kilometers, with complex terrain including tall buildings, roads, and open spaces. Environmental conditions are variable, with wind speeds ranging from 0 to 10 meters per second, visibility between 500 and 3000 meters, and flight altitudes between 100 and 250 meters. The mission status information includes the mission type as multi-target patrol, the number of targets as 5, and the priorities from highest to lowest as vehicles, drones, and motorcycles. The mission coverage area is the designated road and surrounding open space. The UAV collects real-time meteorological conditions, visibility, and flight altitude data through sensors. Simultaneously, it obtains mission type, number of targets, priority, and coverage area information from the mission planning system. Then, it generates an environment-mission joint feature vector, where mission status information is weighted, and environmental information is adjusted using environmental correction weights. The fused joint feature vector represents the UAV's current environmental constraints and mission priority, providing input for beam weight adjustment. The joint feature vector is decomposed into task features and environmental features to generate an initial subarray weight vector. A continuously variable weight update algorithm iteratively adjusts the amplitude and phase of the subarray weight vector to minimize the root mean square error between the target signal and the subarray beam output, while simultaneously satisfying environmental-task constraints. Based on the new beam output signal formed by the updated subarray weight vector, the received signal is weighted and synthesized, and the target signal-to-noise ratio (SNR) is calculated. The target SNR, flight altitude, visibility, and task status information are combined as the input set for power allocation modeling. A power allocation function is constructed using a weighted exponential function, generating curves describing the power allocation relationship between the left and right subarrays under different environmental and task states. The output power of the left and right subarrays is calculated based on these curves, resulting in differentiated power allocation results. For example, for a high-priority vehicle target, the left subarray output power is 12 watts, and the right subarray output power is 10 watts; for a low-priority UAV, the left and right subarrays share an average output power of 7 watts. The subarray weight vector and the differentiated power allocation results are fused to construct a set of basic parameters, generating a dynamic beamforming matrix. This matrix is ​​then used for weighted synthesis and filtering of the received echo signals. Simultaneously, the detection probability, tracking accuracy, and signal-to-noise ratio (SNR) preservation are calculated to provide evaluation metrics for subsequent target prediction, such as a detection probability greater than 95%, tracking accuracy less than 2 meters, and SNR preservation greater than 0.9. Based on the target echo signal, SNR, detection probability, tracking accuracy, and SNR preservation, a target motion trend prediction model is constructed. This model predicts the target's position at the next moment and maps the predicted position to the subarray weight vector corresponding to the dynamic beamforming matrix. The subarray weight vector is pre-offset adjusted to ensure the beam points to the target's future position in advance, thus achieving continuous tracking of fast-moving targets.

[0132] The above description is merely a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the forms disclosed herein and should not be construed as excluding other embodiments. It can be used in various other combinations, modifications, and environments, and can be altered within the scope of the concept described herein through the above teachings or related technologies or knowledge. Modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention should be within the protection scope of the appended claims.

Claims

1. A UAV target perception method based on beam reconstruction and adaptive environmental perception, characterized in that, Includes the following steps: S1. Obtain the external environment parameters and mission status information of the current UAV, and establish a joint feature vector of environment-mission; S2. Based on the joint feature vector of environment and task, the subarray beam weights are dynamically reconstructed. The weight vectors are adjusted in real time through a continuously variable weight update algorithm. Combined with the minimum mean square error criterion and constraint optimization method, the environmental features are mapped to the subarray weight vector. The target signal-to-noise ratio is calculated based on the output signal of the new subarray beam formed by the subarray weight vector. S3. Based on the target signal-to-noise ratio, combined with external environmental parameters and task status information, perform nonlinear modeling of the power allocation function; by constructing power allocation curves, perform differentiated configuration of the output power of the left and right subarrays, and generate differentiated power allocation results for the left and right subarrays; S4. Based on the subarray weight vector and the differential power allocation results, construct a dynamic beamforming matrix, and process the target echo signal according to the constructed dynamic beamforming matrix, and calculate the detection probability, tracking accuracy and signal-to-noise ratio preservation. S5. Based on the processed target echo signal and the calculated target signal-to-noise ratio, predict the target's position at the next moment using the target motion trend prediction model; based on the predicted position, pre-offset the subarray weight vector of the dynamic beamforming matrix to pre-offset the beam pointing to the target's future position.

2. The UAV target perception method based on beam reconstruction and adaptive environmental perception as described in claim 1, characterized in that, The external environmental parameters include weather conditions, flight altitude, and visibility; the mission status information includes the current mission type, number of targets, priority, and mission coverage area.

3. The UAV target perception method based on beam reconstruction and adaptive environmental perception as described in claim 1, characterized in that, In step S1, the specific steps for establishing the joint feature vector of environment and task are as follows: S101. Normalize the external environment parameters and task status information to obtain the external environment vector and task status feature vector; S102. Based on task priority and coverage requirements, assign weights to task status information to obtain a weighted task status feature vector; and based on the evaluation of beam sensitivity by environmental parameters, transform the external environment vector into environmental correction weights. S103. Perform feature fusion between the weighted task state feature vector and the environment correction weight to obtain and output the joint feature vector.

4. The UAV target perception method based on beam reconstruction and adaptive environmental perception as described in claim 3, characterized in that, In step S103, the joint feature vector is specifically constructed through weighted fusion, as follows: ； Among them, the The joint feature vector representing the environment and the task, the The relative weights of environmental features in the joint vector are represented by the following: The weights of task features in the joint vector are represented by the following: The environmental adjustment weights are used to adjust the influence of environmental features on the joint vector. This represents the normalized environmental feature vector, the This represents the weighted feature vector of the task.

5. The UAV target perception method based on beam reconstruction and adaptive environmental perception as described in claim 1, characterized in that, Step S2 specifically includes the following sub-steps: S201. Decompose the joint feature vector to obtain task features and environment features; and generate subarray weight vectors based on task features and environment features; S202. Using the subarray weight vector as the initial value, the subarray weight vector is iteratively adjusted through the weight update algorithm. In the iteration, the joint feature vector of environment-task is used as the constraint condition, and the minimum mean square error is used as the optimization objective to adjust the amplitude and phase distribution of the initial subarray weights and map the environmental features to the subarray weight vector. S203. Based on the new subarray beam formed by the subarray weight vector, the received signal is weighted and synthesized to obtain the synthesized signal. The energy-to-noise power ratio of the synthesized signal is calculated to obtain the target signal-to-noise ratio.

6. The UAV target perception method based on beam reconstruction and adaptive environmental perception as described in claim 5, characterized in that, In step S202, the minimum mean square error is specifically expressed as the optimization objective as follows: ； Among them, the The minimum mean squared error loss function is denoted as . The submatrix weight vector is described above. The joint feature vector representing the environment and the task, the The function represents the expected value. The value of the target echo signal is represented by the following: The conjugate transpose of the submatrix weight vector is given by the following. Represents the received signal vector, the This represents the weight constraint function.

7. The UAV target perception method based on beam reconstruction and adaptive environmental perception as described in claim 5, characterized in that, Step S201 specifically includes the following sub-steps: S2011. Decompose the joint feature vector into task feature sub-vectors and environment feature sub-vectors; extract target priority information based on the task feature sub-vectors, and construct a task weight vector based on the target priority information; extract environmental information based on the environment feature sub-vectors, and construct an environment-corrected weight vector based on the environmental information. S2012. Linearly fuse the task weight vector and the environment correction weight vector to obtain the subarray weight vector.

8. The UAV target perception method based on beam reconstruction and adaptive environmental perception as described in claim 1, characterized in that, Step S3 specifically includes the following sub-steps: S301. Combine the output signal-to-noise ratio, meteorological conditions, flight altitude, and visibility from the external environmental parameters, and the task type, priority, and coverage area from the task status information into the input set for power allocation modeling. S302. Using the input set as the independent variable, construct a power allocation function through a weighted exponential function, and use the power allocation function to generate power allocation curves that describe the power allocation relationship between the left and right subarrays under different environments and task states; S303. Based on the power distribution curve, calculate the output power values ​​of the left and right subarrays respectively to form differentiated power distribution results.

9. The UAV target perception method based on beam reconstruction and adaptive environmental perception as described in claim 1, characterized in that, Step S4 specifically includes the following sub-steps: S401. Merge the subarray weight vector with the differential power allocation results to construct a set of basic parameters; S402. Generate a dynamic beamforming matrix based on the fused set of basic parameters; S403. The received target echo signal is weighted, synthesized, and filtered using a beamforming matrix; S404. Calculate the detection probability, tracking accuracy, and signal-to-noise ratio retention respectively.

10. The UAV target perception method based on beam reconstruction and adaptive environmental perception as described in claim 9, characterized in that, In step S404, the calculation process is as follows: The detection probability is calculated by combining the target signal amplitude with the noise statistical characteristics. Tracking accuracy is calculated based on the target's azimuth and velocity estimation errors; The signal-to-noise ratio (SNR) retention is calculated by comparing the rate of change of SNR before and after beamforming.

11. The UAV target perception method based on beam reconstruction and adaptive environmental perception as described in claim 9, characterized in that, In step S402, the dynamic beamforming matrix is ​​represented as follows: ； Among them, the Represents the dynamic beamforming matrix, the The submatrix weight vector is described above. Indicates the output power of the left subarray, the This indicates the output power of the right subarray.

12. The UAV target perception method based on beam reconstruction and adaptive environmental perception as described in claim 1, characterized in that, Step S5 specifically includes the following sub-steps: S501. Based on the processed target echo signal, signal-to-noise ratio, and calculated detection probability, tracking accuracy, and signal-to-noise ratio preservation, construct a target motion trend prediction model; wherein, the detection probability is used to constrain the target existence confidence level, the tracking accuracy is used to limit the error range of the predicted trajectory, and the signal-to-noise ratio preservation is used to correct the echo signal to reduce the impact of the echo signal on the prediction model. S502. Predict the target's position at the next moment by constructing a target motion trend prediction model; S503. Map the predicted position to the subarray weight vector corresponding to the dynamic beamforming matrix, and pre-offset the subarray weight vector to pre-offset the beam to the future position of the target.

13. The UAV target perception method based on beam reconstruction and adaptive environmental perception as described in claim 12, characterized in that, In step S503, the pre-offset adjustment is specifically expressed as follows: ； ； ； Among them, the This represents the subarray weight vector after pre-offset, the The submatrix weight vector is described above. Indicates the weight adjustment amount, the and These represent the target's predicted position estimate at the next time step and the target's position estimate at the current time step, respectively. This represents the target velocity estimate at the current moment. Indicates the time step.

Images