A fast beam tracking method and system for high dynamic groups

Abstract

The application discloses a kind of fast beam tracking method and system for high dynamic group, belong to wireless communication technical field.The method includes: base station obtains terminal initial angle information by spectral estimation method;Two auxiliary beams are used to send signal, and terminal measures received power and calculates power ratio, and actual physical angle is obtained by inverse solution;Angle is used as the observation of extended Kalman filter, and dynamic beam direction estimation is realized by state prediction and correction;Real-time monitoring received power is carried out to beam misalignment judgment, if misalignment, then trigger reinitialization;The angle estimation residual of extended Kalman filter output is input to long short-term memory network, and the estimation result is compensated by residual modeling and prediction.The system includes corresponding module to realize the above steps.The application combines auxiliary beam pair, extended Kalman filter and long short-term memory network, improves beam tracking precision and robustness, and is suitable for high-speed mobile scenarios such as unmanned aerial vehicle group.

Description

Technical Field

[0001] This invention belongs to the field of wireless communication technology, specifically relating to a fast beam tracking method and system for high dynamic groups. Background Technology

[0002] With the rapid development of UAV technology, UAV swarms are increasingly widely used, making the demand for high-speed, low-latency real-time communication more urgent. Millimeter-wave massive MIMO technology has demonstrated unique value in UAV swarm communication. By forming narrow beams through large-scale antenna arrays, it improves signal gain and anti-interference capabilities, meeting the communication needs of UAV swarms in long-distance, high-dynamic scenarios. However, the high-speed movement of UAV swarms poses a severe challenge to beam tracking. The high speed and drastic angle changes of the terminals mean that traditional beam alignment techniques, requiring periodic full-range scanning, suffer from high overhead and response lag, making real-time beam direction matching difficult. Simultaneously, the complex flight environment of UAV swarms, with significant multipath effects, Doppler frequency shift, and electromagnetic interference, leads to decreased accuracy in received signal observation, further exacerbating the risk of beam misalignment and severely impacting communication link stability. Therefore, it is urgent to design a beam tracking mechanism adapted to the high-dynamic characteristics of UAV swarms, reducing computational complexity while improving tracking accuracy and anti-interference capabilities to ensure the stability and reliability of UAV swarm communication links.

[0003] In beam tracking research for millimeter-wave massive MIMO systems, existing methods are mainly divided into two categories: model-driven and data-driven. Model-driven methods are represented by Kalman filter algorithms, such as the Extended Kalman Filter (EKF), which constructs an angle state model and an observation model to achieve recursive estimation of dynamic angles. Among them, the improved algorithm based on auxiliary beam pairs (ABP) constructs the observation function through power and difference ratio, reducing the impact of interference such as multipath and frequency offset, and improving tracking accuracy in nonlinear scenarios. However, Kalman filter algorithms are highly dependent on accurate state transition models. When the terminal motion mode is complex, such as the sudden change of direction or acceleration of UAV swarms, which leads to model mismatch, the tracking error will accumulate significantly, and even filter divergence may occur. Furthermore, in low signal-to-noise ratio environments, the power and difference ratio of auxiliary beam pair technology are easily affected by noise interference, and the nonlinear characteristics of the observation function may lead to a decrease in accuracy.

[0004] Data-driven approaches include Long Short-Term Memory (LSTM) networks and deep reinforcement learning algorithms. LSTM compensates for EKF estimation errors through residual modeling, enhancing adaptability to complex motion scenarios. Deep Q-Networks (DQNs), based on Markov decision processes, use pilot signal strength as state feedback and achieve dynamic beam direction adjustment through policy optimization, reducing online computational complexity. However, LSTM networks require extensive historical data for training, and residual feature extraction is insufficient in scenarios with abrupt angle changes, limiting the compensation effect. Furthermore, while deep reinforcement learning algorithms reduce model dependence, the priority experience replay mechanism offers limited accuracy improvement, and the average error control capability in highly dynamic scenarios still has room for optimization, making it difficult to fully meet the real-time and robustness requirements of high-speed mobile terminals such as drone swarms. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a fast beam tracking method and system for high dynamic groups, which addresses the shortcomings of the prior art. This method solves the technical problems of insufficient accuracy and link instability caused by traditional algorithms in high dynamic group beam tracking due to model mismatch, interference sensitivity, and tracking failure.

[0006] The present invention adopts the following technical solution:

[0007] A fast beam tracking method for high dynamic groups includes the following steps:

[0008] S1. The base station obtains the initial angle information of the terminal through the spectrum estimation method and completes the initial state estimation; after establishing the communication link, it uses two auxiliary beam pairs to transmit signals, and the terminal measures the received signal power of the two auxiliary beam pairs.

[0009] S2. Based on the received signal power of the two auxiliary beams obtained in step S1, calculate the power ratio, solve the inverse solution to obtain the spatial angular frequency, and then obtain the actual physical angle of the terminal.

[0010] S3. Using the actual physical angle obtained in step S2 as the observation of the extended Kalman filter, construct the state space, and realize the dynamic estimation of the current beam direction through state prediction and correction.

[0011] S4. Monitor the received power in real time to determine if the beam is out of alignment. If it is out of alignment, trigger re-initialization and re-evaluate the angle.

[0012] S5. Input the angle estimation residual of the extended Kalman filter output in step S3 into the long short-term memory network. The long short-term memory network is used to model and predict the residual to compensate and correct the estimation result of the extended Kalman filter. Output the angle estimation result after compensation by the long short-term memory network to complete beam tracking in high dynamic scenes.

[0013] Preferably, in step S1, the spectral estimation method is the MUSIC algorithm.

[0014] Preferably, in step S2, the power ratio is calculated as follows: based on the received signal power of the two auxiliary beams, the power ratio of the auxiliary beam pair is defined as the ratio of the two received signal powers, and the spatial angular frequency is obtained through an inverse function analytical expression.

[0015] Preferably, in step S3, the state prediction and correction includes:

[0016] State prediction: Predicting the angle state based on the state transition equation;

[0017] Observation function linearization: Linearizing the measurement equations;

[0018] Kalman gain calculation: Calculate the Kalman gain based on the linearized observation function and covariance matrix;

[0019] State correction: The predicted state is corrected using Kalman gain and observations.

[0020] Preferably, the state space is constructed based on an angle state model, and the state vector includes angle, angular velocity, and angular acceleration.

[0021] Preferably, in step S4, determining whether the beam is misaligned includes: based on the functional relationship between received power and direction error, when the received power is lower than the power level corresponding to the 3dB beamwidth, it is determined that the beam is misaligned.

[0022] Preferably, the directional error is obtained by back-calculation of the received power, and when the directional error exceeds 3dB beamwidth, re-initialization is triggered.

[0023] Preferably, in step S5, the structure of the long short-term memory network includes: an input layer, a fully connected layer, two stacked LSTM modules, a dropout layer, a ReLU activation layer, and an output fully connected layer; wherein the number of LSTM units in the first layer is 256, and the number of LSTM units in the second layer is 128.

[0024] Preferably, the high-dynamic group is a swarm of unmanned aerial vehicles (UAVs).

[0025] Secondly, embodiments of the present invention provide a fast beam tracking system for high-dynamic groups, comprising:

[0026] In the initial module, the base station obtains the initial angle information of the terminal through the spectrum estimation method and completes the initial state estimation; after establishing the communication link, it uses two auxiliary beam pairs to transmit signals, and the terminal measures the received signal power of the two auxiliary beam pairs.

[0027] The calculation module calculates the power ratio based on the received signal power of the two auxiliary beams obtained by the initial module, and then solves the spatial angular frequency to obtain the actual physical angle of the terminal.

[0028] The estimation module uses the actual physical angles obtained by the calculation module as the observations of the extended Kalman filter to construct the state space. Through state prediction and correction, it realizes the dynamic estimation of the current beam direction.

[0029] The judgment module monitors the received power in real time and determines whether the beam is out of sync. If it is out of sync, it triggers re-initialization and re-evaluates the angle.

[0030] The correction module inputs the angle estimation residual from the extended Kalman filter output in the estimation module into a long short-term memory network. The long short-term memory network models and predicts the residual, thereby compensating and correcting the estimation result of the extended Kalman filter. The module outputs the angle estimation result after compensation by the long short-term memory network, thus completing beam tracking in high dynamic scenes.

[0031] Thirdly, a computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the above-described fast beam tracking method for high dynamic groups.

[0032] Fourthly, embodiments of the present invention provide a computer-readable storage medium including a computer program that, when executed by a processor, implements the steps of the above-described fast beam tracking method for high dynamic groups.

[0033] Fifthly, a chip includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the above-described fast beam tracking method for high dynamic groups.

[0034] In a sixth aspect, embodiments of the present invention provide an electronic device including a computer program, which, when executed by the electronic device, implements the steps of the above-described fast beam tracking method for high dynamic groups.

[0035] Compared with the prior art, the present invention has at least the following beneficial effects:

[0036] A fast beam tracking method for highly dynamic groups is proposed. This method constructs a multi-module collaborative beam tracking system through a complete process: initial angle acquisition, physical angle calculation, EKF dynamic estimation, misalignment detection, and LSTM residual correction. It integrates the advantages of model-driven and data-driven methods, achieving real-time recursive estimation of dynamic angles through EKF while using LSTM to compensate for model mismatch errors in EKF. A misalignment detection mechanism is also added to prevent tracking failure. This method solves the problems of low tracking accuracy and link instability caused by complex motion patterns and numerous interferences in highly dynamic groups, achieving a balance between accurate tracking and low latency, and is suitable for the communication needs of high-speed mobile scenarios such as UAV swarms.

[0037] Furthermore, the MUSIC algorithm, as a high-precision spectral estimation method, can effectively separate the signal subspace and noise subspace, providing accurate initial angle estimates and laying a reliable foundation for subsequent EKF tracking. This reduces the impact of initial errors on overall tracking performance, and especially in low signal-to-noise ratio environments, the MUSIC algorithm's anti-interference capability enhances the system's stability.

[0038] Furthermore, the power ratio is defined as the ratio of the received power of the two auxiliary beams, and the spatial angular frequency is solved using an inverse function analytical expression. This design has clear mathematical theoretical support. Compared to directly relying on single-beam power detection, the power ratio can offset some of the effects of noise and channel fading, enhancing the robustness of the observations. It can maintain good angle calculation accuracy even in low signal-to-noise ratio environments, avoiding physical angle calculation deviations caused by signal interference, and providing high-quality observations for EKF.

[0039] Furthermore, state prediction is based on the motion model to estimate angle changes in advance, the observation function is linearized to handle the nonlinear characteristics of ABP, the Kalman gain dynamically adjusts the weights of prediction and observation, and state correction integrates information to output the optimal estimate. This process improves the convergence speed and accuracy of tracking, especially in scenarios with sudden angle changes, enabling rapid response to dynamic changes and reducing lag.

[0040] Furthermore, by introducing angular velocity and angular acceleration, the state model more accurately describes the motion patterns of highly dynamic groups, enhancing the predictive power of the EKF. This reduces the risk of model mismatch, especially when the UAV swarm accelerates or changes direction, effectively tracking nonlinear angular changes. Figure 8 The results show that the model significantly reduces the estimation error under uniform angular acceleration motion, improving the system's adaptability to complex motion patterns.

[0041] Furthermore, by utilizing the easily observable parameter of received power, low-cost and efficient misalignment detection is achieved without additional hardware. The 3dB threshold, based on beamwidth characteristics, ensures the accuracy and timeliness of detection, preventing communication interruptions caused by beam deviation.

[0042] Furthermore, the design clarifies the logic for back-calculating the directional error and the re-initialization mechanism after misalignment. When the directional error exceeds 3dB of beamwidth, spectral estimation is triggered to re-estimate the main lobe direction. This design forms a closed-loop control of detection-feedback-reset, ensuring rapid recovery of accurate tracking after beam misalignment. It avoids the continuous expansion of errors after misalignment, shortens the link recovery time, enhances the system's adaptability to complex interference and extreme motion scenarios, and improves the robustness of beam tracking.

[0043] Furthermore, the LSTM network comprises two stacked layers, a specific unit count configuration, and functional layer designs, optimized for the characteristics of angular residual sequences. The first LSTM layer captures the sequence memory state, while the second layer extracts long-term dependency information. Combined with Dropout and ReLU activation layers, this effectively avoids network overfitting and enhances feature representation capabilities. This ensures that the LSTM accurately models the variation patterns of the EKF residuals, improving compensation and correction effects and further reducing tracking errors.

[0044] Furthermore, by explicitly defining high-dynamic groups as drone swarms, the application scenarios of the technical solution become more specific and targeted. Adapting to the characteristics of high-speed movement, complex flight environments, and high communication demands of drone swarms, it can more accurately address the core pain point of drone swarm beam tracking compared to general scenario designs. This provides a feasible solution for the implementation of millimeter-wave massive MIMO technology in drone swarm communication, expanding the practical application value of the technology.

[0045] It is understood that the beneficial effects of the second to sixth aspects mentioned above can be found in the relevant descriptions in the first aspect mentioned above, and will not be repeated here.

[0046] In summary, this invention achieves high-precision and robust beam tracking by integrating ABP-EKF-LSTM technology with a misalignment detection mechanism, effectively solving the model mismatch and interference problems in highly dynamic scenarios, improving communication link stability, and maintaining low complexity to meet real-time requirements.

[0047] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0048] Figure 1 A scene diagram illustrating the application of the method of this invention.

[0049] Figure 2 ABP-EKF algorithm beam tracking process

[0050] Figure 3 To assist beamforming of the structure

[0051] Figure 4 LSTM network overall architecture diagram

[0052] Figure 5 Real-time tracking diagram of first-order Markov angle changes based on ABP-EKF over 1000 time slots.

[0053] Figure 6 To determine the RMSE variation under different SNR conditions and its theoretical upper bound in a first-order Markov angle variation mode with a fixed number of antennas N=16.

[0054] Figure 7 The ABP-EKF algorithm provides the state estimation results under sustained angular acceleration.

[0055] Figure 8 To investigate the error variation under different SNR scenarios with a fixed number of antennas N=16 in a continuous angular acceleration variation mode.

[0056] Figure 9 The angle estimation results of the ABP-EKF algorithm under uniform linear motion.

[0057] Figure 10 To investigate the impact of different SNR conditions on estimation error under a constant antenna number N=16 in a uniform linear motion variation mode.

[0058] Figure 11 To compare the angle error performance of the extended Kalman filter algorithm with that of the introduced LSTM correction mechanism, the average angle error results were obtained from 1000 Monte Carlo simulations.

[0059] Figure 12 A schematic diagram of a computer device provided in an embodiment of the present invention;

[0060] Figure 13 This is a block diagram of a chip provided according to an embodiment of the present invention.

[0061] Among them, 60. Computer equipment; 61. Processor; 62. Memory; 63. Computer program; 600. Electronic device; 610. Processing unit; 620. Storage unit; 6201. Random access memory unit; 6202. Cache memory unit; 6203. Read-only memory unit; 6204. Program / utility; 6205. Program module; 630. Bus; 640. Display unit; 650. Input / output interface; 660. Network adapter; 700. External device. Detailed Implementation

[0062] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0063] In the description of this invention, it should be understood that the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0064] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0065] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes such combinations. For example, A and / or B can represent three cases: A alone, A and B simultaneously, and B alone. Additionally, the character " / " in this invention generally indicates that the preceding and following objects have an "or" relationship.

[0066] It should be understood that although terms such as first, second, third, etc., may be used in the embodiments of the present invention to describe the preset range, these preset ranges should not be limited to these terms. These terms are only used to distinguish the preset ranges from one another. For example, without departing from the scope of the embodiments of the present invention, the first preset range may also be referred to as the second preset range, and similarly, the second preset range may also be referred to as the first preset range.

[0067] Depending on the context, the word "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0068] The accompanying drawings illustrate various structural schematic diagrams according to embodiments disclosed in this invention. These drawings are not to scale, and some details have been enlarged for clarity, and some details may have been omitted. The shapes of the various regions and layers shown in the drawings, as well as their relative sizes and positional relationships, are merely exemplary and may deviate from reality due to manufacturing tolerances or technical limitations. Furthermore, those skilled in the art can design regions / layers with different shapes, sizes, and relative positions as needed.

[0069] This invention provides a fast beam tracking method for highly dynamic groups. Combining the environmental characteristics of high-dynamic movement and complex electromagnetic interference of unmanned swarms with the real-time communication requirements, it focuses on the design and implementation of beam tracking technology based on observation signals, aiming to meet the requirements of accurate beam tracking, anti-interference and stable communication, and low-latency link maintenance in high-speed maneuvering scenarios, thereby improving the overall stability, resilience and scalability of the network.

[0070] Please see Figure 1 This invention provides a beam tracking technology using a millimeter-wave massive MIMO system for high-speed mobile terminal scenarios such as drone swarms. All models employ a uniform linear array receiver model, and the base station is equipped with... One antenna, equipped with mobile terminal root antenna, through The radio frequency links are used for hybrid beamforming, with one radio frequency link dedicated to beam tracking.

[0071] This invention combines the ABP observation module with the EKF update mechanism and introduces an error judgment threshold to achieve dynamic anomaly detection. After the algorithm starts, it first performs initial state estimation, using the spectral estimation method to obtain the initial angle information of the terminal and using it as the initial state input of the extended Kalman filter; after entering the iteration process, the system performs EKF update according to the measurement signal, that is, combining the state prediction and the actual observation information to achieve dynamic estimation of the current beam direction.

[0072] Specifically, the filtering module predicts the state based on a preset motion model and corrects the predicted value using ABP observations, thereby achieving continuous and stable tracking of the target angle. Then, the beam misalignment detection module monitors in real time whether the currently estimated direction deviates from the true channel direction. If the observation error... Exceeding the set threshold This means that the current beam direction may have deviated from the main lobe region, triggering a re-initialization process, thereby re-estimating the initial state and starting a new round of tracking.

[0073] Table 1. Complete ABP-EKF Algorithm

[0074]

[0075] Please see Figure 2 The present invention provides a fast beam tracking method for high dynamic groups, comprising the following steps:

[0076] S1. The base station obtains the initial angle information of the terminal through the spectrum estimation method and completes the initial state estimation; after establishing the communication link, it uses two auxiliary beam pairs to transmit signals, and the terminal measures the received signal power of the two auxiliary beam pairs.

[0077] Please see Figure 3 The auxiliary beam pairs the structure with angular parameters. and The two angles, pointing towards adjacent spaces, work together to form a beam pattern, the specific structure of which is as follows:

[0078]

[0079] in, The auxiliary beam pointing angle to the center can be defined in the interval. Inside; This indicates the beamwidth in half. The detection area is defined as... Without loss of generality, the auxiliary beamwidth is set to... .

[0080] The received signal power of the two auxiliary beams is expressed as follows:

[0081]

[0082]

[0083] in, Represents the channel steering vector. , These are the weight vectors for the two auxiliary beams.

[0084] S2. Based on the received signal power of the two auxiliary beams obtained in step S1, calculate the power ratio, solve the inverse solution to obtain the spatial angular frequency, and then obtain the actual physical angle of the terminal.

[0085] During the initial access phase, the MS achieves directional alignment by executing a beam training procedure and obtains complete parameter information for the current channel, indicating that an ideal communication link has been established between the BS and the MS. Subsequently, the BS transmits signals using two auxiliary beams, and the MS measures the received signal power of each of these auxiliary beams. The received power of the two auxiliary beams at the receiving end is expressed as follows:

[0086]

[0087] in, This represents the channel steering vector. Due to noise interference in the channel, the power ratio of the auxiliary beam pair is defined using the sum and difference beams in target detection. Represented as:

[0088]

[0089] exist Unknown, and satisfies When, the above equation is a monotonic function, By adjusting the power ratio of the auxiliary beam pair The inverse function is used to obtain its analytical expression.

[0090]

[0091] The AoD obtained from this inverse solution is the spatial angular frequency, which can be further... Obtain the actual physical angle of the MS.

[0092] S3. Use the angle result obtained in step S2 as the observation of the extended Kalman filter (EKF) to construct the state space. Through state prediction and correction, realize the dynamic estimation of the current beam direction.

[0093] Using the angle measurement results from the ABP as the observations of the extended Kalman filter (EKF), a unified state space is constructed, yielding the state transition equation for the angle:

[0094]

[0095] At any moment The actual angle to be estimated is The spatial angular frequency is The center angle of ABP is Based on the power ratio of the auxiliary beam pair, a ratio index for the received signal strength is obtained as follows:

[0096]

[0097] The above equation is taken as the measurement equation for EKF; the measurement equation is defined as:

[0098]

[0099] in, It is Gaussian white noise.

[0100] Next, the angle can be tracked using EKF; the ABP-EKF algorithm flow is as follows:

[0101] First, state prediction is performed, as follows:

[0102]

[0103] Furthermore, the observation function is linearized as follows:

[0104]

[0105] Furthermore, the Kalman gain is calculated as follows:

[0106]

[0107] S4. Monitor the received power in real time to determine if the beam is out of alignment. If it is out of alignment, trigger re-initialization and re-evaluate the angle.

[0108] By receiving power in real time To determine whether beam misalignment has occurred in the current direction estimation.

[0109] First, establish the received power and direction error. The functional relationship between them. Considering only the horizontal angular direction, using a uniform linear array, the number of array elements is... The array gain function is:

[0110]

[0111] in, In error When I was very young, there were

[0112]

[0113] Substituting the above equation into the pattern gain, we obtain an approximation.

[0114]

[0115] Combined with the formula , and when When the angle is small, a small-angle deviation near the main lobe will have

[0116]

[0117] like And the angle of observation Finally, we get:

[0118]

[0119] The above system can be used to determine if beam misalignment has occurred. Specifically, when the received power falls below a certain threshold, such as the power level corresponding to a 3dB beamwidth, i.e. From the perspective of angle error, given the current received power... Reverse direction error:

[0120]

[0121] When that When the beamwidth reaches 3dB, it is considered misaligned. At this point, relying solely on EKF for angle tracking is insufficient to guarantee estimation accuracy, necessitating retraining of the beam to restore the communication link. Therefore, spectral estimation methods, such as the MUSIC algorithm, are introduced to estimate the current main lobe direction of the beam in real time.

[0122] S5. Input the angle estimation residual output by EKF in step S3 into a Long Short-Term Memory (LSTM) network. Model and predict the residual through LSTM to compensate and correct the EKF estimation result. Output the angle estimation result after LSTM compensation to complete beam tracking in high dynamic scenes.

[0123] The LSTM takes the angle estimation residual sequence (dimension (1, 200)) output by the EKF as input, maps it through a fully connected layer, and then feeds it into a two-layer stacked LSTM module (256 units in the first layer and 128 units in the second layer). The output is the predicted residual.

[0124]

[0125] To achieve compensation and correction, among which Estimate the angle for EKF. For LSTM prediction residuals.

[0126] EKF relies on a priori defined state transition model to predict the system state. However, when the motion model description is not refined enough, its prediction error will increase significantly. To address this, the LSTM algorithm is further combined with EKF to propose the EKF-LSTM algorithm.

[0127] Please see Figure 4 A neural network structure suitable for residual correction was designed; the dimension of the input data is... , This indicates the length of the sample sequence. Each sample consists of a sequence of estimated angles output by EKF.

[0128] First, the input data is mapped to latent space features through a fully connected layer. Enhance the ability to express features;

[0129] Subsequently, the feature sequence is input into two stacked LSTM modules. The first LSTM layer captures the final memory state of the sequence, and the second LSTM layer is used to extract long-term dependency information from the time series.

[0130] Specifically, at each time step The input to LSTM is dimension The single-step estimation results.

[0131] Considering the physical continuity of beam angle changes, this output is used for residual sequence modeling and prediction, predicting the residuals. The activated ReLU units are further mapped to the output sequence via a fully connected layer. That is, the network that jointly predicts the residuals from multiple time steps will ultimately output a residual sequence. Adding back the original EKF estimated sequence, we obtain the corrected tracking results:

[0132]

[0133] The specific parameters of the network are shown in Table 2.

[0134] Table 2 LSTM Network Parameter Table

[0135]

[0136] This structure essentially constructs a residual learning system with EKF as the main filter and LSTM as the auxiliary compensator. During training, the difference between the predicted result and the true trajectory is used to calculate the loss function. We choose the root mean square error as the regression loss, and its calculation formula is as follows:

[0137]

[0138] In another embodiment of the present invention, a fast beam tracking system for high dynamic groups is provided. This system can be used to implement the above-mentioned fast beam tracking method for high dynamic groups. Specifically, the fast beam tracking system for high dynamic groups includes an initialization module, a calculation module, an estimation module, a judgment module, and a correction module.

[0139] In the initial module, the base station obtains the initial angle information of the terminal through the spectrum estimation method and completes the initial state estimation; after establishing the communication link, it uses two auxiliary beam pairs to transmit signals, and the terminal measures the received signal power of the two auxiliary beam pairs.

[0140] The calculation module calculates the power ratio based on the received signal power of the two auxiliary beams obtained by the initial module, and then solves the spatial angular frequency to obtain the actual physical angle of the terminal.

[0141] The estimation module uses the actual physical angles obtained by the calculation module as the observations of the extended Kalman filter to construct the state space. Through state prediction and correction, it realizes the dynamic estimation of the current beam direction.

[0142] The judgment module monitors the received power in real time and determines whether the beam is out of sync. If it is out of sync, it triggers re-initialization and re-evaluates the angle.

[0143] The correction module inputs the angle estimation residual from the extended Kalman filter output in the estimation module into a long short-term memory network. The long short-term memory network models and predicts the residual, thereby compensating and correcting the estimation result of the extended Kalman filter. The module outputs the angle estimation result after compensation by the long short-term memory network, thus completing beam tracking in high dynamic scenes.

[0144] This invention provides a terminal device comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, graphics processing units (GPUs), tensor processing units (TPUs), digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions to achieve a corresponding method flow or corresponding function. The processor described in this embodiment can be used for the operation of a fast beam tracking method for high-dynamic groups, including:

[0145] The base station obtains the initial angle information of the terminal through spectral estimation to complete the initial state estimation. After establishing the communication link, it transmits signals using two auxiliary beam pairs, and the terminal measures the received signal power of the two auxiliary beams. Based on the obtained received signal power of the two auxiliary beams, the power ratio is calculated, and the spatial angular frequency is obtained through inverse solving, thus obtaining the actual physical angle of the terminal. The obtained actual physical angle is used as the observation of the extended Kalman filter to construct the state space. Through state prediction and correction, dynamic estimation of the current beam direction is achieved. The received power is monitored in real time to determine whether the beam is out of sync. If it is out of sync, re-initialization is triggered, and angle estimation is performed again. The angle estimation residual output by the extended Kalman filter is input into the long short-term memory network. The long short-term memory network models and predicts the residual to compensate and correct the estimation result of the extended Kalman filter. The angle estimation result after compensation by the long short-term memory network is output to complete beam tracking in high dynamic scenarios.

[0146] Please see Figure 12 The terminal device is a computer device. In this embodiment, the computer device 60 includes a processor 61, a memory 62, and a computer program 63 stored in the memory 62 and executable on the processor 61. When executed by the processor 61, the computer program 63 implements the fast beam tracking method for high-dynamic groups described in this embodiment. To avoid repetition, these details are not elaborated here. Alternatively, when executed by the processor 61, the computer program 63 implements the functions of each model / unit in the fast beam tracking system for high-dynamic groups described in this embodiment. To avoid repetition, these details are not elaborated here.

[0147] Computer device 60 can be a desktop computer, laptop, handheld computer, cloud server, or other computing device. Computer device 60 may include, but is not limited to, a processor 61 and a memory 62. Those skilled in the art will understand that... Figure 12 This is merely an example of computer device 60 and does not constitute a limitation on computer device 60. It may include more or fewer components than shown, or combine certain components, or different components. For example, computer device may also include input / output devices, network access devices, buses, etc.

[0148] The processor 61 may be a Central Processing Unit (CPU), or other general-purpose processors, graphics processing units (GPUs), tensor processing units (TPUs), digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0149] The memory 62 can be an internal storage unit of the computer device 60, such as a hard disk or RAM of the computer device 60. The memory 62 can also be an external storage device of the computer device 60, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., provided on the computer device 60.

[0150] Furthermore, the memory 62 may include both internal storage units of the computer device 60 and external storage devices. The memory 62 is used to store computer programs and other programs and data required by the computer device. The memory 62 can also be used to temporarily store data that has been output or will be output.

[0151] Please see Figure 13 The terminal device is an electronic device 600, which is manifested in the form of a general-purpose computing device. The components of the electronic device may include, but are not limited to: at least one processing unit 610, at least one storage unit 620, a bus 630 connecting different platform components (including storage unit 620 and processing unit 610), a display unit 640, etc.

[0152] The storage unit stores program code, which can be executed by the processing unit 610 to perform the steps described in the method section of this specification according to various exemplary embodiments of the present invention. For example, the processing unit 610 can perform actions such as... Figure 2 The steps are shown in the figure.

[0153] Storage unit 620 may include a readable medium in the form of a volatile storage unit, such as random access memory (RAM) 6201 and / or cache memory 6202, and may further include a read-only memory (ROM) 6203.

[0154] Storage unit 620 may also include a program / utility 6204 having a set (at least one) program module 6205, such program module 6205 including but not limited to: operating system, one or more application programs, other program modules and program data, each or some combination of these examples may include an implementation of a network environment.

[0155] Bus 630 can represent one or more of several types of bus structures, including a memory cell bus or memory cell controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of the multiple bus structures.

[0156] Electronic device 600 can also communicate with one or more external devices 700 (e.g., keyboard, pointing device, Bluetooth device, etc.), and with one or more devices that enable a user to interact with electronic device 600, and / or with any device that enables electronic device 600 to communicate with one or more other computing devices (e.g., router, modem). This communication can be performed via input / output interface 650. Furthermore, electronic device 600 can also communicate with one or more networks (e.g., local area network, wide area network, and / or public network, such as the Internet) via network adapter 660. Network adapter 660 can communicate with other modules of electronic device 600 via bus 630. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with electronic device 600, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage platforms.

[0157] Example 4

[0158] This invention also provides a storage medium, specifically a computer-readable storage medium, which is a memory device in a terminal device for storing programs and data. It is understood that the computer-readable storage medium here can include both built-in storage media in the terminal device and extended storage media supported by the terminal device; it can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, the storage space also stores one or more instructions suitable for loading and execution by a processor, which can be one or more computer programs (including program code). More specific examples of the computer-readable storage medium include: an electrical connection with one or more wires, a portable disk, a hard disk, random access memory, read-only memory, erasable programmable read-only memory, optical fiber, portable compact disk read-only memory, optical storage device, magnetic storage device, or any suitable combination thereof.

[0159] Computer-readable storage media also include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A readable storage medium can also be any readable medium other than a readable storage medium that can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the readable storage medium can be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, radio frequency, etc., or any suitable combination thereof.

[0160] Program code for performing the operations of this invention can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can execute entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0161] One or more instructions stored in a computer-readable storage medium can be loaded and executed by a processor to implement the corresponding steps of the fast beam tracking method for high dynamic groups in the above embodiments; one or more instructions in the computer-readable storage medium are loaded by the processor and executed as follows:

[0162] The base station obtains the initial angle information of the terminal through spectral estimation to complete the initial state estimation. After establishing the communication link, it transmits signals using two auxiliary beam pairs, and the terminal measures the received signal power of the two auxiliary beams. Based on the obtained received signal power of the two auxiliary beams, the power ratio is calculated, and the spatial angular frequency is obtained through inverse solving, thus obtaining the actual physical angle of the terminal. The obtained actual physical angle is used as the observation of the extended Kalman filter to construct the state space. Through state prediction and correction, dynamic estimation of the current beam direction is achieved. The received power is monitored in real time to determine whether the beam is out of sync. If it is out of sync, re-initialization is triggered, and angle estimation is performed again. The angle estimation residual output by the extended Kalman filter is input into the long short-term memory network. The long short-term memory network models and predicts the residual to compensate and correct the estimation result of the extended Kalman filter. The angle estimation result after compensation by the long short-term memory network is output to complete beam tracking in high dynamic scenarios.

[0163] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0164] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present 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 present invention without inventive effort are within the scope of protection of the present invention.

[0165] The beam tracking capability of this invention will be discussed next under three motion modes: linear motion, angular acceleration motion, and angle variation based on a first-order Markov model. The amplitude of the angle variation based on the first-order Markov model will be further analyzed. for( ).

[0166] Secondly, consider linear motion as an object starting from coordinates (0,1) and moving at a constant velocity along a straight line with a slope of 45° towards the positive x-axis. The state variable includes the distance relative to the origin BS. ,angle and linear velocity Then the state vector is defined as The equation of state assumes the object moves at a constant velocity. For motion in the positive x-axis direction, the motion relationship and state transition equations in polar coordinates are as follows:

[0167]

[0168] Consider another state equation for pure angle tracking. To describe the angle change of the terminal relative to the base station, we define the system state vector as follows: .in Indicates the first The departure angle within each time slot Indicates the first angular velocity within each time slot Then it means the first The angular acceleration within each time slot. In the considered DOA tracking scenario, the system state is a three-dimensional vector. Since the terminal angle change usually has a certain inertia, we can use a kinematic model to characterize the change law of the terminal DOA. The specific system state transition equation is as follows:

[0169]

[0170] Wherein, the state transition matrix Defined as ,here The sampling interval is... It is the Gaussian model noise of the system.

[0171] To improve training performance, a variable learning rate and the Adam optimizer are used when training the EKF-LSTM network. The L2 norm is used to prevent overfitting. Gradient thresholding is also used to prevent excessive impact during training. Specific training parameters are shown in Table 3.

[0172] Table 3 LSTM Network Parameter Training Table

[0173]

[0174] Please see Figure 5 and Figure 6 In angular change motion based on a first-order Markov model, Figure 5The state estimation trajectory of the ABP-EKF algorithm over 1000 time slots is presented. As can be seen from the figure, the Kalman filter, supported by the ABP observation structure, can more closely approximate the actual state trajectory. Figure 6 Further analysis was conducted on the RMSE variation under different SNR conditions with a fixed number of antennas N=16. It can be observed that at low SNR, the ABP-EKF scheme has better error control capability than the receiving signal scheme, with a smaller overall error curve value and faster convergence speed, indicating that it is more robust in noisy interference scenarios.

[0175] Please see Figure 7 and Figure 8 In the uniform angular acceleration motion model, Figure 7 The figure shows the state estimation results of the ABP algorithm under sustained angular acceleration. As can be seen from the figure, even when the real state exhibits a significant nonlinear growth trend, the Kalman filter, supported by the ABP structure, can still achieve accurate tracking with almost no lag, and the estimation curve closely follows the real trajectory, reflecting the algorithm's high adaptability to dynamic models. In contrast, the scheme based on the received signal exhibits certain estimation delays and error accumulation in similar scenarios. Figure 8 This describes the error variation under different SNR scenarios with an N=16 antenna configuration. It can be observed that ABP-EKF maintains a low error level under various signal-to-noise ratio conditions, and its RMSE is lower than the theoretical upper limit, outperforming similar results in the aforementioned received signal algorithms.

[0176] Please see Figure 9 and Figure 10 In the model based on uniform linear motion, Figure 9 The corresponding angle estimation results are shown, with blue circles representing the estimated results and red solid lines representing the actual angles. Throughout the prediction process, the angle estimation converged rapidly with small errors, and the fluctuation amplitude was slightly smaller than that of the received signal scheme. Figure 10 The impact of different signal-to-noise ratios (SNRs) on angle estimation error was further evaluated under a fixed array N=16 condition. The results show that the ABP algorithm exhibits good stability and error control within an SNR range of 0 dB to 20 dB, with the error curve consistently remaining near the theoretical upper limit. However, in a strong noise environment with an SNR of -10 dB, the estimation error increases, and the error curve exceeds the theoretical upper limit. This indicates that although ABP-EKF possesses superior performance under medium-to-high SNR conditions, its robustness significantly decreases in low SNR scenarios due to noise interference. The main reason is that the ABP observation model's extraction of the power difference ratio is highly dependent on signal quality. Under low SNR conditions, the power ratio information is easily masked by noise, leading to drastic fluctuations in the observed values ​​and compromising the estimation stability of ABP-EKF.

[0177] Please see Figure 11 Based on the LSTM-corrected EKF algorithm, this study mainly examines the magnitude of angle change when the angle conforms to a first-order Markov transformation. The correction status. Figure 11 The figure shows a comparison of the angle error performance between the Extended Kalman Filter (EKF) algorithm and the algorithm after introducing an LSTM correction mechanism. The average angle error was obtained from 1000 Monte Carlo simulations. The figure shows that before the correction mechanism was activated, the error curves of the two algorithms completely overlapped for indices 0 to 200. This indicates that without the learning model correction, the system relied solely on the EKF for state estimation, resulting in a continuously increasing error that accumulated over time, eventually reaching a large value of nearly 4° near index 200. After introducing the LSTM compensation model at index 200, the average angle error decreased from approximately 3.7° to 1.3°, a reduction of about 65%, preventing further expansion of the original EKF error. However, even after introducing the LSTM correction mechanism, the average angle estimation error still showed an increasing trend, failing to significantly suppress further error accumulation. This phenomenon may be related to the fact that the correction strategy used in the simulation did not fully consider the dynamic characteristics of error evolution, especially under trajectories with highly random angle changes, where system errors are difficult to effectively suppress through correction at a single moment. However, in practical applications, the changes in terminal angle often have certain physical regularities, and at this time, LSTM's ability to learn error trends will have more room to play its role.

[0178] In summary, this invention provides a fast beam tracking method and system for highly dynamic groups. It constructs an auxiliary beam pair with its power and difference ratio as the observation function, combines this with an extended Kalman filter to achieve dynamic angle estimation, and introduces a long short-term memory network to compensate for the estimation residuals. It also features a beam misalignment detection and reset mechanism. This not only improves beam tracking accuracy and reduces the impact of multipath and frequency offset interference in high-speed moving scenarios such as UAV swarms, but also enhances robustness under complex motion modes, meeting the requirements of low complexity and real-time performance.

[0179] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0180] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0181] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0182] In the embodiments provided by this invention, it should be understood that the disclosed devices / terminals and methods can be implemented in other ways. For example, the device / terminal embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0183] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0184] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0185] If the integrated module / unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random-access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0186] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus, and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0187] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0188] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0189] The above content is only for illustrating the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solution based on the technical concept proposed in this invention shall fall within the scope of protection of the claims of this invention.

Claims

1. A fast beam tracking method for high-dynamic groups, characterized in that, Includes the following steps: S1. The base station obtains the initial angle information of the terminal through the spectrum estimation method and completes the initial state estimation; after establishing the communication link, it uses two auxiliary beam pairs to transmit signals, and the terminal measures the received signal power of the two auxiliary beam pairs. S2. Based on the received signal power of the two auxiliary beams obtained in step S1, calculate the power ratio, solve the inverse solution to obtain the spatial angular frequency, and then obtain the actual physical angle of the terminal. S3. Using the actual physical angle obtained in step S2 as the observation of the extended Kalman filter, construct the state space, and realize the dynamic estimation of the current beam direction through state prediction and correction. S4. Monitor the received power in real time to determine if the beam is out of alignment. If it is out of alignment, trigger re-initialization and re-evaluate the angle. S5. Input the angle estimation residual of the extended Kalman filter output in step S3 into the long short-term memory network. The long short-term memory network is used to model and predict the residual to compensate and correct the estimation result of the extended Kalman filter. Output the angle estimation result after compensation by the long short-term memory network to complete beam tracking in high dynamic scenes.

2. The fast beam tracking method for high-dynamic groups according to claim 1, characterized in that, In step S1, the spectral estimation method is the MUSIC algorithm.

3. The fast beam tracking method for high-dynamic groups according to claim 1, characterized in that, In step S2, the power ratio is calculated as follows: based on the received signal power of the two auxiliary beams, the power ratio of the auxiliary beam pair is defined as the ratio of the two received signal powers, and the spatial angular frequency is obtained through an inverse function analytical expression.

4. The fast beam tracking method for high dynamic groups according to claim 1, characterized in that, In step S3, the state prediction and correction includes: State prediction: Predicting the angle state based on the state transition equation; Observation function linearization: Linearizing the measurement equations; Kalman gain calculation: Calculate the Kalman gain based on the linearized observation function and covariance matrix; State correction: The predicted state is corrected using Kalman gain and observations.

5. The fast beam tracking method for high dynamic groups according to claim 4, characterized in that, The state space is constructed based on the angle state model, and the state vector includes angle, angular velocity, and angular acceleration.

6. The fast beam tracking method for high dynamic groups according to claim 1, characterized in that, In step S4, determining whether the beam is misaligned includes: based on the functional relationship between received power and direction error, when the received power is lower than the power level corresponding to the 3dB beamwidth, it is determined that the beam is misaligned.

7. The fast beam tracking method for high dynamic groups according to claim 6, characterized in that, The directional error is obtained by back-calculation of the received power. When the directional error exceeds 3dB beamwidth, reinitialization is triggered.

8. The fast beam tracking method for high dynamic groups according to claim 1, characterized in that, In step S5, the structure of the Long Short-Term Memory network includes: an input layer, a fully connected layer, two stacked LSTM modules, a Dropout layer, a ReLU activation layer, and an output fully connected layer; wherein the first LSTM layer has 256 units and the second LSTM layer has 128 units.

9. The fast beam tracking method for high dynamic groups according to any one of claims 1 to 8, characterized in that, The high-dynamic group refers to a swarm of unmanned aerial vehicles (UAVs).

10. A fast beam tracking system for high-dynamic groups, characterized in that, include: In the initial module, the base station obtains the initial angle information of the terminal through the spectrum estimation method and completes the initial state estimation; after establishing the communication link, it uses two auxiliary beam pairs to transmit signals, and the terminal measures the received signal power of the two auxiliary beam pairs. The calculation module calculates the power ratio based on the received signal power of the two auxiliary beams obtained by the initial module, and then solves the spatial angular frequency to obtain the actual physical angle of the terminal. The estimation module uses the actual physical angles obtained by the calculation module as the observations of the extended Kalman filter to construct the state space. Through state prediction and correction, it realizes the dynamic estimation of the current beam direction. The judgment module monitors the received power in real time and determines whether the beam is out of sync. If it is out of sync, it triggers re-initialization and re-evaluates the angle. The correction module inputs the angle estimation residual from the extended Kalman filter output in the estimation module into a long short-term memory network. The long short-term memory network models and predicts the residual, thereby compensating and correcting the estimation result of the extended Kalman filter. The module outputs the angle estimation result after compensation by the long short-term memory network, thus completing beam tracking in high dynamic scenes.

Images