An ultra-wideband antenna beamforming method and system

Abstract

The application discloses a kind of ultra-wideband antenna beam forming method and system, it is related to ultra-wideband communication technical field.The method includes: obtaining the static parameter of ultra-wideband antenna array, constructs initial beam forming matrix;Synchronous acquisition array element received radio frequency analog signal, carries out analog-digital conversion to radio frequency analog signal, obtains discrete time sequence signal;Carry out space-time domain joint preprocessing, extract multipath signal characteristic parameter, and generate beam optimization objective function in combination with preset communication quality index;Using particle swarm optimization algorithm, with beam optimization objective function as target, to initial beam forming matrix Iterative optimization, obtain optimized beam forming matrix;The radiation pattern of dynamic adjustment ultra-wideband antenna array is formed focusing beam pointing to target receiving end;Real-time monitoring communication performance parameter, and according to monitoring result feedback adjustment beam optimization objective function.The application effectively improves the beam pointing precision and communication stability of ultra-wideband antenna.

Description

Technical Field

[0001] This invention relates to the field of ultra-wideband communication technology, and specifically to an ultra-wideband antenna beamforming method and system. Background Technology

[0002] With the widespread application of ultra-wideband (UWB) technology in wireless communication, radar detection, and precise positioning, the beamforming performance of UWB antennas has become a key factor determining the communication quality and detection accuracy of the system. In existing technologies, UWB antenna beamforming often employs fixed parameter design or preset direction control strategies, relying on the array's static geometric parameters and a preset target direction for beam pointing control.

[0003] However, traditional beamforming methods fail to fully consider the impact of multipath propagation, dynamic interference, and environmental changes on signal quality in real-world scenarios. This makes it difficult for the formed beam to be focused accurately on the target receiver in dynamic or complex environments in real time, resulting in problems such as beam pointing deviation and communication quality fluctuations. Summary of the Invention

[0004] This invention provides a method and system for ultra-wideband antenna beamforming, aiming to solve the technical problems of low beam pointing accuracy and insufficient communication stability of ultra-wideband antennas in the prior art.

[0005] In view of the above problems, the present invention provides an ultra-wideband antenna beamforming method and system.

[0006] In a first aspect, the present invention provides an ultra-wideband antenna beamforming method, comprising: Obtain the static parameters of the ultra-wideband antenna array of the target scene and construct an initial beamforming matrix. The ultra-wideband antenna array contains M array elements, and the static parameters include the geometric positions, preset target directions and operating frequency bands of the M array elements. Simultaneously acquire M radio frequency analog signals received by the M array elements, perform analog-to-digital conversion on the M radio frequency analog signals, and obtain M discrete time series signals; The M discrete time series signals are subjected to joint spatiotemporal preprocessing to extract multipath signal feature parameters, and a beam optimization objective function is generated by combining preset communication quality indicators. The particle swarm optimization algorithm is used to iteratively optimize the initial beamforming matrix with the beam optimization objective function as the optimization objective to obtain the optimized beamforming matrix. Based on the optimized beamforming matrix, the radiation pattern of the ultra-wideband antenna array is dynamically adjusted to form a focused beam pointing towards the target receiver. The communication performance parameters are monitored in real time, and the beam optimization objective function is adjusted based on the monitoring results.

[0007] In a second aspect, the present invention provides an ultra-wideband antenna beamforming system, comprising: The array parameter initialization module is used to obtain the static parameters of the ultra-wideband antenna array of the target scene and construct the initial beamforming matrix. The ultra-wideband antenna array contains M array elements, and the static parameters include the geometric position, preset target direction and operating frequency band of the M array elements. The radio frequency signal acquisition module is used to synchronously acquire the M radio frequency analog signals received by the M array elements, and perform analog-to-digital conversion on the M radio frequency analog signals to obtain M discrete time series signals. The optimization function construction module is used to perform spatiotemporal joint preprocessing on the M discrete time series signals, extract multipath signal feature parameters, and generate a beam optimization objective function in combination with preset communication quality indicators. The matrix iterative optimization module is used to perform iterative optimization of the initial beamforming matrix using the particle swarm optimization algorithm, with the beamforming objective function as the optimization target, to obtain the optimized beamforming matrix. The beamforming module is used to dynamically adjust the radiation pattern of the ultra-wideband antenna array based on the optimized beamforming matrix, so as to form a focused beam pointing towards the target receiver. The performance closed-loop feedback module is used to monitor communication performance parameters in real time and adjust the beam optimization objective function based on the monitoring results.

[0008] One or more technical solutions provided in this invention have at least the following technical effects or advantages: This invention provides a method and system for beamforming an ultra-wideband antenna. First, an initial beamforming matrix is ​​constructed based on the array geometry and a preset target, providing a starting point for optimization. Second, array signals are simultaneously acquired and digitized to ensure spatiotemporal consistency. Then, multipath features are extracted through joint spatiotemporal preprocessing and combined with communication quality indicators to form a quantifiable optimization target. Subsequently, a particle swarm optimization algorithm is used to efficiently iteratively search the beamforming matrix to obtain an optimized solution for the current channel environment. Afterward, the array radiation pattern is adjusted in real time using the optimized matrix to generate a focused beam pointing towards the target receiver. Finally, by monitoring communication performance and adjusting the optimization target accordingly, the system can adapt to time-varying environments. This invention achieves integrated processing from signal acquisition, channel feature extraction, intelligent optimization calculation to dynamic beam adjustment and closed-loop adaptive processing, effectively improving the beam pointing accuracy and communication stability of ultra-wideband antennas. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 This is a flowchart illustrating an ultra-wideband antenna beamforming method provided in an embodiment of the present invention. Figure 2 This is a schematic diagram of the structure of an ultra-wideband antenna beamforming system provided in an embodiment of the present invention; The components represented by each number in the attached diagram are explained below: Array parameter initialization module 11, radio frequency signal acquisition module 12, optimization function construction module 13, matrix iterative optimization module 14, beam dynamic shaping module 15, and performance closed-loop feedback module 16. Detailed Implementation

[0011] This invention provides a method and system for ultra-wideband antenna beamforming, which addresses the technical problems of low beam pointing accuracy and insufficient communication stability in existing ultra-wideband antennas.

[0012] Example 1, as Figure 1 As shown, the present invention provides an ultra-wideband antenna beamforming method, the method comprising: S100: Obtain the static parameters of the ultra-wideband antenna array of the target scene and construct an initial beamforming matrix, wherein the ultra-wideband antenna array contains M array elements, and the static parameters include the geometric positions, preset target directions and operating frequency bands of the M array elements.

[0013] In this embodiment of the invention, the static parameters of an ultra-wideband antenna array in a target scene are obtained, and an initial beamforming matrix is ​​constructed. The ultra-wideband antenna array comprises M elements, and the static parameters include the geometric positions of the M elements, the preset target direction, and the operating frequency band. Ultra-wideband antennas have advantages such as a wide operating frequency range, high transmission rate, and strong anti-interference capability. However, in practical applications, the wide bandwidth characteristic leads to differences in the response of antenna elements at different frequencies. If beamforming optimization is performed directly, problems such as beam pointing deviation, excessively high sidelobe levels, and poor frequency domain consistency are likely to occur. Furthermore, the beamforming matrix is ​​the core of controlling the radiation pattern of the antenna array. Without a reasonable initial matrix, subsequent optimization algorithms may get trapped in local optima, increasing optimization complexity, prolonging convergence time, and even failing to form a focused beam that meets communication requirements. Therefore, it is necessary to first obtain the static parameters of the antenna array in the target scene and construct an initial beamforming matrix through scientific steps. This lays a reliable foundation for subsequent beam optimization and dynamic adjustment, ensuring the stability and accuracy of ultra-wideband antenna beamforming.

[0014] Step S100 in the method provided in this embodiment of the invention includes: Obtain the geometric positions, preset target directions, and operating frequency bands of the M elements of the ultra-wideband antenna array; Within the operating frequency band, N discrete frequency points are selected at uniform intervals; Based on the geometric positions of the M array elements and the preset target direction, N initial beamforming weight vectors corresponding to the N discrete frequency points are calculated respectively, wherein each of the N initial beamforming weight vectors is an M-dimensional vector. The N initial beamforming weight vectors are arranged in columns to form an initial beamforming matrix with M rows and N columns.

[0015] First, the geometric positions, preset target directions, and operating frequency bands of the M elements of the ultra-wideband antenna array are obtained. An ultra-wideband antenna array is an antenna system composed of multiple antenna elements arranged according to a specific pattern. It not only meets the requirements of ultra-wideband operation but also allows for beam pointing control and gain adjustment by adjusting the signals of each element. The geometric position of an element refers to its specific location in space, usually marked in coordinate form with a specific element as a reference. The preset target direction refers to the beam pointing direction determined in advance according to communication or detection requirements, usually expressed as an angle with the array normal plane as a reference. The operating frequency band refers to the frequency range in which the antenna array can operate normally, specifying the start and end frequencies to cover the signal transmission requirements of the target application scenario. The geometric position data of the M elements, preset target direction parameters, and operating frequency band range are obtained by consulting the design drawings and technical manuals of the ultra-wideband antenna array, or by using actual measuring equipment to perform spatial coordinate measurements and frequency response tests on the constructed array.

[0016] For example, the ultra-wideband antenna array is set as a uniform linear array containing four elements, which are uniformly arranged along a straight line with a spacing of 0.1 meters between them. Spatial coordinates are established with the first element as the origin. The specific positions of the four elements are: element 1 (0,0,0), element 2 (0,0.1,0), element 3 (0,0.2,0), and element 4 (0,0.3,0). The preset target direction is (0°, 90°), meaning the signal needs to be transmitted horizontally to the right. The operating frequency band is 3GHz to 10GHz. Using the above method, the complete static parameters of the array can be obtained.

[0017] Secondly, within the operating frequency band, N discrete frequency points are selected at uniform intervals. Discrete frequency points refer to several discrete frequency values ​​selected within the operating frequency band, which serve as the reference frequencies for subsequent beamforming weight calculations. Uniform intervals mean that the frequency differences between the selected N frequency points are equal, ensuring a uniform distribution of frequency points within the operating frequency band and avoiding calculation deviations caused by concentrated or sparse frequency points. First, the start and end frequencies of the operating frequency band are determined, and the total width of the frequency interval is calculated. Then, based on the preset number of discrete frequency points N, the frequency interval between two adjacent frequency points is calculated. Finally, starting from the start frequency of the operating frequency band, the frequency intervals are sequentially accumulated to obtain N discrete frequency points.

[0018] For example, based on the operating frequency band from 3GHz to 10GHz, N=5 discrete frequency points are selected. First, the total width of the frequency range is calculated to be 10GHz-3GHz=7GHz, and the frequency interval is 7GHz / (5-1)=1.75GHz. Starting from the initial frequency of 3GHz, 1.75GHz is added sequentially to obtain 5 discrete frequency points, namely 3GHz, 4.75GHz, 6.5GHz, 8.25GHz, and 10GHz.

[0019] Next, based on the geometric positions of the M array elements and the preset target direction, N initial beamforming weight vectors corresponding to the N discrete frequency points are calculated respectively, wherein each of the N initial beamforming weight vectors is an M-dimensional vector.

[0020] Specifically, based on the geometric positions of the M array elements and the preset target direction, N initial beamforming weight vectors corresponding to the N discrete frequency points are calculated, including: Sort the N discrete frequency points in ascending order of frequency value, and select the first discrete frequency point as the target frequency point. Based on the geometric position coordinates of the M array elements and the preset target direction, combined with the wavelength of the target frequency point, the M relative phase delays of the M array elements are calculated. Based on the M relative phase delays, construct an M-dimensional array manifold vector; The M-dimensional array manifold vector is normalized to obtain the M-dimensional initial beamforming weight vector; By traversing the N discrete frequency points, N initial beamforming weight vectors corresponding to the N discrete frequency points are obtained; The weights include complex coefficients used to control the signal amplitude and phase of the corresponding array element at N discrete frequency points.

[0021] First, the N discrete frequency points are sorted in ascending order of frequency value, and the first discrete frequency point is selected as the target frequency point. The target frequency point is the reference frequency selected from the discrete frequency points, which serves as the reference frequency for subsequent phase delay calculations. The N discrete frequency points are sorted in ascending order of frequency value, and the first frequency point in the sorted sequence is selected as the target frequency point. For example, if N=5 discrete frequency points are 3GHz, 4.75GHz, 6.5GHz, 8.25GHz, and 10GHz, sorting these 5 frequency points in ascending order yields a target frequency point of 3GHz.

[0022] Secondly, based on the geometric coordinates of the M array elements and the preset target direction, combined with the wavelength of the target frequency point, the M relative phase delays of the M array elements are calculated. Relative phase delay refers to the phase difference between signals from different array elements when they reach the preset target direction; it is a core parameter for calculating beamforming weights. Wavelength refers to the electromagnetic wave wavelength corresponding to the target frequency point, calculated using the formula: wavelength = speed of light / frequency; it is a key physical quantity for calculating phase delay. Based on the geometric coordinates of the M array elements and the preset target direction, combined with the wavelength of the target frequency point, the relative phase delays between the M array elements are calculated using the phase delay calculation formula. For example, the wavelength corresponding to the target frequency point 3GHz is 10cm. The geometric positions of array elements 1 to 4 are (0,0,0), (0,0.1,0), (0,0.2,0), and (0,0.3,0), respectively, and the preset target direction is (0°, 90°). Through the phase delay calculation method, the relative phase delays of each array element are obtained as 0, 2π, 4π, and 6π.

[0023] Next, based on the M relative phase delays, an M-dimensional array manifold vector is constructed. The array manifold vector is a vector reflecting the phase relationship between the positions of the ultra-wideband antenna array elements and the preset target direction, and it forms the basis for constructing the initial beamforming weight vector. The calculated M relative phase delays are converted into phase values ​​and arranged according to the spatial order of the array elements to form the M-dimensional array manifold vector. For example, converting the relative phase delays into phase values ​​and arranging them according to the array element order yields a 4-dimensional array manifold vector of [0, 2π, 4π, 6π].

[0024] Then, the M-dimensional array manifold vector is normalized to obtain the M-dimensional initial beamforming weight vector. Normalization means adjusting the amplitude of the array manifold vector to 1 to ensure the consistency of the amplitude of each element's weight, avoiding beam pointing deviation caused by amplitude differences. The amplitude of the M-dimensional array manifold vector is calculated, and each element in the vector is divided by this amplitude to obtain the normalized M-dimensional initial beamforming weight vector. For example, if the calculated amplitude of the 4-dimensional array manifold vector is 2π√30, dividing each element of the vector by this amplitude yields the normalized M-dimensional initial beamforming weight vectors: 0, 2π / √30, 4π / √30, 6π / √30.

[0025] Subsequently, the N discrete frequency points are traversed to obtain N initial beamforming weight vectors corresponding to the N discrete frequency points. The initial beamforming weight vector is a set of complex coefficients used to control the signal amplitude and phase of each element in the ultra-wideband antenna array at the corresponding frequency point; the dimension of each vector is consistent with the number of elements M. The above operation is repeated for each of the N discrete frequency points to obtain N initial beamforming weight vectors corresponding to the N discrete frequency points. For example, repeating the above steps for 4 frequency points ultimately yields 5 four-dimensional initial beamforming weight vectors.

[0026] Finally, the N initial beamforming weight vectors are arranged column-wise to form an M-row, N-column initial beamforming matrix. The initial beamforming matrix is ​​a matrix composed of N M-dimensional initial beamforming weight vectors arranged column-wise. The rows of the matrix correspond to the elements of the ultra-wideband antenna array, and the columns correspond to discrete frequency points, used to uniformly control the signal amplitude and phase of each element at different frequency points. Using each initial beamforming weight vector as a column of the matrix, all N vectors are arranged sequentially to form an M-row, N-column matrix structure. The i-th row of the matrix corresponds to the weight parameters of the i-th element, and the j-th column corresponds to the weight parameters of the j-th discrete frequency point. For example, based on five 4-dimensional initial beamforming weight vectors, arranging each vector as a column yields a 4-row, 5-column initial beamforming matrix. The first row of the matrix corresponds to the weight parameters of element 1, the second row corresponds to the weight parameters of element 2, and so on; the first column corresponds to the weight parameters of the 3GHz frequency point, the second column corresponds to the weight parameters of the 4.75GHz frequency point, and so on.

[0027] In this embodiment of the invention, an initial beamforming matrix adapted to the operating frequency band is constructed by accurately acquiring the static parameters of the ultra-wideband antenna array. This matrix provides a reliable initial reference for subsequent dynamic beamforming operations, ensuring coordinated control of signal amplitude and phase of each array element at different frequency points. This effectively improves the beam pointing accuracy and frequency adaptability of the ultra-wideband antenna array, laying the foundation for efficient signal transmission and reception in the target scenario.

[0028] S200: Synchronously acquire the M radio frequency analog signals received by the M array elements, perform analog-to-digital conversion on the M radio frequency analog signals, and obtain M discrete time sequence signals.

[0029] In this embodiment of the invention, M radio frequency analog signals received by the M array elements are synchronously acquired, and analog-to-digital conversion is performed on the M radio frequency analog signals to obtain M discrete time sequence signals. The radio frequency analog signals received by the front-end array elements cannot be directly used for subsequent digital signal processing steps such as spatiotemporal joint preprocessing and multipath feature extraction. Synchronous acquisition is necessary to ensure signal timing consistency, and analog signals are converted into digital signals through analog-to-digital conversion. At the same time, interference is suppressed and signal amplitude is conditioned to meet the requirements of subsequent processing. Ultra-wideband signals are characterized by wide bandwidth, susceptibility to high-frequency interference, and significant multipath effects. If the acquisition is not synchronized, it will lead to signal phase deviation between array elements. If anti-aliasing processing is not performed or the sampling frequency is insufficient, signal distortion will be introduced. If the gain is inappropriate, it will lead to signal saturation or excessively low amplitude, all of which will seriously affect the accuracy of the subsequent beam optimization objective function and the final beamforming effect. Therefore, standardized synchronous acquisition, signal conditioning, analog-to-digital conversion, and sequence combination steps are required to provide a high-quality, timing-consistent digital signal source for subsequent processing.

[0030] Step S200 in the method provided in this embodiment of the invention includes: Synchronization trigger signals are sent to the M receiving channels corresponding to the M array elements to obtain M original radio frequency analog signals; The M original radio frequency analog signals are subjected to anti-aliasing filtering and gain conditioning to obtain M radio frequency analog signals; The M radio frequency analog signals are synchronously sampled and quantized at a sampling frequency of not less than 2.2 times the working bandwidth to generate M discrete digital sequences; The M discrete digital sequences are combined according to the array element numbers to output M discrete time sequence signals.

[0031] First, synchronization trigger signals are sent to the M receiving channels corresponding to the M array elements to obtain M original radio frequency analog signals. Receiving channel: A signal path corresponding one-to-one with each array element of the ultra-wideband antenna array, used for transmitting and initially carrying the received signals of the array element, including core components such as signal transmission lines and interface modules.

[0032] Synchronization trigger signal: A reference control signal used to control all receiving channels to start signal acquisition simultaneously, ensuring that the signals acquired by each channel are strictly aligned in the time dimension without timing deviation.

[0033] Raw RF analog signal: The RF band analog signal directly received by the array element without any filtering, gain conditioning or other processing, and contains components such as target signal and environmental interference signal.

[0034] The implementation method involves configuring a synchronous trigger control module (such as a synchronous pulse generator), which establishes signal connections with M receiving channels; generating a timing-precise synchronous trigger signal (such as a digital trigger signal with a pulse width of 1μs and a level of 3.3V) through the trigger control module, and simultaneously sending it to the M receiving channels; after each receiving channel receives the trigger signal, it immediately starts signal acquisition, captures the signal received by the corresponding array element, and outputs M channels of original radio frequency analog signals.

[0035] For example, using the ultra-wideband antenna array (M=4 elements, operating frequency band 3-10GHz) mentioned earlier, each element corresponds to one receiving channel (channel 1-channel 4). A synchronization pulse generator is configured as a trigger control module to generate a synchronization trigger signal (pulse width 1μs, high level 3.3V, low level 0V), which is simultaneously sent to channels 1-4. All four receiving channels start acquisition the instant they receive the trigger signal, respectively capturing the original RF analog signals received by elements 1-4, which include the target signal (pointing to 0°, 90° direction) and environmental interference, ultimately obtaining four channels of original RF analog signals.

[0036] Secondly, the M original RF analog signals are subjected to anti-aliasing filtering and gain conditioning to obtain M RF analog signals. Anti-aliasing filtering refers to the process of suppressing high-frequency interference components in the original signal above a preset cutoff frequency through a filtering circuit, avoiding signal distortion caused by high-frequency signals folding into the baseband during subsequent sampling. Gain conditioning refers to the process of adjusting the amplitude of the original signal through an amplifier, adjusting the signal amplitude to the optimal input range of the subsequent analog-to-digital conversion module, avoiding saturation distortion due to excessively high signal amplitude or excessive quantization error due to excessively low signal amplitude. The RF analog signal refers to the RF band analog signal that has undergone anti-aliasing filtering and gain conditioning to remove high-frequency interference and whose amplitude is adapted to subsequent processing; it is the input signal for analog-to-digital conversion. Each original RF analog signal is configured with an anti-aliasing filter module and a gain conditioning module; the cutoff frequency of the anti-aliasing filter is set according to the operating frequency band of the ultra-wideband antenna array and the subsequent sampling frequency; the anti-aliasing filter removes high-frequency interference components in the original signal that are higher than the cutoff frequency; the gain conditioning module then conditions the amplitude of the filtered signal to the standard input range of the analog-to-digital converter module, such as 0-5V; finally, M conditioning RF analog signals are output.

[0037] For example, based on four original RF analog signals, each signal is connected in series with an anti-aliasing filter and a programmable gain amplifier. The operating frequency band is 3-10GHz, and the subsequent sampling frequency is set to 16GHz. Therefore, the cutoff frequency of the anti-aliasing filter is set to 8GHz to filter out high-frequency interference above 8GHz in the four original signals. The amplitude of each filtered signal is conditioned from the original 0.1-0.5V to 2-4V by the programmable gain amplifier to match the 0-5V input range of the subsequent analog-to-digital conversion module, ultimately obtaining four conditioned RF analog signals.

[0038] Next, the M-channel RF analog signals are synchronously sampled and quantized at a sampling frequency no less than 2.2 times the operating bandwidth to generate M-channel discrete digital sequences. The sampling frequency refers to the number of times the analog-to-digital converter (ADC) samples the analog signal per second. In ultra-wideband scenarios, this must be at least 2.2 times the operating bandwidth; the core purpose is to avoid signal aliasing and ensure signal sampling integrity. Synchronous sampling means that the ADC simultaneously samples the M-channel RF analog signals under a unified sampling clock control, ensuring that the sampled values ​​of each signal at the same time point strictly correspond, maintaining the spatiotemporal consistency of the signal. Quantization is the process of converting the sampled continuous amplitude values ​​into discrete digital quantities; the number of quantization bits determines the precision of the digital quantity. A discrete digital sequence is a sequence composed of a series of discrete digital values ​​obtained after sampling and quantization of each signal; it contains only digital information and has no continuous analog characteristics.

[0039] Specifically, a multi-channel synchronous analog-to-digital converter module is configured, with the number of channels matching the number of array elements M. M conditioned RF analog signals are connected to the module's M input channels. The minimum sampling frequency is calculated based on the operating bandwidth (minimum sampling frequency = operating bandwidth × 2.2), and the module's sampling frequency is set to be no lower than the minimum sampling frequency. Simultaneously, the module's sampling clock is calibrated. Synchronous sampling is initiated, and the module samples the M signals simultaneously under unified clock control, acquiring the continuous amplitude of each signal at different time points. The amplitude of each sampling point is quantized, such as with 12-bit quantization, converting the amplitude into binary numbers from 0 to 4095. After sampling and quantization, each signal generates a discrete digital sequence composed of discrete digital values, ultimately outputting M discrete digital sequences.

[0040] For example, a 4-channel synchronous analog-to-digital converter (12-bit quantization precision) is configured to connect 4 conditioned RF analog signals to the 4 input channels of the module; the sampling frequency is set to 16GHz, and the sampling clock is calibrated to ensure that the clocks of the 4 channels are synchronized; after sampling is started, the module simultaneously samples the 4 signals every 1 / 16GHz to obtain the continuous amplitude of each signal; the amplitude of each sampling point is converted into a binary number from 0 to 4095 through 12-bit quantization, such as the number 2048 corresponding to an amplitude of 3V; finally, the 4 channels generate 4 discrete digital sequences, each sequence containing several binary numbers, corresponding to the signal amplitude at different time points.

[0041] Finally, the M discrete digital sequences are combined according to their element numbers to output M discrete-time sequence signals. The element number is a unique identifier assigned to the M elements of the ultra-wideband antenna array, such as 1, 2, ..., M, used to distinguish different elements and their corresponding signal channels and sequences. The discrete-time sequence signal is a digital signal sequence that associates discrete digital sequences with the time dimension, arranges them in the order of sampling time, and binds them to corresponding element number identifiers. It can intuitively reflect the signal change characteristics of each element at different time points and serves as the direct input for subsequent spatiotemporal joint preprocessing.

[0042] Specifically, the array element numbering rules are defined to clarify the array element number corresponding to each discrete digital sequence; the sampling timestamp of each discrete digital sequence is extracted; each discrete digital sequence is sorted in order from earliest to latest according to the sampling timestamp, and the corresponding array element number is marked; the sorted M discrete digital sequences are combined and encapsulated in order from 1 to M according to the array element number, and finally output M discrete time sequence signals.

[0043] For example, define array element numbers 1 to 4, corresponding to the sequences of channels 1 to 4; extract the sampling timestamp of each sequence, such as starting from 0ps, with a timestamp every 62.5ps: 0ps, 62.5ps, 125ps...; sort each discrete digital sequence by timestamp from earliest to latest, and label them as array element 1 - discrete time sequence, array element 2 - discrete time sequence, array element 3 - discrete time sequence, array element 4 - discrete time sequence; combine these 4 sequences in the order of array element numbers 1 to 4 to output 4 discrete time sequence signals, which can then be directly input into the spatiotemporal domain joint preprocessing stage.

[0044] In this embodiment of the invention, synchronous triggering acquisition ensures the timing consistency of the M-channel signals and avoids signal phase deviation between array elements; anti-aliasing filtering and gain conditioning effectively suppress high-frequency interference and optimize signal amplitude, providing high-quality input for analog-to-digital conversion; synchronous sampling and quantization with a working bandwidth of more than 2.2 times ensure signal integrity and digital accuracy; the discrete time sequence signal output by combining array element numbers accurately matches the input requirements of subsequent spatiotemporal joint preprocessing and multipath signal feature extraction, laying a reliable signal foundation for the efficient advancement of the entire beamforming process, while improving the accuracy of subsequent beam optimization objective function construction and the overall stability of beamforming.

[0045] S300: Perform spatiotemporal joint preprocessing on the M discrete time series signals, extract multipath signal feature parameters, and generate a beam optimization objective function in combination with preset communication quality indicators.

[0046] In this embodiment of the invention, the M discrete-time series signals undergo joint spatiotemporal preprocessing to extract multipath signal feature parameters, and a beam optimization objective function is generated by combining these parameters with preset communication quality indicators. In ultra-wideband communication scenarios, the M discrete-time series signals contain target signals, multipath interference signals, and environmental noise. Directly using these signals for beamforming matrix optimization can lead to optimization direction deviation and fail to meet preset communication quality indicators. Joint spatiotemporal preprocessing can achieve spatiotemporal noise reduction and normalization of the signals. Multipath signal feature extraction can accurately distinguish the propagation characteristics of target signals and interference signals. The beam optimization objective function is the core guide for subsequent particle swarm optimization algorithms. Only by clearly defining the optimization objective of maximizing the signal-to-interference-plus-noise ratio (SINR) can the beamforming matrix be adjusted in a targeted manner to ensure that the final beam main lobe points to the desired signal and suppresses interference and noise. Furthermore, the array response model, as the core model connecting the geometric position of array elements and signal propagation characteristics, directly determines the reliability of the objective function. If preprocessing or feature extraction is lacking, or if the objective function is poorly constructed, subsequent iterative optimization will converge slowly, the beam pointing will deviate, and ultimately affect the formation effect of the focused beam. Therefore, S300 is a key link between signal preprocessing and matrix optimization, providing accurate optimization basis for efficient beamforming.

[0047] Step S300 in the method provided in this embodiment of the invention includes: The M discrete time series signals are divided into multiple spatiotemporal data blocks, where each spatiotemporal data block has a dimension of M×L, and L is the number of snapshots within a preset time window; Spatiotemporal joint parameter estimation is performed on multiple spatiotemporal data blocks to extract the multipath signal feature parameters, wherein the multipath signal feature parameters include the direction of arrival parameters, relative time delay parameters, and complex path gain feature parameters of the main multipath components; Based on the preset communication quality index, the multipath signal characteristic parameters are classified into desired signal paths and interference paths, and corresponding optimization weights are assigned to the desired signal paths and the interference paths. Based on the array response model, the multipath signal characteristic parameters, and the optimization weights, a beam optimization objective function is constructed with the beamforming matrix as the variable, wherein the objective function is used to maximize the signal-to-interference-plus-noise ratio.

[0048] The array response model is determined based on the geometric position of the ultra-wideband antenna array.

[0049] First, the M discrete-time series signals are divided into multiple spatiotemporal data blocks, each with a dimension of M×L, where L is the number of snapshots within a preset time window. A spatiotemporal data block is a signal data unit that integrates spatial and temporal dimensions, with a dimension of M×L, where M is the number of array elements and L is the number of snapshots within the preset time window. This process breaks down the continuous time series signal into regular processing units, facilitating subsequent spatiotemporal joint parameter estimation. The number of snapshots (L) refers to the number of signal sampling points per array element within the preset time window. The number of snapshots needs to balance processing accuracy and efficiency; too large a L results in long processing times, while too small a L leads to large parameter estimation errors. Set a fixed time window, such as 1 μs, and calculate the number of sampling points within this time window based on the sampling frequency of S200, which is taken as the snapshot number L. Divide the M discrete time series signals into segments according to their chronological order, using the time window as the unit. Each segment contains signal data of M array elements at L sampling points. Arrange each segment of data according to the rule that rows correspond to array elements and columns correspond to sampling points, forming a spatiotemporal data block with a dimension of M×L. Repeat the segmentation process until all M discrete time series signals are covered, resulting in multiple spatiotemporal data blocks.

[0050] For example, the time window is set to 1μs, and the number of snapshots is calculated as L=1μs / 62.5ps=16000, meaning that each time window contains 16000 sampling points. The four discrete time series signals output by S200 are divided into segments every 1μs according to the time sequence. Each segment of data is arranged with rows 1 to 4 corresponding to array elements 1 to 4, and columns 1 to 16000 corresponding to sampling points 1 to 16000, forming a spatiotemporal data block with a dimension of 4×16000. If the total signal duration is 100μs, a total of 100 spatiotemporal data blocks are obtained, which are used for subsequent parameter estimation.

[0051] Secondly, spatiotemporal joint parameter estimation is performed on multiple spatiotemporal data blocks to extract the multipath signal characteristic parameters. These multipath signal characteristic parameters include the direction of arrival (DOA) parameter, relative delay parameter, and complex path gain characteristic parameter of the main multipath components. Spatiotemporal joint parameter estimation refers to the process of accurately estimating the key parameters of the multipath signal by combining the spatial and temporal characteristics of the signal. This differs from estimation based on a single spatial or temporal dimension and improves the accuracy of parameter estimation. The DOA parameter refers to the spatial angle, such as azimuth and elevation angle, when the multipath signal arrives at the ultra-wideband antenna array; it is a core spatial feature distinguishing different multipath signals. The relative delay parameter refers to the time difference between the arrival times of different multipath signals at the array; it is a temporal characteristic distinguishing multipath signals. The complex path gain characteristic parameter is a complex parameter characterizing the amplitude attenuation and phase shift of the multipath signal in its propagation path. The real part corresponds to amplitude attenuation, and the imaginary part corresponds to phase shift, reflecting the propagation loss characteristics of the multipath signal.

[0052] Specifically, a two-dimensional spatiotemporal parameter estimation algorithm, such as a multiple signal classification algorithm or a rotation-invariant subspace algorithm, is used to jointly process multiple spatiotemporal data blocks. First, each spatiotemporal data block is preprocessed for noise reduction, such as moving average noise reduction, to reduce the impact of environmental noise on parameter estimation. Based on the denoised spatiotemporal data blocks, the direction of arrival parameters and relative time delay parameters of each multipath signal are estimated. Combining signal amplitude and phase information, the complex path gain characteristic parameters of each multipath signal are calculated. All estimation results are summarized, and the characteristic parameters of the main multipath components are extracted to form a set of multipath signal characteristic parameters.

[0053] For example, based on 100 4×16000 spatiotemporal data blocks, a multi-signal classification algorithm is used for spatiotemporal joint parameter estimation. First, moving average noise reduction is performed on each data block, with a window size of 10 sampling points. Through spatial spectrum estimation, the direction of arrival parameters of the three main multipath components are determined: multipath 1 (0°, 90°, consistent with the preset target direction), multipath 2 (30°, 90°), and multipath 3 (-30°, 90°). Using multipath 1 as a reference, the relative time delay of multipath 2 is estimated to be 0.2μs, and the relative time delay of multipath 3 is estimated to be 0.3μs. Combining the signal amplitude and phase, the complex path gain characteristic parameters of the three types of multipaths are calculated as follows: multipath 1 (0.8+j0.2), multipath 2 (0.3+j0.1), and multipath 3 (0.2+j0.05). Finally, the three sets of parameters are extracted as the main multipath signal characteristic parameters.

[0054] Next, based on the preset communication quality indicators, the multipath signal characteristic parameters are classified into desired signal paths and interference paths, and corresponding optimization weights are assigned to the desired signal paths and the interference paths. The preset communication quality indicators refer to pre-defined standards for evaluating communication performance, including signal-to-interference-plus-noise ratio (SINR) thresholds and bit error rate (BER) thresholds, used to distinguish between desired signal paths and interference paths. A desired signal path is a multipath signal path that meets the preset communication quality indicators and carries the target communication information; typically, its direction of arrival is consistent with the preset target direction, and its complex path gain amplitude is large. An interference path is a multipath signal path that does not meet the preset communication quality indicators and interferes with the target communication; its direction of arrival deviates from the preset target direction, weakening the reception quality of the desired signal. Optimization weights are weight coefficients assigned to each path, used to distinguish the priority of desired signals and interference signals in the objective function; desired signal paths have higher weights, and interference paths have lower weights, achieving an optimization orientation that strengthens the desired signal and suppresses interference.

[0055] Specifically, define the preset communication quality indicators, such as signal-to-interference-plus-noise ratio ≥20dB and complex path gain amplitude ≥0.5; compare the multipath signal characteristic parameters with the preset indicators, and select multipath signal paths that meet the indicators as desired signal paths; identify multipath signal paths that do not meet the indicators as interference paths; set optimization weight allocation rules based on path priority: assign higher positive optimization weights (e.g., 0.8~1.0) to desired signal paths and lower negative optimization weights (e.g., 0.1~0.3) to interference paths, and adjust the total weights according to the objective function construction requirements; record the classification results and corresponding optimization weights for each path.

[0056] For example, preset communication quality indicators are set: the deviation of the direction of arrival from the preset target direction (0°, 90°) is ≤5°, and the complex path gain amplitude is ≥0.5. The three multipath components are compared: Multipath 1: direction of arrival 0°, 90°, deviation 0°; complex path gain amplitude √(0.8²+0.2²)≈0.82≥0.5, which meets the indicators and is determined to be the desired signal path, and is assigned an optimization weight of 0.9; Multipath 2 and Multipath 3 have deviations of 30°, which do not meet the indicators and are determined to be interference paths, and are assigned optimization weights of 0.2 and 0.1 respectively; the final classification result is: 1 desired path (weight 0.9) and 2 interference paths (weights 0.2 and 0.1), which provides a weight basis for the subsequent construction of the objective function.

[0057] Furthermore, based on the array response model, the multipath signal characteristic parameters, and the optimization weights, a beam optimization objective function is constructed with the beamforming matrix as the variable. This objective function maximizes the signal-to-interference-plus-noise ratio (SINR). The array response model is determined based on the geometric position of the ultra-wideband antenna array. The array response model is a mathematical model describing the response characteristics of an ultra-wideband antenna array to signals from different directions of arrival. It is constructed based on the geometric positions of the array elements and reflects the correlation between element positions and signal phase and amplitude. The beam optimization objective function is a mathematical expression that characterizes communication performance by using the beamforming matrix as a variable and integrating multipath signal characteristic parameters and optimization weights. It is the optimization criterion of the particle swarm optimization algorithm, and its objective is to maximize the function value. The SINR is the ratio of the desired signal power to the sum of the interference signal power and noise power. It is a core indicator for measuring communication quality; a higher SINR indicates better communication performance.

[0058] Specifically, firstly, based on the geometric positions of the M elements of the ultra-wideband antenna array, an array response model is constructed, including parameters such as element coordinates, signal wavelength, and direction of arrival, reflecting the response of signals with different directions of arrival on each element. Then, multipath signal characteristic parameters are substituted into the array response model to calculate the array response vectors corresponding to the desired signal path and the interference path, respectively. Combining optimization weights, a signal-to-interference-plus-noise ratio (SIR) expression is constructed with the beamforming matrix as the variable, multiplying the desired signal power term by a positive weight and the interference signal power term by a negative weight. This SIR expression is defined as the beam optimization objective function, with the optimization direction of the objective function being to maximize the function value, i.e., to maximize the SIR.

[0059] For example, M = 4 array elements with coordinates (0,0,0), (0,0.1,0), (0,0.2,0), and (0,0.3,0). An array response model is constructed based on these geometric positions, including the relationship between the coordinates of each element and the signal direction of arrival and wavelength. Substituting the multipath characteristic parameters into the model, the array response vector for the desired path is obtained as [1, e^(j2π), e^(j4π), e^(j6π)]. The array response vectors for interference path 1 (30°, 90°) and interference path 2 (-30°, 90°) are also shown. The quantities are [1, e^(jπ), e^(j2π), e^(j3π)] and [1, e^(jπ / 2), e^(jπ), e^(j3π / 2)], respectively. Combining the optimization weights, a beam optimization objective function is constructed: objective function value = 0.9 × desired signal power - 0.2 × interference path 1 power - 0.1 × interference path 2 power, where the signal power is calculated by multiplying the beamforming matrix and the corresponding array response vector. The core of this objective function is to maximize the function value by adjusting the beamforming matrix, that is, to maximize the signal-to-interference-plus-noise ratio.

[0060] In this embodiment of the invention, signal normalization is achieved through spatiotemporal data block partitioning, providing reliable data units for parameter estimation; spatiotemporal joint parameter estimation accurately extracts key features of multipath signals, laying the foundation for path classification; the classification and weight allocation of desired and interfering paths clarify the optimization orientation of strengthening the desired path and suppressing the interference; the objective function for maximizing the signal-to-interference-plus-noise ratio (SINR) constructed based on the array response model provides a clear and reliable optimization criterion for subsequent particle swarm optimization algorithms. This effectively weakens the impact of noise and interference, improves the targeting and efficiency of beamforming optimization, and provides support for obtaining the optimized beamforming matrix and forming a focused beam.

[0061] S400: Using the particle swarm optimization algorithm, with the beam optimization objective function as the optimization target, the initial beamforming matrix is ​​iteratively optimized to obtain the optimized beamforming matrix.

[0062] In this embodiment of the invention, a particle swarm optimization algorithm is employed. Using the beamforming objective function as the optimization target, the initial beamforming matrix is ​​iteratively optimized to obtain the optimized beamforming matrix. S300 has already constructed a beamforming objective function with maximizing the signal-to-interference-plus-noise ratio (SINR) as its core. However, the initial beamforming matrix obtained in S100 is only a baseline matrix based on static parameters, failing to consider dynamic factors such as multipath interference and environmental noise in real-world scenarios, and thus cannot directly achieve the optimal beamforming effect. Particle swarm optimization is a highly efficient swarm intelligence optimization algorithm suitable for handling optimization problems with high-dimensional parameters such as beamforming matrices. Iterative search by a swarm of particles can quickly approximate the global optimum. Furthermore, addressing the problems of high particle similarity, low search efficiency, and premature convergence inherent in traditional particle swarm optimization, this step adds particle similarity judgment and cooperation strategies, discarding redundant particles and enhancing the search capability of high-quality particles, further improving optimization efficiency and global optimality.

[0063] Step S400 in the method provided in this embodiment of the invention includes: The initial beamforming matrix is ​​used as an initial particle, and multiple beamforming matrices are randomly generated as an initial particle swarm, where each particle represents an M-row N-column beamforming matrix. The function value of the beam optimization objective function is defined as the fitness of the particle; Iterative optimization is performed based on the particle swarm optimization algorithm, updating the velocity and position of each particle in each iteration; During the iteration process, the similarity between particles is calculated. For particle pairs with similarity exceeding a preset similarity threshold, a cooperative strategy is implemented: particles with lower fitness are discarded, and optimization privileges are granted to the retained particles. The optimization privileges include increasing the number of derived new solutions of the retained particles in subsequent iterations. When the preset convergence condition is met, the beamforming matrix corresponding to the globally optimal particle is output as the optimized beamforming matrix.

[0064] First, the initial beamforming matrix is ​​used as an initial particle, and multiple beamforming matrices are randomly generated as an initial particle swarm. Each particle represents an M-row, N-column beamforming matrix. Particle: In this step, the particle is the basic search unit of the algorithm. Each particle uniquely corresponds to an M-row, N-column beamforming matrix, and all elements of the matrix are the particle's characteristic parameters. The particle search process is the beamforming matrix parameter adjustment process. Initial particle: This refers to the particle directly used as the initial beamforming matrix obtained from S100. It is the baseline search starting point of the algorithm and possesses basic beam pointing capability. Particle swarm: This refers to a group search set composed of multiple particles. Through the cooperative iteration of particles within the swarm, the optimal solution is searched simultaneously from multiple starting points, avoiding the problem of getting trapped in local optima by a single search.

[0065] Specifically, the size of the particle swarm needs to be determined, taking into account both search efficiency and computational cost, typically 10-50 particles; the M x N initial beamforming matrix output by S100 is added to the particle swarm as one initial particle; based on the parameter constraints of the ultra-wideband antenna array, within a reasonable range of complex coefficients, an M x N beamforming matrix with the same dimension as the initial particle is randomly generated as the remaining particles; the initial particles and the randomly generated particles are integrated to form a complete initial particle swarm, completing the search initialization of the algorithm.

[0066] For example, M=4, N=5, the beamforming matrix dimension is 4×5; the particle swarm size is set to 20 particles; the initial 4×5 beamforming matrix obtained by S100 is used as particle 1; within the complex coefficient range [-1+j1, 1-j1], 19 beamforming matrices with a dimension of 4×5 are randomly generated as particles 2-20; particles 1-20 are integrated to form an initial particle swarm containing 20 4×5 beamforming matrices, preparing for subsequent iterative optimization.

[0067] Secondly, the function value of the beam optimization objective function is defined as the particle's fitness. Fitness is an indicator used to evaluate the quality of particles; a higher fitness value indicates that the beamforming matrix corresponding to the particle better satisfies the optimization objective of maximizing the signal-to-interference-plus-noise ratio (SIR), resulting in better beamforming performance. The objective function value refers to the numerical value calculated by substituting the beamforming matrix into the beam optimization objective function, directly reflecting the SIR level corresponding to that matrix. The unique calculation rule for determining fitness is: particle fitness = function value obtained by substituting the beamforming matrix corresponding to the particle into the beam optimization objective function; each particle in the initial particle swarm is then substituted into the beam optimization objective function, and the objective function value of each particle is calculated one by one; the calculation result is directly assigned as the fitness of the corresponding particle, completing the initial evaluation of the quality of all particles.

[0068] For example, based on 20 initial particles, each particle is substituted into the beam optimization objective function in turn; 20 objective function values ​​are calculated and directly used as the fitness of each particle: the fitness of particle 1 is 85, corresponding to a signal-to-interference-plus-noise ratio of 85dB, the fitness of particles 2 to 20 is between 42 and 80, and no particle exceeds the fitness of particle 1. At this time, particle 1 is the optimal particle in the initial particle swarm.

[0069] Furthermore, iterative optimization is performed based on the particle swarm optimization algorithm, updating the velocity and position of each particle in each iteration. Iterative optimization refers to the process of gradually approaching the global optimum through repeated parameter adjustments and performance evaluations; each repetition constitutes one iteration. Particle velocity is a variable characterizing the magnitude and direction of parameter adjustments in the search space, consistent with the dimension of the beamforming matrix, and its value determines the adjustment amount of matrix elements. Particle position refers to the coordinates of the particle in the search space, uniquely corresponding to all elements of the beamforming matrix. Updating the particle position is equivalent to adjusting the parameters of the beamforming matrix, forming a new beamforming matrix.

[0070] Specifically, the basic parameters of the particle swarm optimization algorithm are set, including the maximum number of iterations, velocity weights, and learning factors. In each iteration, the individual optimal position of each particle is first recorded, which is the position / beamforming matrix with the highest fitness of the particle so far. The global optimal position is also recorded in the particle swarm, which is the position / beamforming matrix with the highest fitness of all particles so far. According to the classic velocity-position update formula of particle swarm optimization, the new velocity of each particle is calculated by combining the individual optimal position, the global optimal position, and the current velocity. The current position of each particle is adjusted based on the new velocity to obtain the new position, which is the new M-row N-column beamforming matrix. The new position is used as the current particle, and the velocity and position update of one iteration is completed.

[0071] For example, the algorithm's basic parameters are set as follows: maximum number of iterations 100, velocity weight 0.729, and learning factor 1.494. Iterative optimization is initiated for 20 particles: In the first iteration, based on fitness, particle 1 is the individual optimum and the global optimum. The new velocities of the 20 particles are calculated according to the update formula, and 20 new 4×5 beamforming matrices are obtained based on the new velocities. The fitness is recalculated, and the new fitness of particle 12 is increased to 88, becoming the new global optimum. Each particle updates its own individual optimum. In the second to tenth iterations, the velocity is updated, the position is adjusted, the fitness is calculated, and the individual global optimum is updated. The global optimum fitness is gradually increased to 92, and no particle gets stuck in a local optimum.

[0072] Then, during the iteration process, the similarity between particles is calculated. For particle pairs with a similarity exceeding a preset similarity threshold, a cooperative strategy is implemented: particles with lower fitness are discarded, and the retained particles are granted optimization privileges. These optimization privileges include increasing the number of derived new solutions for the retained particles in subsequent iterations. Particle similarity characterizes the degree of similarity between two beamforming matrices, quantified by the similarity calculation method of matrix elements. A higher value indicates that the parameters of the two matrices are closer, and the search directions are more redundant. The preset similarity threshold is a critical value for judging whether a particle pair is redundant. It is set based on engineering experience. If the value is exceeded, the two particles are considered to have duplicate search directions, and the cooperative strategy must be implemented. The cooperative strategy is an optimization strategy for redundant particle pairs, discarding redundant particles with lower fitness, reducing invalid searches, and improving algorithm efficiency. Optimization privileges refer to granting enhanced search rights to the retained high-quality particles, specifically increasing the number of derived new solutions in subsequent iterations, allowing high-quality particles to explore the optimal solution more fully in their search directions.

[0073] Specifically, a suitable matrix similarity calculation method is selected, such as cosine similarity or Euclidean distance similarity. After each iteration's position update, all particle pairs in the particle swarm are traversed, and their similarity is calculated one by one. The calculation results are compared with a preset similarity threshold, and redundant particle pairs with similarity exceeding the threshold are filtered out. For each redundant particle pair, their fitness is compared, and particles with lower fitness are discarded, while only particles with higher fitness are retained. The retained high-quality particles are given optimization privileges: in the velocity-position update stage of subsequent iterations, the number of new solutions derived from them is increased from the default value to a set value, thus completing the execution of the cooperation strategy.

[0074] For example, the parameters are set as follows: cosine similarity is used to calculate particle similarity, the preset similarity threshold is 0.85, the default number of derived new solutions is 1, and the number of derived new solutions with optimization privilege is 3. After the 5th iteration: all particle pairs of 20 particles are traversed, and it is found that the cosine similarity between particle 3 and particle 12 is 0.88 (exceeding the threshold of 0.85), which is a redundant particle pair. The fitness is compared: the fitness of particle 3 is 72, and the fitness of particle 12 (currently the global best) is 90. Particle 3 is discarded, and the particle swarm size is temporarily changed to 19. Privilege is granted: particle 12 is granted optimization privilege, and the number of derived new solutions in subsequent iterations increases from 1 to 3, allowing particle 12 to explore the optimal solution more fully in its search direction.

[0075] Finally, when the preset convergence condition is met, the beamforming matrix corresponding to the globally optimal particle is output as the optimized beamforming matrix. The preset convergence condition is the criterion for stopping the algorithm's iteration; it must balance optimization accuracy and computational efficiency. If this condition is met, the globally optimal solution has been found, and further iteration is unnecessary. The globally optimal particle refers to the particle whose fitness remains the highest throughout the iteration process; its corresponding beamforming matrix is ​​the currently searched optimal matrix, achieving the maximum signal-to-interference-plus-noise ratio (SNR). The optimized beamforming matrix refers to the M-row, N-column beamforming matrix corresponding to the globally optimal particle when the preset convergence condition is met, providing the optimal weights for subsequent beam pattern adjustments.

[0076] Specifically, a preset convergence condition is set, typically when the rate of change of the global optimal fitness is less than a preset threshold after multiple consecutive iterations, or when the number of iterations reaches the maximum number of iterations. After each iteration, the preset convergence condition is checked: if it is not met, the next iteration continues; if it is met, the iteration is stopped immediately, and the global optimal particle in the current particle swarm is extracted. The M-row N-column beamforming matrix corresponding to the global optimal particle is output as the optimized beamforming matrix.

[0077] For example, a preset convergence condition is set: the rate of change of the global optimal fitness is less than 0.1% for 10 consecutive iterations; in the 25th iteration, the global optimal fitness is 98.2, and in the subsequent 26th to 34th iterations, the global optimal fitness fluctuates only slightly between 98.2 and 98.5, with a rate of change of less than 0.1%, satisfying the preset convergence condition; iteration stops: the current global optimal particle is extracted, which is particle 12, which was given optimization privilege in the 5th iteration. After multiple derivations of new solutions, it is optimized, and its corresponding 4-row 5-column beamforming matrix is ​​the optimized beamforming matrix. The signal-to-interference-plus-noise ratio of this matrix reaches 98.5dB, which is much higher than the initial matrix's 85dB.

[0078] In this embodiment of the invention, a particle swarm is constructed based on an initial beamforming matrix, avoiding the blindness of random initialization and allowing the algorithm to start searching from a starting point with basic beam pointing capability, reducing the number of iterations. Simultaneously, through particle similarity judgment and cooperation strategies, redundant particles are eliminated and the search capability of high-quality particles is enhanced, effectively solving the problems of premature convergence and low search efficiency in traditional particle swarm optimization, ensuring that the algorithm can approach the global optimum. The optimized beamforming matrix incorporates the multipath signal characteristics of the actual scene, no longer relying solely on static parameters, and can accurately match the communication requirements of the target scene, achieving the goal of strengthening the desired signal and suppressing interference and noise. The output M-row N-column optimized beamforming matrix directly matches the number of elements and discrete frequency points of the ultra-wideband antenna array, and can be directly used for subsequent dynamic adjustment of the radiation pattern, providing optimal weight control parameters for forming a focused beam pointing towards the target receiver, fundamentally guaranteeing the final beamforming effect.

[0079] S500: Based on the optimized beamforming matrix, dynamically adjust the radiation pattern of the ultra-wideband antenna array to form a focused beam pointing towards the target receiver.

[0080] In this embodiment of the invention, the radiation pattern of the ultra-wideband antenna array is dynamically adjusted based on the optimized beamforming matrix to form a focused beam pointing towards the target receiver. In ultra-wideband communication scenarios, signal propagation at the target receiver suffers from multipath attenuation and is accompanied by environmental interference. If the optimized matrix is ​​directly applied to the array elements without standardized analysis and configuration, signal coordination between array elements will fail, making it impossible to achieve precise pointing of the main lobe of the beam. Furthermore, if weighting is not performed according to the transmit / receive modes, signal transmission distortion will occur, reducing communication quality. Therefore, standardized weighting analysis and configuration, mode-specific signal weighting, and dynamic adjustment of the radiation pattern are necessary to ensure that the radiation characteristics of the ultra-wideband antenna array accurately match the control requirements of the optimized matrix, ultimately forming a focused beam pointing towards the target receiver, achieving beamforming that enhances the target signal and suppresses interference.

[0081] Step S500 in the method provided in this embodiment of the invention includes: The optimized beamforming matrix with M rows and N columns is analyzed into M sets of weights, and the M sets of weights are respectively assigned to M array element channels; If in signal transmission mode, one signal to be transmitted will be allocated to M array element channels according to the M sets of weights for weighted transmission; If in signal receiving mode, the signals received by the M array element channels are weighted and summed according to the M sets of weights to form a single output signal; Based on the optimized beamforming matrix, the radiation pattern of the ultra-wideband antenna array is dynamically adjusted so that the main lobe of the radiation pattern continuously points to the target receiver and forms a focused beam.

[0082] First, parse the optimized M-row N-column beamforming matrix into M groups of weights, and configure the M groups of weights to M array element channels respectively. Weight parsing refers to the process of splitting the optimized M-row N-column beamforming matrix by rows, and extracting the complex coefficients corresponding to all discrete frequency points for each array element. Each group of weights after splitting corresponds to an array element one to one. The M groups of weights refer to a set of weights obtained by parsing the optimized beamforming matrix and matching the M array elements one to one. Each group of weights contains N complex coefficients, each of which carries amplitude and phase control information, and is used to control the signal characteristics of the corresponding array element at different frequency points. An array element channel is a signal path matched with an array element that has both signal transmitting and receiving functions. It can receive weight configuration instructions and accurately adjust the amplitude and phase of the signal according to the weights.

[0083] Specifically, a weight parsing and configuration module is provided, and the module establishes a two-way communication connection with M array element channels; the optimized M-row N-column beamforming matrix output by S400 is input to the parsing module, and according to the rule that matrix rows correspond to array elements and in-row elements correspond to discrete frequency points, the matrix is split row by row, and M groups of weights are obtained after parsing, where the i-th row is parsed into the weight of the i-th array element, and each group contains N complex coefficients; the parsing module distributes the M groups of weights to the corresponding M array element channels respectively through a communication link. After each array element channel receives the weights, it writes the weights into the local weight control unit, completes the solidified configuration of the weights, and prepares for subsequent signal weighting.

[0084] For example, the optimized beamforming matrix output by S400 is a 4-row 5-column complex matrix; the matrix is input to the weight parsing module, and split into 4 groups of weights by rows, each group contains 5 complex coefficients: Weight group for array element 1: [0.9+j0.1, 0.85+j0.12, 0.8+j0.15, 0.75+j0.18, 0.7+j0.2]; Weight group for array element 2: [0.88+j0.11, 0.83+j0.13, 0.78+j0.16, 0.73+j0.19, 0.68+j0.21]; Weight group for array element 3: [0.85+j0.13, 0.8+j0.15, 0.75+j0.18, 0.7+j0.2, 0.65+j0.22]; Weight group for array element 4: [0.82+j0.15, 0.77+j0.17, 0.72+j0.2, 0.67+j0.22, 0.62+j0.24]. The parsing module distributes the 4 groups of weights to the corresponding channels of array elements 1 to 4 respectively, and each channel completes the solidified configuration of the weights.

[0085] Secondly, if in signal transmission mode, one signal to be transmitted is allocated to M array element channels according to the M sets of weights for weighted transmission. Signal transmission mode refers to the operating mode of the ultra-wideband antenna array as a signal transmitter, transmitting ultra-wideband communication signals to the target receiver, converting a single baseband signal into multiple coordinated radio frequency transmission signals. The signal to be transmitted refers to the original ultra-wideband signal to be transmitted to the target receiver, covering the entire operating frequency band of the array, without any amplitude or phase modulation. Weighted transmission refers to the process of splitting the single signal to be transmitted, using the M sets of weights for each array element channel to perform amplitude scaling and phase shifting on the split signals respectively, and then having the corresponding array elements transmit synchronously. The weighted signals can achieve spatial coordinated superposition.

[0086] Specifically, the operating mode of the ultra-wideband antenna array is detected. If it is determined to be in signal transmission mode, the transmitter signal processing flow is initiated. A single ultra-wideband signal covering the operating frequency band is input to the array's signal splitting module. The splitting module divides the single signal into M equal paths, which are then transmitted to M array element channels. Each array element channel calls its locally configured weighting group to perform frequency matching weighting on the received split signals. For N discrete frequency points in the signal, the corresponding complex coefficients in the weighting group are used to scale the amplitude and shift the phase of the signal at each frequency point. The M weighted radio frequency signals are then synchronously transmitted by the M array elements according to a unified timing sequence, achieving spatial coordinated transmission of the signal.

[0087] For example, the array is determined to be in transmit mode, and the signal to be transmitted is an ultra-wideband communication signal of 3-10GHz. The splitting module divides the signal into four equal paths, which are transmitted to the channels of array elements 1-4 respectively. Each channel calls the configured weight group to perform frequency matching weighting: for the split signal of array element 1 channel, the 3GHz frequency component is amplified to 0.9 times and phase shifted by 10° according to the weight 0.9+j0.1, and the 4.75GHz component is amplified to 0.85 times and phase shifted by 12° according to the weight 0.85+j0.12. The remaining frequency points are weighted in sequence. Array elements 2-4 channels use their respective weight groups to perform amplitude and phase modulation on the five discrete frequency points of the split signal according to the above rules. Finally, the four array elements synchronously transmit the four weighted radio frequency signals, and the signals are superimposed in space along the preset target direction (0°, 90°).

[0088] Secondly, in signal reception mode, the signals received by the M array element channels are weighted and summed according to the M sets of weights to synthesize a single output signal. Signal reception mode refers to the working mode where the ultra-wideband antenna array acts as a signal receiver, capturing the ultra-wideband signal transmitted by the target receiver and suppressing interference signals. The core is to convert the multi-element received signals into a single high signal-to-interference-plus-noise ratio (SNR) output signal. Weighted summation refers to the process where the M array element channels separately weight the received multiple signals by amplitude and phase, and then combine all the weighted signals into a single signal through a combining module. This achieves the superposition and enhancement of the desired signal while suppressing the cancellation of interference signals. The combined output signal is the single ultra-wideband signal obtained after weighted summation of the multiple received signals. It retains the desired signal in the target direction, weakens interference signals and noise, and possesses high SNR characteristics, allowing it to be directly input into subsequent communication signal processing stages.

[0089] Specifically, the operating mode of the ultra-wideband antenna array is detected. If it is determined to be in signal reception mode, the receiver signal processing flow is initiated. M array elements synchronously capture radio frequency signals in space, which are then converted into baseband signals by the corresponding M array element channels, resulting in M ​​received signals. Each array element channel calls its locally configured weight group to perform frequency matching weighting on the received baseband signals, consistent with the weighting rules of the transmission mode, and performs amplitude and phase modulation at N discrete frequency points respectively. The weighted M signals are then transmitted to the array's signal combining module, which sums the M signals in a linear superposition manner to obtain a combined signal. After amplitude conditioning and filtering, the combined signal is output as the final combined output signal, completing signal reception and optimization.

[0090] For example, when the array is in receive mode, the four array elements synchronously capture radio frequency signals in the space and convert them into four baseband receive signals through the array element channels. Each channel calls the weight group to weight the received signal: the array element 1 channel weights the 3GHz component of the received signal by 0.9 + j0.1 to enhance the target direction signal and weaken the interference signal, and the other frequency points are processed in sequence; the array elements 2 to 4 complete the weighting of the corresponding frequency points according to their respective weight groups, so that the desired signals in the four signals are in the same phase and have superimposed amplitudes, while the interference signals have opposite phases and cancel each other out in amplitude; the combining module linearly superimposes the four weighted signals to obtain a combined output signal. The signal-to-interference-plus-noise ratio of this signal is maintained at 98.5dB, which is consistent with the target value of the S400 optimization matrix, and the interference signal is suppressed by more than 90%.

[0091] Furthermore, based on the optimized beamforming matrix, the radiation pattern of the ultra-wideband antenna array is dynamically adjusted so that the main lobe of the radiation pattern continuously points towards the target receiver and forms a focused beam. The radiation pattern is a graphical representation of the signal radiation / reception gain of the ultra-wideband antenna array in different spatial directions, comprising the main lobe, side lobes, and nulls. The main lobe is the direction with the highest gain, i.e., the main direction of the beam; the side lobes refer to the small gain regions outside the main lobe; and the nulls are the directions with zero gain, used to suppress interference. Dynamic adjustment refers to the real-time fine-tuning of the signal amplitude and phase of the M array elements based on the weight control of the optimized beamforming matrix, ensuring that the main lobe of the array's radiation pattern continuously tracks the target receiver and the nulls continuously align with the interference direction, unlike the fixed pointing of a static beam. A focused beam refers to a beamform with a narrow main lobe, high gain, and the main lobe precisely pointing towards the target receiver while the nulls align with the interference direction. This allows signal energy to be highly focused in the target direction, increasing the signal strength in that direction while effectively suppressing interference and noise.

[0092] Specifically, a radiation pattern monitoring and adjustment module is configured. This module collects the radiation pattern data of the array in real time and compares it with the ideal radiation pattern corresponding to the optimized beamforming matrix. Based on the weighted transmit / receive status of the M array elements and combined with the spatial propagation characteristics of electromagnetic waves, the array forms an initial radiation pattern, with its main lobe initially pointing towards the target receiver and the nulls aligned with the interference path direction. The monitoring module detects the deviation between the actual main lobe and the target receiver, as well as the matching degree between the nulls and the interference direction in real time. If the deviation exceeds a preset threshold, the signal amplitude and phase of each array element are dynamically adjusted by fine-tuning the weight coefficients of the array element channels. Through continuous monitoring and fine-tuning, the main lobe of the radiation pattern is continuously and accurately pointed towards the target receiver, the sidelobe gain is reduced to a minimum, and the nulls are always aligned with the interference direction, ultimately forming a highly focused beam in the target direction.

[0093] For example, based on the weighted transmit and receive states of M=4 array elements, the array initially forms a radiation pattern, with the main lobe initially pointing at 0° and 90° (target receiver direction), and the nulls aligned at 30° and -30° (interference direction). The monitoring module collects the radiation pattern data in real time. If, due to environmental changes, the main lobe pointing shifts to 2° or 90°, and the deviation exceeds the preset 1° threshold, the adjustment module makes a small correction to the weight coefficients of array elements 2-4 (e.g., the 3GHz weight of array element 2 is slightly adjusted from 0.88+j0.11 to...). (0.89+j0.1), fine-tuning the signal phase of each array element; after weight fine-tuning, the main lobe of the array's radiation pattern returns to the 0° and 90° directions, and the main lobe width is compressed from the initial 15° to 8°, the signal gain in the target direction is increased to 98.5dB, and the gain in the 30° / -30° interference direction is reduced to -20dB; through continuous monitoring and fine-tuning, the array forms a focused beam with a narrow main lobe, high gain, and strong anti-interference in the 0° and 90° directions, and the signal energy is highly focused on the target receiver.

[0094] In this embodiment of the invention, by optimizing the beamforming matrix row by row, a precise one-to-one configuration of M sets of weights and M array element channels is achieved. The N complex coefficients of each set of weights can match the N discrete frequency points of an ultra-wideband array, allowing the array elements to adjust the signal according to optimization requirements throughout the entire operating frequency band, thus solving the problem of ultra-wideband frequency band adaptation. For the two core operating modes of transmission and reception, differentiated processing flows for split-weighted transmission and weighted summation reception are designed, allowing the same set of optimized weights to adapt to the bidirectional communication requirements of the array, improving the method's versatility and eliminating the need to design separate optimization strategies for transmission and reception modes. Through real-time monitoring of the radiation pattern and... By slightly adjusting the weights, the radiation pattern is dynamically adjusted, ensuring that the main lobe continuously tracks the target receiver and the null is always aligned with the direction of interference. This solves the problems of traditional static beams being susceptible to environmental influences and pointing deviations, improving the beam's anti-interference capability and scene adaptability. The resulting focused beam has the characteristics of a narrow main lobe, high gain, and deep null, allowing signal energy to be highly focused in the direction of the target receiver, improving the signal-to-interference-plus-noise ratio (SNR) in the target direction. At the same time, it effectively suppresses interference signals and noise, reducing the communication bit error rate. This achieves the optimization goal of maximizing the SNR at the hardware level, providing key beam support for the efficient and stable transmission of ultra-wideband communication.

[0095] S600: Monitors communication performance parameters in real time and adjusts the beam optimization objective function based on the monitoring results.

[0096] In this embodiment of the invention, communication performance parameters are monitored in real time, and the beam optimization objective function is adjusted based on the monitoring results. Ultra-wideband communication scenarios are dynamic; the target receiver may experience slight positional shifts, and environmental interference intensity may change over time. Even if S500 has formed a focused beam pointing towards the target, after long-term operation, the adaptability of the originally optimized beamforming matrix and beam optimization objective function may decrease due to dynamic changes in the scenario, leading to communication performance degradation. S600 is the dynamic closed-loop feedback link of the entire beamforming method. By monitoring communication performance parameters in real time, it can promptly capture performance degradation trends, adjust the weight coefficients of the beam optimization objective function through feedback, and trigger a new round of matrix optimization. This ensures that the beamforming effect always adapts to the dynamic scenario requirements, avoiding communication quality degradation due to scenario changes. It achieves a closed-loop process of optimization-beamforming-monitoring-adjustment, guaranteeing the long-term stability and efficiency of ultra-wideband communication.

[0097] Step S600 in the method provided in this embodiment of the invention includes: Real-time monitoring of communication link performance parameters, which include at least signal-to-interference-plus-noise ratio and bit error rate; The communication link performance parameters are compared with a preset quality threshold, and an adjustment decision for the beam optimization objective function is generated based on the comparison result. Based on the adjustment decision, the weight coefficients in the beam optimization objective function are adjusted to generate the adjusted beam optimization objective function; Based on the adjusted beam optimization objective function, a new round of iterative optimization of the beamforming matrix is ​​triggered and executed.

[0098] First, real-time monitoring of communication link performance parameters is conducted. These parameters include at least the signal-to-interference-plus-noise ratio (SINR) and bit error rate (BER). Communication link performance parameters are core indicators used to quantitatively evaluate the transmission quality of ultra-wideband (UWB) communication links. They directly reflect the support capability of beamforming effects for communication and mainly include SINR and BER. BER refers to the ratio of the number of erroneous signal symbols transmitted in the communication link to the total number of transmitted symbols. The lower the value, the higher the signal transmission accuracy and the better the communication quality. It is typically required to be below a preset threshold. A communication link performance monitoring module is configured, establishing a bidirectional communication connection with the signal transceiver end and the target receiver of the UWB antenna array. A monitoring frequency is set, balancing real-time performance and resource consumption, typically 10ms / time to 100ms / time. SINR and BER data of the communication link are collected in real-time at a fixed frequency. The collected parameter data is preprocessed, such as through filtering and noise reduction, and the removal of transient fluctuations, to ensure the authenticity and stability of the parameter data. The preprocessed performance parameters are then uploaded to the decision adjustment module in real-time, providing reliable data support for subsequent comparison and judgment.

[0099] For example, a performance monitoring module is configured with a monitoring frequency of 50ms / time to collect the link signal-to-interference-plus-noise ratio (SIR) and bit error rate (BER) in real time; after preprocessing, stable parameter data is obtained: the initial monitoring SIR is 98.5dB, and the BER is 8×10⁻⁶. -7 Subsequently, data was continuously collected at 50ms intervals to track parameter changes in real time. For example, if the signal-to-interference-plus-noise ratio dropped to 88dB and the bit error rate rose to 1.2×10⁻⁶ at a certain moment, the data was collected. -5 Immediately upload to the decision adjustment module.

[0100] Secondly, the communication link performance parameters are compared with preset quality thresholds, and an adjustment decision for the beam optimization objective function is generated based on the comparison results. The preset quality thresholds are pre-defined communication performance standards that need to be set according to actual communication requirements, including signal-to-interference-plus-noise ratio (SINR) and bit error rate (BER) thresholds. The adjustment decision refers to generating instructions on whether to adjust the objective function and how to adjust it based on the comparison results between the performance parameters and the preset thresholds, categorized into three scenarios: no adjustment required, slight adjustment, and significant adjustment.

[0101] Specifically, based on the requirements of ultra-wideband communication scenarios, preset quality thresholds are set: signal-to-interference-plus-noise ratio (SIR) threshold and bit error rate (BER) threshold. The decision adjustment module receives the preprocessed performance parameters and compares the SIR and BER with their respective preset thresholds. Based on the comparison results, an adjustment decision is generated: if the SIR is greater than or equal to the threshold and the BER is less than or equal to the threshold, the communication performance is deemed satisfactory, and a decision requiring no adjustment is generated; if the parameters deviate slightly from the threshold, the performance is deemed slightly degraded, and a decision to slightly adjust the objective function weight coefficients is generated; if the parameters deviate significantly from the threshold, the performance is deemed severely degraded, and a decision to significantly adjust the objective function weight coefficients is generated. The generated adjustment decision is then sent to the objective function adjustment module in real time.

[0102] For example, preset quality thresholds are set as follows: signal-to-interference-plus-noise ratio ≥ 90dB, bit error rate ≤ 1×10⁻⁶. -5 The parameter data received by the decision adjustment module is divided into three scenarios: Scenario 1: Signal-to-interference-plus-noise ratio 98.5dB, bit error rate 8×10 -7 All meet the threshold requirements, and no adjustment is needed for the generated decision; Scenario 2: Signal-to-interference-plus-noise ratio 89dB, bit error rate 1.1×10 -5 Slight deviation from the threshold, generating a decision by slightly adjusting the weight coefficients of the objective function; Scenario 3: Signal-to-interference-plus-noise ratio 82dB, bit error rate 6×10 -5 If the value deviates significantly from the threshold, a decision to significantly adjust the weight coefficients of the objective function is generated; assuming the current scenario is scenario 2, a slight adjustment decision is sent to the objective function adjustment module.

[0103] Next, based on the adjustment decision, the weight coefficients in the beam optimization objective function are adjusted to generate the adjusted beam optimization objective function. The adjusted beam optimization objective function refers to the new objective function formed after correcting the weight coefficients based on the adjustment decision. It still focuses on maximizing the signal-to-interference-plus-noise ratio (SINR), but the optimization orientation is more adapted to the current dynamic scenario. The objective function adjustment module receives the adjustment decision, extracts the weight coefficients in the current beam optimization objective function, and sets weight adjustment rules according to the decision type: slight adjustment: the weight of the desired signal path increases by 5%~10%, and the weights of each interference path decrease by 10%~20%; large adjustment: the weight of the desired signal path increases by 10%~20%, and the weights of each interference path decrease by 20%~30%. The weight coefficients are corrected according to the adjustment rules to ensure that the corrected weight coefficients conform to the objective function construction logic. The corrected weight coefficients are substituted into the original beam optimization objective function to reconstruct the adjusted beam optimization objective function, which is then output to the iterative optimization trigger module.

[0104] For example, the current weight coefficients of the beam optimization objective function are: 0.9 for the desired signal path (0°, 90°), 0.2 for interference path 1 (30°, 90°), and 0.1 for interference path 2 (-30°, 90°). After receiving a slight adjustment decision, the weight coefficients are adjusted according to the rules: the weight of the desired signal path increases by 8%, 0.9 × 1.08 = 0.972; the weights of the two interference paths decrease by 15%, 0.2 × 0.85 = 0.17 and 0.1 × 0.85 = 0.085 respectively. Substituting the adjusted weight coefficients into the original objective function, the adjusted beam optimization objective function is reconstructed. This function focuses more on strengthening the desired signal and suppressing the interference signal, adapting to the scenario of slight performance degradation.

[0105] Furthermore, based on the adjusted beamforming objective function, a new round of iterative optimization of the beamforming matrix is ​​triggered and executed. This new round of iterative optimization refers to repeating the S400 iterative optimization process using the adjusted beamforming objective function as the new optimization criterion, re-optimizing the beamforming matrix to adapt it to the performance requirements of the current dynamic scene. Optimization triggering refers to the instruction that automatically initiates iterative optimization, triggered by the adjusted objective function, ensuring that the optimization process responds quickly to scene changes without manual intervention.

[0106] Specifically, the system receives the adjusted beamforming objective function and automatically generates a new round of iterative optimization trigger command; it calls the particle swarm optimization algorithm of S400, uses the currently used beamforming matrix as the new initial particles, and constructs a new particle swarm; it uses the function value of the adjusted objective function as the new particle fitness evaluation criterion and performs iterative optimization; when the preset convergence condition is met, such as the fitness change rate < 0.1% for 10 consecutive iterations, it outputs the beamforming matrix after a new round of optimization; and it sends the new matrix to the weight parsing and configuration module of S500 to replace the original matrix, thereby realizing the dynamic update of the beamforming effect.

[0107] For example, the adjusted beamforming objective function is received, triggering a new round of iterative optimization; a swarm of 20 particles is constructed using the currently used 4×5 beamforming matrix as the initial particles; iterative optimization is performed using the function value of the new objective function as the fitness; after the 18th iteration, the convergence condition is met, and the fitness change rate is <0.1% for 10 consecutive iterations; a new round of optimized 4×5 beamforming matrix is ​​output; the signal-to-interference-plus-noise ratio corresponding to this new matrix recovers to 96dB, and the bit error rate drops to 7×10⁻⁶. -7 It is then sent to the S500 to replace the original matrix and reconfigure the array element channel weights, adjust the radiation pattern, and allow the focusing beam to be re-adapted to the current scene.

[0108] In this embodiment of the invention, by real-time monitoring of the signal-to-interference-plus-noise ratio (SINR) and bit error rate (BER), dynamic changes in the communication scenario can be captured in a timely manner, ensuring that the beamforming effect always adapts to the current scenario requirements and avoiding performance degradation. By adjusting the objective function weight coefficients in stages, the desired signal is strengthened and interference is suppressed in a targeted manner. Combined with a new round of iterative optimization, the degraded communication performance can be quickly brought back to the acceptable range, ensuring that the SINR and BER remain stable within the preset threshold for a long time, thereby improving the reliability of communication transmission. A complete closed loop is formed, consisting of initial optimization, beamforming, performance monitoring, decision adjustment, objective function update, and a new round of optimization. This gives the entire beamforming method the ability to self-correct and continuously optimize, improving the long-term working stability and anti-interference capability of the ultra-wideband communication system.

[0109] Through the specific implementation methods described above, the embodiments of the present invention achieve the following technical effects: This invention provides an ultra-wideband antenna beamforming method and system. It lays a reliable optimization foundation through the construction of an initial beamforming matrix; it achieves accurate differentiation between the desired signal and interference by leveraging spatio-temporal joint preprocessing and multipath feature extraction; it rapidly approximates the globally optimal beamforming matrix using a particle swarm optimization algorithm incorporating particle similarity judgment and cooperative strategies, ultimately forming a focused beam with a narrow main lobe, high gain, and zero depth of trapping; and it effectively solves problems such as multipath interference suppression, beam pointing deviation, and insufficient scene dynamic adaptability in ultra-wideband communication by real-time monitoring of signal-to-interference ratio (SIR) and bit error rate (BER) and feedback adjustment of the beam optimization objective function, triggering a new round of iterative optimization. This ensures that the SIR of the communication link is consistently maintained at a high level and the BER is consistently below a preset threshold, achieving efficient, stable, and interference-resistant directional transmission of ultra-wideband signals, and improving the accuracy, dynamic adaptability, and communication quality stability of ultra-wideband antenna beamforming.

[0110] Example 2, as Figure 2 As shown, the present invention provides an ultra-wideband antenna beamforming system, the system comprising: The array parameter initialization module 11 is used to obtain the static parameters of the ultra-wideband antenna array of the target scene and construct an initial beamforming matrix. The ultra-wideband antenna array contains M array elements, and the static parameters include the geometric position, preset target direction and operating frequency band of the M array elements. RF signal acquisition module 12 is used to synchronously acquire M RF analog signals received by the M array elements, and perform analog-to-digital conversion on the M RF analog signals to obtain M discrete time series signals; The optimization function construction module 13 is used to perform spatiotemporal joint preprocessing on the M discrete time series signals, extract multipath signal feature parameters, and generate a beam optimization objective function in combination with preset communication quality indicators. The matrix iterative optimization module 14 is used to perform iterative optimization of the initial beamforming matrix using the particle swarm optimization algorithm with the beam optimization objective function as the optimization objective, so as to obtain the optimized beamforming matrix. The beamforming module 15 is used to dynamically adjust the radiation pattern of the ultra-wideband antenna array based on the optimized beamforming matrix to form a focused beam pointing towards the target receiver. The performance closed-loop feedback module 16 is used to monitor communication performance parameters in real time and adjust the beam optimization objective function based on the monitoring results.

[0111] In one embodiment, the array parameter initialization module 11 is further configured to: Obtain the geometric positions, preset target directions, and operating frequency bands of the M elements of the ultra-wideband antenna array; Within the operating frequency band, N discrete frequency points are selected at uniform intervals; Based on the geometric positions of the M array elements and the preset target direction, N initial beamforming weight vectors corresponding to the N discrete frequency points are calculated respectively, wherein each of the N initial beamforming weight vectors is an M-dimensional vector. The N initial beamforming weight vectors are arranged in columns to form an initial beamforming matrix with M rows and N columns.

[0112] Specifically, based on the geometric positions of the M array elements and the preset target direction, N initial beamforming weight vectors corresponding to the N discrete frequency points are calculated, including: Sort the N discrete frequency points in ascending order of frequency value, and select the first discrete frequency point as the target frequency point. Based on the geometric position coordinates of the M array elements and the preset target direction, combined with the wavelength of the target frequency point, the M relative phase delays of the M array elements are calculated. Based on the M relative phase delays, construct an M-dimensional array manifold vector; The M-dimensional array manifold vector is normalized to obtain the M-dimensional initial beamforming weight vector; By traversing the N discrete frequency points, N initial beamforming weight vectors corresponding to the N discrete frequency points are obtained; The weights include complex coefficients used to control the signal amplitude and phase of the corresponding array element at N discrete frequency points.

[0113] In one embodiment, the radio frequency signal acquisition module 12 is further configured to: Synchronization trigger signals are sent to the M receiving channels corresponding to the M array elements to obtain M original radio frequency analog signals; The M original radio frequency analog signals are subjected to anti-aliasing filtering and gain conditioning to obtain M radio frequency analog signals; The M radio frequency analog signals are synchronously sampled and quantized at a sampling frequency of not less than 2.2 times the working bandwidth to generate M discrete digital sequences; The M discrete digital sequences are combined according to the array element numbers to output M discrete time sequence signals.

[0114] In one embodiment, the optimization function construction module 13 is further configured to: The M discrete time series signals are divided into multiple spatiotemporal data blocks, where each spatiotemporal data block has a dimension of M×L, and L is the number of snapshots within a preset time window; Spatiotemporal joint parameter estimation is performed on multiple spatiotemporal data blocks to extract the multipath signal feature parameters, wherein the multipath signal feature parameters include the direction of arrival parameters, relative time delay parameters, and complex path gain feature parameters of the main multipath components; Based on the preset communication quality index, the multipath signal characteristic parameters are classified into desired signal paths and interference paths, and corresponding optimization weights are assigned to the desired signal paths and the interference paths. Based on the array response model, the multipath signal characteristic parameters, and the optimization weights, a beam optimization objective function is constructed with the beamforming matrix as the variable, wherein the objective function is used to maximize the signal-to-interference-plus-noise ratio.

[0115] The array response model is determined based on the geometric position of the ultra-wideband antenna array.

[0116] In one embodiment, the matrix iterative optimization module 14 is further configured to: The initial beamforming matrix is ​​used as an initial particle, and multiple beamforming matrices are randomly generated as an initial particle swarm, where each particle represents an M-row N-column beamforming matrix. The function value of the beam optimization objective function is defined as the fitness of the particle; Iterative optimization is performed based on the particle swarm optimization algorithm, updating the velocity and position of each particle in each iteration; During the iteration process, the similarity between particles is calculated. For particle pairs with similarity exceeding a preset similarity threshold, a cooperative strategy is implemented: particles with lower fitness are discarded, and optimization privileges are granted to the retained particles. The optimization privileges include increasing the number of derived new solutions of the retained particles in subsequent iterations. When the preset convergence condition is met, the beamforming matrix corresponding to the globally optimal particle is output as the optimized beamforming matrix.

[0117] In one embodiment, the beamforming module 15 is further configured to: The optimized beamforming matrix with M rows and N columns is analyzed into M sets of weights, and the M sets of weights are respectively assigned to M array element channels; If in signal transmission mode, one signal to be transmitted will be allocated to M array element channels according to the M sets of weights for weighted transmission; If in signal receiving mode, the signals received by the M array element channels are weighted and summed according to the M sets of weights to form a single output signal; Based on the optimized beamforming matrix, the radiation pattern of the ultra-wideband antenna array is dynamically adjusted so that the main lobe of the radiation pattern continuously points to the target receiver and forms a focused beam.

[0118] In one embodiment, the performance closed-loop feedback module 16 is further configured to: Real-time monitoring of communication link performance parameters, which include at least signal-to-interference-plus-noise ratio and bit error rate; The communication link performance parameters are compared with a preset quality threshold, and an adjustment decision for the beam optimization objective function is generated based on the comparison result. Based on the adjustment decision, the weight coefficients in the beam optimization objective function are adjusted to generate the adjusted beam optimization objective function; Based on the adjusted beam optimization objective function, a new round of iterative optimization of the beamforming matrix is ​​triggered and executed.

[0119] It should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for beamforming an ultra-wideband antenna, characterized in that, The method includes: Obtain the static parameters of the ultra-wideband antenna array of the target scene and construct an initial beamforming matrix. The ultra-wideband antenna array contains M array elements, and the static parameters include the geometric positions, preset target directions and operating frequency bands of the M array elements. Simultaneously acquire M radio frequency analog signals received by the M array elements, perform analog-to-digital conversion on the M radio frequency analog signals, and obtain M discrete time series signals; The M discrete-time series signals are subjected to joint spatiotemporal preprocessing to extract multipath signal feature parameters. Combined with preset communication quality indicators, a beam optimization objective function is generated, including: The M discrete time series signals are divided into multiple spatiotemporal data blocks, where each spatiotemporal data block has a dimension of M×L, and L is the number of snapshots within a preset time window; Spatiotemporal joint parameter estimation is performed on multiple spatiotemporal data blocks to extract the multipath signal feature parameters, wherein the multipath signal feature parameters include the direction of arrival parameters, relative time delay parameters, and complex path gain feature parameters of the main multipath components; Based on the preset communication quality index, the multipath signal characteristic parameters are classified into desired signal paths and interference paths, and corresponding optimization weights are assigned to the desired signal paths and the interference paths. Based on the array response model, the multipath signal characteristic parameters, and the optimization weights, a beam optimization objective function is constructed with the beamforming matrix as the variable, wherein the objective function is used to maximize the signal-to-interference-plus-noise ratio; The array response model is determined based on the geometric position of the ultra-wideband antenna array; The particle swarm optimization algorithm is used to iteratively optimize the initial beamforming matrix with the beam optimization objective function as the optimization objective to obtain the optimized beamforming matrix. Based on the optimized beamforming matrix, the radiation pattern of the ultra-wideband antenna array is dynamically adjusted to form a focused beam pointing towards the target receiver. The communication performance parameters are monitored in real time, and the beam optimization objective function is adjusted based on the monitoring results.

2. The ultra-wideband antenna beamforming method as described in claim 1, characterized in that, Obtain the static parameters of the ultra-wideband antenna array in the target scene and construct the initial beamforming matrix, including: Obtain the geometric positions, preset target directions, and operating frequency bands of the M elements of the ultra-wideband antenna array; Within the operating frequency band, N discrete frequency points are selected at uniform intervals; Based on the geometric positions of the M array elements and the preset target direction, N initial beamforming weight vectors corresponding to the N discrete frequency points are calculated respectively, wherein each of the N initial beamforming weight vectors is an M-dimensional vector. The N initial beamforming weight vectors are arranged in columns to form an initial beamforming matrix with M rows and N columns.

3. The ultra-wideband antenna beamforming method as described in claim 2, characterized in that, Based on the geometric positions of the M array elements and the preset target direction, calculate the N initial beamforming weight vectors corresponding to the N discrete frequency points, including: Sort the N discrete frequency points in ascending order of frequency value, and select the first discrete frequency point as the target frequency point. Based on the geometric position coordinates of the M array elements and the preset target direction, combined with the wavelength of the target frequency point, the M relative phase delays of the M array elements are calculated. Based on the M relative phase delays, construct an M-dimensional array manifold vector; The M-dimensional array manifold vector is normalized to obtain the M-dimensional initial beamforming weight vector; By traversing the N discrete frequency points, N initial beamforming weight vectors corresponding to the N discrete frequency points are obtained; The initial beamforming weight vector includes complex coefficients used to control the signal amplitude and phase of the corresponding array element at N discrete frequency points.

4. The ultra-wideband antenna beamforming method as described in claim 1, characterized in that, Simultaneously acquire M channels of radio frequency analog signals received by the M array elements, perform analog-to-digital conversion on the M channels of radio frequency analog signals to obtain M channels of discrete time series signals, including: Synchronization trigger signals are sent to the M receiving channels corresponding to the M array elements to obtain M original radio frequency analog signals; The M original radio frequency analog signals are subjected to anti-aliasing filtering and gain conditioning to obtain M radio frequency analog signals; The M radio frequency analog signals are synchronously sampled and quantized at a sampling frequency of not less than 2.2 times the working bandwidth to generate M discrete digital sequences; The M discrete digital sequences are combined according to the array element numbers to output M discrete time sequence signals.

5. The ultra-wideband antenna beamforming method as described in claim 1, characterized in that, Using a particle swarm optimization algorithm, with the beamforming objective function as the optimization target, the initial beamforming matrix is ​​iteratively optimized to obtain the optimized beamforming matrix, including: The initial beamforming matrix is ​​used as an initial particle, and multiple beamforming matrices are randomly generated as an initial particle swarm, where each particle represents an M-row N-column beamforming matrix. The function value of the beam optimization objective function is defined as the fitness of the particle; Iterative optimization is performed based on the particle swarm optimization algorithm, updating the velocity and position of each particle in each iteration; During the iteration process, the similarity between particles is calculated. For particle pairs with similarity exceeding a preset similarity threshold, a cooperative strategy is implemented: particles with lower fitness are discarded, and optimization privileges are granted to the retained particles. The optimization privileges include increasing the number of derived new solutions of the retained particles in subsequent iterations. When the preset convergence condition is met, the beamforming matrix corresponding to the globally optimal particle is output as the optimized beamforming matrix.

6. The ultra-wideband antenna beamforming method as described in claim 1, characterized in that, Based on the optimized beamforming matrix, the radiation pattern of the ultra-wideband antenna array is dynamically adjusted to form a focused beam pointing towards the target receiver, including: The optimized beamforming matrix is ​​analyzed into M sets of weights, and the M sets of weights are respectively assigned to M array element channels; If in signal transmission mode, one signal to be transmitted will be allocated to M array element channels according to the M sets of weights for weighted transmission; If in signal receiving mode, the signals received by the M array element channels are weighted and summed according to the M sets of weights to form a single output signal; Based on the optimized beamforming matrix, the radiation pattern of the ultra-wideband antenna array is dynamically adjusted so that the main lobe of the radiation pattern continuously points to the target receiver and forms a focused beam.

7. The ultra-wideband antenna beamforming method as described in claim 1, characterized in that, Real-time monitoring of communication performance parameters, and adjustment of the beam optimization objective function based on the monitoring results, including: Real-time monitoring of communication link performance parameters, which include at least signal-to-interference-plus-noise ratio and bit error rate; The communication link performance parameters are compared with a preset quality threshold, and an adjustment decision for the beam optimization objective function is generated based on the comparison result. Based on the adjustment decision, the weight coefficients in the beam optimization objective function are adjusted to generate the adjusted beam optimization objective function; Based on the adjusted beam optimization objective function, a new round of iterative optimization of the beamforming matrix is ​​triggered and executed.

8. An ultra-wideband antenna beamforming system, characterized in that, The system for implementing the ultra-wideband antenna beamforming method according to any one of claims 1-7, the system comprising: The array parameter initialization module is used to obtain the static parameters of the ultra-wideband antenna array of the target scene and construct the initial beamforming matrix. The ultra-wideband antenna array contains M array elements, and the static parameters include the geometric position, preset target direction and operating frequency band of the M array elements. The radio frequency signal acquisition module is used to synchronously acquire the M radio frequency analog signals received by the M array elements, and perform analog-to-digital conversion on the M radio frequency analog signals to obtain M discrete time series signals. The optimization function construction module is used to perform spatiotemporal joint preprocessing on the M discrete time series signals, extract multipath signal feature parameters, and generate a beam optimization objective function in combination with preset communication quality indicators. The matrix iterative optimization module is used to perform iterative optimization of the initial beamforming matrix using the particle swarm optimization algorithm, with the beamforming objective function as the optimization target, to obtain the optimized beamforming matrix. The beamforming module is used to dynamically adjust the radiation pattern of the ultra-wideband antenna array based on the optimized beamforming matrix, so as to form a focused beam pointing towards the target receiver. The performance closed-loop feedback module is used to monitor communication performance parameters in real time and adjust the beam optimization objective function based on the monitoring results.

Images