Portable satellite communication station signal optimization method and system

Abstract

The invention discloses a portable satellite communication station signal optimization method and system, relates to the field of wireless communication, and is used for synchronously realizing dynamic adaptive management of multiple scattering paths. The method comprises the following steps: generating an earth surface image through multi-modal detection data, and carrying out layered calibration on scattering intensity; identifying a high-reflection sub-region according to the surface image and the layered calibration information thereof, and performing unified management on high-reflection and weak echo paths; based on the gain, interference and angle information of the multiple scattering paths, performing sub-array division on the antenna array according to priorities; phase stepping correction is carried out on the antenna units of the sub-arrays, and emission excitation is combined in adjacent path coverage areas; through a dynamic power scheduling and beam forming algorithm, the direction gain of each sub-array is optimized and sidelobe leakage is controlled. According to the invention, synchronous management of high reflection and weak echo paths can be realized, and the dynamic adaptability to scattering scenes is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of wireless communication, and in particular to a portable satellite communication station signal optimization method and system. BACKGROUND

[0002] Multiple-input multiple-output technology is widely used in the field of wireless communication to improve transmission rate and anti-interference performance. Existing technical solutions generally use array antennas in combination with channel estimation to perform parallel processing of signals from multiple transmission and reception channels. By implementing channel equalization and beamforming at the physical layer, higher spatial multiplexing efficiency can be achieved under fixed frequency spectrum resources. In addition, a cooperative scheduling strategy can also be combined to enable flexible power distribution of multiple antenna units for different scattering environments and dynamic adjustment of transmission gain when fading or blockage is detected. Such strategies are usually based on known channel characteristics and weight multiple paths to extract more effective signal components in complex propagation paths.

[0003] To meet the demand for multi-mode detection, known solutions attempt to explore the configuration of different types of sensors in the same system, such as radar, optical, radio frequency, etc. modules to comprehensively utilize their respective target detection range and resolution characteristics. Such fusion solutions often use interface protocols or interface modules to aggregate detection information from multiple sources and align them at a certain time reference. Subsequently, at the high-level data level, filtering and interpolation methods are used for fusion to obtain a relatively fine scene distribution. In practical applications, these multi-sensor units usually work independently, and the algorithms mainly focus on data alignment and multi-source information fusion. For temporary interference and specific scattering effects, most still rely on traditional channel compensation mechanisms or offline modeling methods.

[0004] In the processing of multiple scattering scenarios, some systems further introduce geometric positioning or beam mapping functional modules to attempt to extract primary and secondary reflection signals in multiple subspaces, thereby preliminarily labeling different reflection paths. The known approach is to measure the time delay and angle information of the received signal to locate high-intensity scattering sources, and then correct weak reflections or secondary reflections by attaching them to the records of stronger paths. With changes in the multiple scattering environment, existing solutions usually perform one or more parameter revisions at the subarray level, but in small-scale array settings, only the main reflection paths are often modeled, and weak echoes or unstable paths are often given secondary status.

[0005] In terms of coding and resource allocation, known techniques are often based on real-time evaluation of path gain, assigning paths with higher channel gain to higher order modulation, or concentrating dominant power within certain subcarrier intervals to reduce the waste of useless bandwidth. When encountering sudden fading or environmental changes, the error correction code rate is generally adjusted within a given bandwidth convergence range to balance the bit error rate and throughput efficiency as much as possible. Some advanced schemes also use carrier aggregation or multiplexing to improve overall throughput, and implement aggregated decoding on the receiving side. In this way, regardless of the number and type of sensors, the reliability of the overall link can be maintained under multiple scattering or complex terrain conditions. Although these techniques can reasonably schedule the slowly changing reflective environment in a steady state scenario, there are still some differences in design and implementation in terms of capturing and updating more scattering paths in a timely manner, fine-grained differentiation of different scattering levels, and dynamic correction of multi-dimensional parameters.

[0006] In a multiple-input multiple-output system, although the use of array antennas and channel estimation has the ability to improve throughput and reduce bit error rate, existing technologies are mostly based on equalization and weighting of the main reflection path, and the processing of weak reflections or multiple scattering is often general, making it difficult to capture weak signal characteristics in a timely manner. This design may not adapt well when the terrain environment changes significantly, resulting in a communication link that cannot accurately respond to sudden obstructions or interference. On the other hand, although multi-mode detection methods have emerged, common systems focus on the static effects of data fusion and do not provide a complete dynamic observation and correction process for transient scattering environments. For high reflection areas and secondary reflection areas, a unified approach is often used for processing. Accordingly, in terms of parallel processing of multi-sensor data and layered identification of multiple scattering paths, it is difficult to take into account the differences between different types of echoes. When it comes to spatial filtering or beamforming, it is often limited to local scheduling output. When faced with complex scattering source distribution and multi-dimensional interference factors, it is difficult to quickly adjust the subarray configuration or power distribution strategy of the antenna array.

[0007] In terms of resource scheduling and coding allocation, known existing technologies are mainly based on periodic sampling of instantaneous channel gain, adjusting modulation and error correction depth based on a wide range of statistical models, but the rapid fading or phase drift caused by multiple scattering is often not fine enough. Known approaches generally allow high-order modulation for high-gain paths and redundant coding for fading paths to protect them, but due to the lack of comprehensive judgment of fine-grained factors such as scattering path intensity level, surface structure changes, and even sudden interference points, it is often difficult to make more flexible optimizations in parallel multiplexing and bandwidth slicing. SUMMARY

[0008] The application aims at providing a portable satellite communication station signal optimization method and system to realize dynamic adaptive management of multiple scattering paths.

[0009] The technical scheme adopted by the application is as follows: The application provides a portable satellite communication station signal optimization method, which comprises the following steps: S1, checking, time sequence slicing and spatial coordinate registration of radar, optical and infrared three-mode detection data of a detection area; fusing the detection data of various modes to generate a ground surface image, performing attenuation correction on the ground surface image based on meteorological observation data; and performing layered calibration of scattering intensity on the ground surface image; S2, identifying a high-reflection sub-area according to the ground surface image and layered calibration information thereof, marking a geometric scatterer in the high-reflection sub-area; detecting a weak echo path in the high-reflection sub-area, adding the weak echo path to a multiple scattering path set; and performing discrete distribution mapping and secondary time delay review on the azimuth angle and the elevation angle of each scattering path to generate a multiple scattering path record with gain distribution and interference identification; S3, dividing an antenna array into sub-arrays according to the priority of gain, interference and angle information of the multiple scattering paths; performing phase step correction on the antenna units of each sub-array and combining transmission excitation in adjacent path coverage areas; and optimizing the directional gain of each sub-array and controlling the sidelobe leakage through dynamic power scheduling and beam forming algorithms.

[0010] The application further provides a portable satellite communication station signal optimization system, which comprises a storage medium and a processor, the storage medium stores a computer program, and the processor executes the computer program to perform the portable satellite communication station signal optimization method.

[0011] As described above, due to the adoption of the above technical scheme, the application has the following advantages: (1) The application first introduces radar, optical and infrared detection modules into the same portable satellite communication station through the fusion means of multi-mode detection, and establishes a unified interface and checking mechanism. On the one hand, the output consistency of various sensors between different time sequences and coordinate systems is ensured, and on the other hand, meteorological attenuation factors can be dynamically corrected in the detection results in the early stage. Compared with the method of relying on only a single waveband or lacking attenuation compensation algorithm, the present application can form a more detailed layered calibration basis in the ground surface sub-area coverage analysis, thereby providing more accurate input data for the subsequent identification and management of multiple scattering paths.

[0012] (2) In the aspect of multi-scattering path calibration and phased management, geometric scatterers in high reflection blocks are marked, and low-intensity weak echoes are identified synchronously, and both are included in the path set for analysis. The boundary of different scattering sources is divided by using the discrete mapping mode of azimuth angle and elevation angle, and the determination error caused by the overlapping of path arrival times is reduced by secondary time delay review. In this process, not only the gain and interference of each path are mapped, but also the interference sources are tracked and positioned based on the mutual interference relationship between adjacent paths, which is different from the method of only focusing on the main reflection path or a single high-intensity signal. Through this "high reflection + weak reflection" strategy, timely monitoring and differentiation of multi-stage scattering channels in complex environments are realized.

[0013] (3) In the subsequent antenna array and beam forming design, the application prioritizes the paths according to their gain, interference, and angle matching degree, and then logically divides the antenna array into several parallel sub-arrays, so that each sub-array adaptively focuses on high-priority paths and merges and excites adjacent path regions. Compared with the general static or single-point correction strategy, this application continuously collects monitoring data at the receiving end link, feeds back the phase drift and energy gain anomalies to the sub-array in a timely manner, and finally realizes adaptive control of the multi-scattering environment under the dynamic updating of beam forming and power distribution mechanism, so as to maintain the identification and support of key paths in the presence of terrain obstruction or sudden interference. BRIEF DESCRIPTION OF DRAWINGS

[0014] The application will be described by way of example and with reference to the accompanying drawings, in which: Figure 1 is a flowchart of a signal optimization method for a portable satellite communication station.

[0015] Figure 2 is a flowchart of a signal optimization method for a portable satellite communication station in one embodiment. DETAILED DESCRIPTION

[0016] All features disclosed in this specification, or all steps of any method or process specified in this specification, can be combined in any combination, except combinations where at least some features and / or steps are mutually exclusive.

[0017] Any feature disclosed in this specification, unless stated otherwise, can be replaced by alternative features or equivalents having the same or a similar effect. That is, unless stated otherwise, each feature is one example only of a number of alternative or comparable features.

[0018] In the embodiments of the present application, a signal optimization method for a portable satellite communication station is provided, referring to the accompanying drawings Figure 1 The method comprises the following processes: S1, checking, time series slicing and spatial coordinate registration of the radar, optical and infrared three-mode detection data of the detection area; fusing the detection data of each mode to generate a ground surface image, and performing attenuation correction on the ground surface image based on meteorological observation data; and performing layered calibration of the scattering intensity of the ground surface image.

[0019] As an optional implementation, as shown in Figure 2 , step S1 includes: S11, checking the output consistency of each detection unit of the detection data of each mode.

[0020] In the portable satellite communication station, multi-mode detection units meeting the requirements of low power consumption and high detection accuracy are deployed, including radar detection modules, optical detection modules and infrared detection modules, to detect radar, optical and infrared three-mode detection data of the target area (usually 5km field of view range near the communication station). Based on a unified communication interface and power-on test mechanism, the output consistency of each detection unit is preliminarily checked. The checking method is: Formula (1): ; In the formula, i represents the number of detection units, j represents the number of sub-regions in the target area detected, represents the spatial range of the i-th sub-region, j represents the radiation count of the i-th detection unit at the pixel point coordinate , i represents the effective coverage function of the i-th detection unit at the coordinate , DN represents the integral result of the coverage of the i-th detection unit to the j-th sub-region, and the dimension is . i Based on the threshold, the coverage of the sub-regions is tested and screened out, and the arrangement position of the detection units is optimized, so that the coverage of at least the detection area meets the threshold requirement. The optimization of the distribution of the detection units is targeted at , and is constrained by the total power consumption and mechanical travel. The optimization method is: under the condition of screening out the sub-regions with insufficient coverage, the positions of the detection units are iteratively moved by using the simulated annealing method, so that j is minimized, and , is the threshold for testing, which is called the empirical radiation channel threshold. The finally arranged detection units meet the mechanical travel constraint and achieve the optimal global uniform coverage.

[0021] ​​​​​​

[0022] In some possible embodiments, the threshold value is adaptively updated by the 95% quantile of the in-orbit radiation baseline in the last several (e.g., 3) detection periods, taking into account the low-power consumption and high-dynamic-range detection requirements. The so-called in-orbit radiation baseline refers to the reference curve formed by the radiation counts obtained during the recent several clear sky passages of the satellite, after normalization by the on-board internal standard source and cross-orbit cross-calibration . It is updated by removing cloud and rain anomalies (the mean and standard deviation of the radiation counts in each detection period) and then compared with the ground radiation target to write into the FPGA LUT (Look Up Table), which is updated every new period. The mean and standard deviation of the radiation counts in each detection period are removed, and then compared with the ground radiation target to write into the FPGA LUT (Look Up Table), which is updated every new period.

[0023] S12, based on the expected coverage range, the detection area is divided into sub-regions, and each sub-region is time-series sliced, spatially coordinated and scanned segment indexed.

[0024] The method of sub-region division is derived from the six-frequency chirp gridding of global DEM (Digital Elevation Model) by the on-board mission planning system, with a grid resolution of The grid is used as a static template in the mission planning interface, without the need for multiple dynamic recalculation, only the patrol scanning sequence and grid ID need to be corrected.

[0025] The so-called time-series slice is based on the visible window of the satellite to the same ground area , The start and end times are respectively. The scanning band obtained by detecting multi-modal data is sorted according to the time sequence and divided into equal time intervals (3 seconds by default, which can be adjusted in real time according to the angular velocity jitter of the on-board IMU (Inertial Measurement Unit)). Each scanning segment is divided into segments, each of which maintains quasi-steady state in attitude and speed, thereby ensuring the consistency of radiation response and geometric resolution within each segment.

[0026] The spatial coordinate registration of the sub-region refers to the unification of the spatial coordinates of each sub-region in the same coordinate system.

[0027] According to the division results of each sub-region and the sub-regions covered by each scanning segment, each scanning segment is assigned an independent index. The scanning segment itself does not increase the on-site remote sensing data, but is a time index that supports subsequent sub-array beam optimization and power scheduling.

[0028] ​S13. Integrate multimodal detection data into a unified coordinate system, and weight and fuse the multimodal detection data of each scan segment into a surface image.

[0029] The radar, optical, and infrared detection data from each scan segment are fused into the same coordinate system. The distribution intensity of the fused multimodal detection data is calculated separately for each scan segment, resulting in a more refined depiction of undulations and obstacle distribution contours in the dynamic map. The multimodal detection data fusion method for each scan segment is as follows: Equation (2): ; In the formula, This indicates that the k-th scan segment is at pixel point The fused radiation intensity represents the topographic perception, which is the weighted overlay result of radiation representations of the same surface under multiple temporal phases and different observation conditions. The unit remains 1. ; Indicates the first i Each detection unit at the pixel The original radiation intensity at that location, after irradiance normalization, is dimensionless. K The total number of detection units. Indicates the first i The weighting coefficients of each detection unit are derived from a comprehensive score of attitude control stability, signal-to-noise ratio (SNR), and solar zenith angle, and satisfy... By weighted and fused multimodal detection data from multiple detection units, single-phase noise can be effectively suppressed and the inherent reflection differences of the terrain can be highlighted.

[0030] S14. Attenuation correction is performed on the surface images of each scan segment based on meteorological observation data.

[0031] Meteorological observation data, including rainfall, wind direction, and atmospheric attenuation, has a significant impact on the detection of surface radiation intensity. This involves mapping and correlating meteorological observation data with surface images of the scanned segments, and then adding meteorological labels to the scanned segments.

[0032] Based on the degree of influence of meteorological interference during surface detection, the attenuation of surface images is corrected using meteorological observation data through the following methods: Equation (3): ; In the formula, This represents the total scattered energy of the k-th scan segment at time t after meteorological attenuation correction, still expressed in units of 1. ; This represents the meteorological attenuation factor, which is linearly interpolated to the radiation intensity of the scanned segment after sampling over time t, and mapped to the total scattered energy. It changes dynamically based on atmospheric attenuation characteristics such as rainfall; This represents the spatial range of the k-th scan segment.

[0033] right Real-time motion averaging analysis is performed to accurately assess the impact of weather on the detection path, providing attenuation correction for subsequent block layering and path identification.

[0034] The so-called motion average analysis refers to the analysis of... Perform a moving average. For example, in one specific embodiment, a moving window covering N (N is a positive integer greater than 2, such as N=5) scan segments is used. The moving average is obtained by performing mean processing. This is the motion average analysis result for the current scan segment, which can reduce the amplitude of random fluctuations while ensuring response performance. The sliding window can start at the current scan segment or be located at the middle of the current scan segment.

[0035] S15. Perform layered aggregation of scattering intensity levels on the surface images of each scan segment.

[0036] Layered aggregation of surface images is performed based on the scattering attenuation correction factor for each scan segment. In some optional implementations, the scattering attenuation correction factor is calculated using the following formula: Equation (4): ; In the formula, Indicates the first k The scattering attenuation correction factor for each scan segment For the first k The most recent clear-sky baseline for each scan segment. This clear-sky baseline is calculated by the onboard equipment in the FPGA and stored in the LUT corresponding to each scan segment. A coefficient threshold is set. ,determination This scenario involves significant differences and triggers a hierarchical aggregation process.

[0037] The so-called hierarchical aggregation process includes: using sub-regions as clustering objects, and performing a correction based on their scattering attenuation coefficients within the same scan segment. Perform hierarchical clustering. Then, based on a set threshold, calculate the Euclidean distance across the scan segment scale. Adjacent scan segments that reach the threshold are merged to avoid excessive fragmentation caused by noise islands. All data is acquired along with the on-orbit commissioning records.

[0038] S2, identifying high-reflection sub-regions from the ground surface image and its layered calibration information, marking the high-reflection sub-regions as geometric scatterers, detecting weak echo paths in the high-reflection sub-regions, and adding the weak echo paths to a multi-scattering path set; and mapping discrete distribution of each scattering path in azimuth and elevation and secondary time delay calibration to generate a multi-scattering path record with gain distribution and interference identification.

[0039] As an optional implementation, step S2 includes: S21, calculating reflection intensity of each sub-region respectively, and screening high-reflection sub-regions based on a first threshold.

[0040] In this step, the reflection intensity of each sub-region is counted and accumulated respectively, a high-reflection index (i.e. threshold) is constructed, and potential regions of multi-scattering are selected based on the high-reflection index.

[0041] Specifically, the sub-region set is defined as , wherein j represents the jth sub-region, and m represents the total number of sub-regions. For each sub-region, the reflection intensity integral is calculated as : Equation (5): ; wherein represents the ground surface reflection coefficient at pixel point .

[0042] The first threshold is set as , and the size of the sub-region reflection intensity integral and the first threshold is compared. When > , the sub-region is determined as a high-reflection sub-region, otherwise, the sub-region is classified into a single-scattering normal ray module (instead of being deleted). In this way, all terrain geometric information is retained, and only statistically significant sub-regions in optics or microwaves are ensured to participate in subsequent multi-path reasoning.

[0043] In some preferred embodiments, the first threshold is set adaptively: Equation (6): ; wherein are the mean and standard deviation of the reflection intensity integral set of the same target region, respectively. The coefficient is derived from in-orbit calibration data, and is usually about 1.8.

[0044] S22, marking the high-reflection sub-regions as geometric scatterers.

[0045] The edge position and the center position of the high-reflection sub-region are extracted, for example, the edge of the high-reflection sub-region is extracted by using the edge detection algorithm of the Canny operator, and then the center is detected according to the principle of the connected domain, that is, the center position is obtained. The center position is the position of the geometric scatterer.

[0046] A unique index is established for each labeled geometric scatterer, and is written into the scatterer index table to avoid the problem of attribution confusion of multiple adjacent geometric scatterers.

[0047] For example, the position coordinates of each geometric scatterer are mapped to the WGS-84 ellipsoid system Triplet, where respectively represent the latitude, longitude and ellipsoid height, , , h is in m, taken from the GNSS height corrected from the EGM96 gravity field. Insert the R*-tree spatial index, and the index number is the geometric scatterer ID. In subsequent ray intersection operations, only Table lookup can determine the unique attribution of the scattering source.

[0048] S23, detecting a weak echo path in the high-reflection sub-region of the labeled geometric scatterer, and adding the detected weak echo path to the multiple scattering path set.

[0049] After the positioning and labeling of the geometric scatterer are completed, the low-intensity echoes that may exist in the same high-reflection sub-region are detected, and the weak echo path is supplemented into the multiple scattering path set to be observed, and the completeness of the path information is expanded.

[0050] The weak echo path detection method includes: The sliding window energy is calculated according to the following formula: Formula (7): ; In the formula, represents the sliding window energy of the echo sequence of the same geometric scatterer from the echo pulse p , with a length (i.e. the number of consecutive connected echo pulses) of W ; represents the energy function of the echo pulse, that is, the amplitude of the echo pulse.

[0051] The energy threshold and , are set, if , it is determined that the echo path is transient noise, which is removed; if , it is determined that the echo path is a weak echo path, and if , it is determined that the echo path is a main reflection path.

[0052] Using sliding window energy to determine rather than single echo pulse amplitude to determine can effectively exclude the influence of isolated noise peaks on the determination result, and improve the accuracy of classification.

[0053] S24, mapping and separating the discrete distribution of each geometric scatterer in azimuth and elevation, separating the main reflection path and the weak echo path.

[0054] For each scatterer, record its angle coordinates to realize the boundary division of adjacent scatter sources and prevent the main reflection path from overlapping with the adjacent weak echo path in the angle dimension.

[0055] S25, secondary time delay correction of the offset of each separated scattering path to reduce the false determination caused by the overlapping of different paths.

[0056] Define the scattering path set as , The p th path in the set is represented as n , and the total number of paths in the set is represented as . For any two paths in the set, calculate the optimal alignment offset in the time domain: ; In the formula, the normalized signal envelope of path at time t is represented as , the normalized signal envelope of path under the additional offset m is represented as , and L represents the observation time considered by the secondary time delay correction. Through the secondary time delay correction, the path arrival time error is controlled at sub-chip level, and then the geometric constraint in the time domain is given.

[0057] S26, gain distribution and interference identification processing of each scattering path to generate a gain-interference mapping table.

[0058] Project all scattering paths after secondary time delay correction onto the angle-time delay three-dimensional kernel function , where represent the angle and time delay respectively. Calculate the comprehensive gain of path by the following formula: ; ; In the formula, the projection domain of path in the plane is represented as .

[0059] Based on the threshold and the threshold , if Or the cross-correlation value with the main reflection path is lower than Then the path is marked as hidden interference and suppressed before beamforming; otherwise the path is reserved for subsequent MIMO beam design. Wherein, The 10% quantile of the gain set of the last several (e.g. 3) in-orbit scattering paths is taken ( The gain of a single scattering path) and linearly compensated with the noise floor to filter out paths below the power threshold. The 95% confidence lower bound under the clear sky baseline of the cross-correlation sequence of the main reflection path and the adjacent paths is determined and updated in real time with the IMU drift coefficient. Through gain analysis and interference coordination analysis, the direction of subsequent MIMO beam design and power allocation can be corrected.

[0060] S3, based on the gain, interference and angle information of multiple scattering paths, the antenna array is divided into subarrays according to priority; phase stepping correction is performed on the antenna elements of each subarray, and adjacent path coverage areas are merged for transmission excitation; through dynamic power scheduling and beamforming algorithm, the directional gain of each subarray is optimized and the sidelobe leakage is controlled.

[0061] As an optional implementation, step S3 includes: S31, based on the gain, interference and angle matching information of multiple scattering paths, the priority score of each scattering path is calculated respectively, and a predetermined number / percentage of scattering paths with the highest priority score are extracted.

[0062] In the set of scattering paths P , for any path , based on the data of gain, interference and angle information, the priority of each path is calculated according to the importance level, and the calculation method is: Equation (10): ; In the equation, represents the gain distribution of the path at the angle and the delay ; represents the interference intensity of the path at the angle and the delay , which is obtained by taking the cross-correlation of the complex envelope of the path with the remaining echoes in the window and taking the amplitude at zero delay after averaging and normalizing to the value of to the value of ; represents the path​ At angle And delay The angle matching function, which is related to the coupling degree of the system beam layout, points to the path azimuth-elevation And the array design main lobe Calculate the cosine similarity ( The included angle between the two directions), and then linearly calibrate and normalize to obtain the angle matching function; Indicates the path Integral domain in The plane. For each path, a scalar weight , Normalized by link loss, Reflects the angle domain matching degree, Measures the interference margin. The three coefficients are in the normalized interval of , which is convenient for direct comparison. Specifically, The historical data of the nearest Link capacity, bit error rate and sidelobe leakage are obtained by multi-objective Bayesian optimization. The optimization goal is to maximize The correlation of the three indicators (gain distribution, interference strength, and angle matching). The initial value , every Frame (variable) iterations, if ( Indicates the global evaluation value of the priority score of the entire scattering path set under the current scalar weight , Indicates the change of Evaluated by adjacent two rounds, that is , Indicates the J of the rth round), then adaptively adjust , otherwise keep the last round To prevent excessive swing. Always keep And limit the three weights to the interval .

[0063] The calculation method of the above In each round of iteration is: Assume that in the rth round of iteration, the scattering path set of the tth frame is , which contains K available scattering paths. Calculate the aggregate weight of the kth scattering path: , Indicates the energy of the kth scattering path in the tth frame. For each scattering path, calculate and normalize three vectors: the gain vector , the interference strength vector , and the angle matching degree vector , for The value corresponding to the k-th scattering path in the t-th frame; according to the set The weighted calculation yields: Based on the scalar weighting of the current round Calculate the global evaluation value for this frame: ,in .

[0064] After calculating the priority score of the scattering path Afterwards, based on For the set of scattering paths P The scattering paths are sorted in descending order to provide location anchors for subarray power redistribution. The highest-scoring scattering paths of a predetermined number / proportion are extracted and used as the key configuration targets for subsequent subarray power redistribution.

[0065] S32. Divide the antenna array into subarrays based on the principle that each subarray focuses on the extracted scattering path.

[0066] The antenna array is logically divided into multiple parallel subarrays to avoid wasting transmit power on secondary paths. A subarray refers to an antenna cluster that is continuously divided by address on the array surface and independently fed. z This indicates the number of antennas in a single subarray, for example, Indicates the first k The number of antennas in a subarray. The set of subarray antennas is defined as... , Indicates the first subarray z One antenna.

[0067] Constructing a subarray partitioning matrix based on a multi-dimensional partitioning strategy S This allows each subarray to focus on several high-priority scattering paths, and imposes determinant constraints on the mutual interference between different subarrays. k The internal correlation is measured.

[0068] Subarrays are constructed using the following formula. k Cross-correlation matrix of internal antenna elements : Equation (11): ; In the formula, Subarray k Inside r Branch 1 (antenna) and s The correlation coefficient between branches 1 and 2 , u,v Representing subarrays k Number of rows and columns of the antenna; cross-correlation matrix Then the correlation matrix of the subarray k .

[0069] For Threshold screening based on the set threshold and combined with iterative update strategy for subarray division update, the internal structure of each subarray tends to be independent, providing a good foundation for subsequent phase correction and power allocation.

[0070] The above threshold screening and iterative updating process, wherein the threshold screening first checks In , set the upper limit ; if , the corresponding antenna pair is forced to be divided into different subarrays. Subsequently, with as the target, the address table is iteratively rearranged using the simulated annealing method until the overall array matrix gain converges. After completion, the correlation coefficient in each subarray is generally low and has zero overlap with other subarrays, i.e., it tends to be independent, so that subsequent phase correction and power allocation are not coupled, and the beam shaping efficiency is highest.

[0071] S33, matching the expected phase of the extracted scattering path, corrects the phase step of each subarray antenna unit.

[0072] The phase of the antenna unit of the subarray k is marked as , the expected phase vector of the target path is marked as , and the phase difference is minimized based on the following formula: Formula (12): ; In the formula, represents the phase difference obtained by optimizing the subarray k , i.e., the minimum phase difference; represents the translation amount of the overall phase drift between the subarray k and the high-priority scattering path (referred to as this path) , the initial value is the carrier phase at the starting time of the path, and the iterative process only updates this scalar, and the relative phase of the antenna unit within the subarray is unchanged, so the calculation cost of the optimization process can be greatly reduced; represents the target phase of the th antenna unit corresponding to the path u , represents the original phase of the k th antenna unit in the subarray u . Through the above process, the phase offset can be optimized, and the correction of the phase step of the subarray antenna unit is completed.

[0073] S34, combine the transmit excitations of the subarrays that exist in the adjacent path coverage area.

[0074] If there is an adjacent path coverage area between the subarrays, the transmit excitations of the relevant subarrays are combined to jointly improve the detectability of the weak echo path.

[0075] The shared excitation vector of the subarrays is represented as , and amplitude accumulation is performed on the vector to obtain the comprehensive transmit power described by the following formula : Formula (13): ; In the formula, , represents the normalized excitation amplitude of the corresponding antenna element; , and represents the number of antennas of the corresponding subarray , as described above.

[0076] Only transmit excitation combination will superimpose the complex current envelope of subarray k in the same scattering direction and perform interpolation alignment after frequency domain convolution. The comprehensive transmit power obtained by the combination directly enters the power constraint optimization step.

[0077] S35, based on the phase stepping correction vector and the transmit excitation combination result, calculate the array response function of each subarray in the three-dimensional angle space using the beamforming algorithm, and search for the scattering path corresponding to the maximum gain in the angle domain to focus the main lobe in the direction of the scattering path.

[0078] In step S35, the phase stepping correction vector in step S33 and the transmit excitation combination result in step S34 are input into the beamforming algorithm to calculate the array response function of each subarray in the three-dimensional angle space , and the complex weight value of the subarray k is , and the maximum gain is searched in the angle domain : Formula (14): ; In the formula, K , represents the number of subarray divisions, and represents the maximum gain.

[0079] The definitions of the array response function and the complex weight value both belong to the commonly used models in the classical beamforming algorithm. The array response function (array manifold) is determined by the coordinates of the antenna elements of the subarrays and the direction cosine , and is defined as a column vector , , and represents the operating wavelength; the complex weight value takes the amplitude-phase vector inner product of the manifold , denotes the conjugate vector.

[0080] By the above method, the angle corresponding to the maximum gain can be searched (including at most two dimensions of search domain), so that the main lobe is focused on the direction of the high gain path at the angle, effectively reducing the sidelobe leakage and strengthening the control of the multiple scattering channel.

[0081] S36, aiming at minimizing the total power of the antenna array, dynamically scheduling the real-time power of each subarray based on the real-time channel efficiency.

[0082] Combined with the real-time observed subarray output power and channel efficiency, high frequency power distribution scheduling is performed. The output power of subarray k at time t is denoted as , and the channel efficiency is denoted as The optimal allocation of global power consumption is performed by the following method: Equation (15): ; Constraint condition: ; Among them, denotes the maximum power that the system can provide, denotes the optimization result of the output power of each subarray.

[0083] Through continuous updating of , the influence of topographic changes or temporary obstacles can be adaptively adapted to enhance the beam coverage efficiency in complex scenarios.

[0084] S4, monitor the signal-to-noise ratio and bit error rate of the output signal of each subarray; compare the monitoring results with the multiple scattering path record, and perform real-time update of the gain coefficient and arrival time for the scattering path with accumulated deviation or phase drift; isolate the burst interference or invalid reflection signal.

[0085] As an optional implementation, step S4 includes: S41, monitor the output signal of each subarray, and collect the signal-to-noise ratio and bit error rate of the output signal of each subarray.

[0086] In some specific embodiments, a separate link monitoring component is started at the receiving end to collect the signal-to-noise ratio and bit error rate of each subarray output signal at a high frequency.

[0087] S42, associate the monitoring results with the scattering path identification determined in step S2 and the optimization result of the subarray in step S3, and identify the accumulated deviation between the monitoring results and the configuration value in the multiple scattering path.

[0088] For path The associated monitoring indicators are recorded as follows The reference pointer that matches the subarray output is marked as The cumulative deviation is calculated using the following formula. It depicts the differences in the dimensions of space and time.

[0089] Equation (16): ; In the formula, Indicates the angular coordinates within the monitoring range; Representation and path The matched multidimensional angle domain, i.e., the monitoring range; Indicates a path At time t, angle The actual observations of the monitoring indicators are derived from the echo power sequence detected in real time by the subarray; Indicates a path At time t, angle The expected value of the reference power index at that location is derived from the theoretical apparent value calculated from the calibration constant in the previous frame. and The difference is obtained after double integration in the time and angle domains. This is the cumulative bias, used to measure whether a systematic drift occurs in the measurement link over a long period. Index tolerance threshold based on offline calibration Detect the degree of prediction bias, once Exceeding This allows for early intervention, preventing adhesion errors from being gradually amplified in subsequent subarray scheduling.

[0090] S43. Perform multi-point sampling and monitoring of the phase quantity of each scattering path, and identify the cumulative phase drift between the monitoring results and the expected reference phase in the multiple scattering paths.

[0091] After a preliminary comparison between observed and theoretical values, the phase of each scattering path is sampled and monitored at multiple points to screen for significant phase drift. The subarray... k At any moment t The phase vector is denoted as The originally expected phase reference vector is marked as Calculate the cumulative phase shift : Equation (17): ; In the formula, T This indicates the monitoring duration, typically set to 128 PRIs, which can be automatically scaled based on the RMS reading from the flight track. Indicates calculation The second norm of . This reflects the overall phase drift of the scattering path over time, and is independent of the corresponding power drift. Based on a set phase tolerance threshold. ,once Exceed If the location of the geometric scatterer or the distribution of obstacles in the scattering path are found to have drifted, an alarm message is sent to the subsequent path indicator correction module to shorten the time required for fault path tracing and location.

[0092] S44. Based on the real-time monitoring of gain correction and time delay correction, update the gain coefficient and arrival time for scattering paths with accumulated deviation or phase drift to ensure that the scattering signal can still be accurately identified after abnormal terrain changes.

[0093] Current path The gain (signal-to-noise ratio) is recorded as The time delay (the arrival time of the pulse envelope) is recorded as follows: Combined with the gain correction amount calculated by real-time monitoring and delay correction amount Update the gain and latency of the path: Equation (18): .

[0094] Gain correction and delay correction amount Depend on The values ​​are calculated simultaneously using the minimum mean square error criterion, and the calculation methods are all defined within the onboard firmware parameter description module for easy traceability. Among them:

[0095] L is the number of time steps for calculation, i.e., the number of frames; This is the measured value of the m-th frame. To calculate the average of the measured values, updates are performed periodically. (i.e., gain coefficient and arrival time) can maintain accurate path positioning even under dynamic fluctuations in the scattering environment, and reduce jitter and bit error risks at the link level. The update condition is that any of the following conditions are met: 1) , 2) , 3) A total of L pulses are accumulated.

[0096] Threshold All calculations are calibrated by on-board noise and can be completed in a single FPGA cycle.

[0097] S45. The updated parameters and gain information of the multiple scattering paths are fed back to step S3 to optimize the transmit power and directional gain of each subarray.

[0098] After correcting for gain and time delay, the parameters of the scattering path (such as angle and interference) and gain information are fed back to step S3 to perform short-cycle optimization of the subarray beam pointing and transmit power. The scattering path parameters are obtained for each scattering path. First, press the meridians to obtain the pericardium. Then, use FFT (Fast Fourier Transform) to calculate the Doppler frequency offset, and use sliding window cross-correlation to obtain the time delay. The spatial spectrum of beamforming output Taking the peak value yields the azimuth and pitch. Its amplitude is higher than the threshold. Peak value for gain The interpeak difference is the degree of interference. The algorithm can be implemented using MUSIC / ESPRIT or ML, both of which have been implemented on the intermediate quantities generated in steps S31-S34 above, without the need for additional measurements.

[0099] In step S45, a beam control vector is introduced. Its elements correspond sequentially to subarrays. k Antenna weights in the array. Beamforming function. Based on the latest optimized subarray parameters, secondary beam allocation is performed, and phase coherence optimization is quickly achieved using the following formula: Equation (19): ; In the formula, Representation and angle Related wavenumber functions; Indicates the first u Spatial coordinates of each antenna element This is the phase delay term. The wavenumber function is designed as follows: .

[0100] Additionally, if sidelobe suppression is required, a window function can be added to the optimization function described above: ,in Let represent the main lobe direction angle, i.e., the reference angle to which the desired beam points. Then equation (19) can be optimized as: S46. Based on the predefined interference determination method, identify the sudden interference (or transient interference) or invalid reflection signals that occur in steps S44 and S45, and prioritize the suppression of scattering paths with sudden interference or invalid reflection signals to ensure that scattering paths without sudden interference or invalid reflection signals can preferentially obtain complete channel resources.

[0101] In some optional embodiments of this application, a temporary isolation function is established for sudden interference and invalid reflected signals to reduce resource waste caused by noise coupling.

[0102] In some alternative implementations, an interference determination function is defined. The system determines sudden interference or invalid reflected signals on a single path, and makes a comprehensive judgment based on the fused amplitude information and phase difference.

[0103] Equation (20): ; In the formula, For amplitude function, Let A and B represent the observed and expected (reference) values ​​of the path phase, respectively; A and B are the discrimination coefficients for gain and phase difference, determined by open-loop simulation, to ensure a predetermined threshold. Maximum discrimination is achieved at -6dB of the main reflection path signal. If a path is identified as experiencing sudden interference or interference echo (i.e., invalid reflected signal), it is temporarily isolated, and its power resources are immediately reallocated to higher-priority paths still within the controllable range. This is to prioritize suppressing paths with severe interference during synchronization beam or power allocation, ensuring that higher-priority paths receive more complete channel resources. The power reallocation priority is based on... Monotonically decreasing, of which Let K be the short-term pulse power variance for path i. "Short-term" refers to the number of time steps (frames) measured; this number is adjustable. For example, if K = 32 frames, then the time step size... The period is 10ms. K can take values ​​in the range of 16 to 64 frames. This design balances noise smoothing and tracking delay, and has been embedded in the on-board FPGA, adapting to different pulse repetition frequencies via instructions.

[0104] The above threshold After being calibrated offline on the ground, it is updated with the ephemeris and, under normal circumstances, has a linear compensation relationship with the temperature and power supply fluctuations of the airborne (satellite communication station carrier).

[0105] By employing the methods described above, which are based on real-time monitoring and short-term optimization, the subarray beam shape can be guaranteed to remain within the specified range. Within a certain range, with a power error of less than 0.8dB, the encroachment of multipath scattering on the main path resources is suppressed to the greatest extent.

[0106] S5. Utilize spatial filtering and sidelobe energy scheduling to dynamically schedule the main lobe energy of the high-reflectivity sub-region and attenuate or correct the phase of the sidelobe energy; perform directional suppression on sudden interference or invalid reflection signals; map the multiple scattering paths after spatial filtering and interference suppression to the frequency domain, and perform subcarrier slicing based on gain characteristics to identify high-gain carriers and low-gain carriers.

[0107] As an optional implementation, step S5 includes: S51. Based on the updated multiple scattering path record, identify local strong reflection sub-regions based on the second threshold and significant fading sub-regions based on the third threshold.

[0108] The set of scattering paths is ,definition Representing a path At time t, frequency f, angle The combined echo intensity at that location indicates a strong reflection factor. Represented as: Equation (21): ; In the formula, For path In the spatial range formed by the intersection of the time domain, frequency domain, and angular domain.

[0109] like Exceeding the set second threshold (i.e., the upper limit) indicates the existence of a local strong reflection sub-region within this path, requiring adjustment of the main lobe power allocation in subsequent steps; if Below the set third threshold If the lower limit is reached, it indicates that there is a fading region in the path, and gain scheduling needs to be performed in subsequent steps.

[0110] As an optional implementation, the second threshold and the third threshold It can be set based on a two-level strategy: statistical baseline and on-orbit adaptive approach. The strong reflection factor is calculated based on the first-order echo power integral. The global threshold is determined by dividing the data globally and using the 90th percentile as the second threshold. 10% is taken as the third threshold. The threshold can be re-estimated by sliding a window each time the nadir point crosses two longitudes, which can fit the slow drift of terrain and atmospheric decay in real time.

[0111] S52. Use spatial filtering to dynamically and adaptively schedule the main lobe energy of the local strong reflection sub-region.

[0112] Subarray k The set of antenna elements is represented as ,in, Subarray k Antenna unit in u Define subarrays k The beamforming weight vector is ,in, Subarray k Medium antenna elementu At angle The filtering weights at each position are then used to construct a main lobe filtering algorithm for the output signal of the subarray. Perform airspace weighting: Equation (22): ; In the formula, Subarray k antenna element u At angle The baseband signal received at the location.

[0113] Through the Iterative optimization (e.g., using gradient descent) is performed to maintain optimal signal strength in the high-reflectivity sub-region while suppressing signal overflow in non-target regions, thus determining a main lobe enhancement strategy combining spatial filters and subarrays. Experiments have shown that... By using the gradient descent method in three iterative steps, the main lobe gain error can be converged to <0.2dB.

[0114] S53. Based on the constraint of the interference coefficient between subarrays, the sidelobe energy is attenuated or phase corrected.

[0115] In step S53, the sidelobe scheduling effect is evaluated using the inter-array interference coefficient matrix. Equation (23): ; In the formula, Subarray Subarray The actual sidelobe overlap interference intensity between them, K, as mentioned above, represents the total number of subarrays. By setting a threshold, in the matrix M If any off-diagonal element exceeds this threshold, the pattern optimization process is triggered until the matrix... M All off-diagonal elements do not exceed the threshold. For example, if any off-diagonal element is above -25dB, fine-tuning of the pattern is triggered, and this process is repeated until all intersection elements (i.e., off-diagonal elements) are below -30dB. By... Limiting the sidelobe energy within a threshold framework, moderately attenuating or phase correcting the sidelobe energy, and reducing the coupling between parallel channels.

[0116] The fine-tuning process of the above-mentioned radiation pattern includes: Subarray The weight vector is When the matrix Any off-diagonal element above the threshold Then, enter the fine-tuning loop: 1) Calculate the current far-field radiation pattern: , and in the forbidden angle domain Leakage power is obtained: .

[0117] 2) Construct the cost function: , is the weighting factor.

[0118] 3) Use gradient descent or RLS iteration: , where the step size , and the update result is projected to make the total transmit power (or array element amplitude envelope) consistent with the previous frame, that is, the projection constraint of the total power (or amplitude) is unchanged to suppress power mutation; the update occurs in the direction (i.e., phase and relative weight), and the direction update is completed by adjusting the phase and relative weight. First, the temporary update is given according to the gradient / RLS, and then the result is projected back to the "equal power / equal amplitude" feasible region to avoid inter-frame radiation power mutation, meet the PA linear region and spectrum aperture constraints, while still being able to change the pattern. That is, the phase / direction and the relative distribution among the subarrays are updated, and the total power or amplitude envelope remains unchanged; the pattern is mainly determined by the phase difference and the relative weight, so even if the amplitude is unchanged, the leakage power integral can be reduced, and the sidelobes can be reshaped. represents the gradient of .

[0119] 4) If or the number of iterations reaches , exit and return the new weight value to the beam controller, otherwise go to step 1) and continue.

[0120] Through the above micro-amplitude-phase fine tuning, the cross-interference can be suppressed to below within dozens of iterations, while keeping the main lobe gain change <0.2dB.

[0121] S54, directional interference suppression is performed on the burst interference or invalid reflection signal identified in step S4 to suppress the interference directionally, filter out useless signals and retain valuable scattering paths.

[0122] Suppose the set of interference sources is , r is the total number of interference sources, is the index of the rth interference source. In step S54, a suppressor is designed to suppress the interference, and the output function of the suppressor is defined as : Equation (24): ; In the equation, Indicates the source of interference At angle With frequency f Distribution domain in dimensionality; Indicating targeting Targeted suppression kernel function.

[0123] Will The peak value is controlled at a low level (limited by a threshold), meaning it is required to remain within a threshold range, for example, within -35dB above the local thermal noise. If this is not met, null weights are applied within the corner slot (angle-frequency domain) until the threshold requirement is met. In this way, interference can be eliminated to the maximum extent and channel resources can be reserved for subsequent high-gain path communication.

[0124] S55. Map the multiple scattering paths after spatial filtering and interference suppression to the frequency domain, and classify the carrier frequencies into high-gain subcarriers or low-gain subcarriers with the goal of maximizing the carrier gain difference.

[0125] The carrier set is represented as , express F The first in x One carrier frequency, m This represents the total number of carriers. The high-gain subcarrier set is analyzed using the following method. and low-gain subcarrier set To distinguish: Equation (25): ; In the formula, Indicates at carrier frequency The gain function at that point Similarly.

[0126] Distinguishing high-gain subcarrier sets and low-gain subcarrier set In this case, by setting a higher modulation rate and bandwidth configuration for the high-gain subcarrier (higher than the gain subcarrier), a higher spectrum utilization rate can be achieved while ensuring system stability.

[0127] S6. Based on the gain and stability of the multiple scattering paths, they are divided into high-gain paths and fluctuating paths; the modulation order and coding rate are optimized for high-gain paths; the error correction code rate is dynamically adjusted for fluctuating paths; and group scheduling and carrier multiplexing are performed on each scattering path to maximize the overall communication throughput while meeting the requirements of bit error rate and latency.

[0128] As an optional implementation, step S6 includes: S61, respectively weighting and fusing the gain and stability of the scattering paths, and based on the size relationship between the fusion result and the fourth threshold value, respectively classifying each scattering path as a high-gain path or a fluctuation-prone path.

[0129] This step S61 mainly maps the estimated quantities of path gain and stability into a unified frequency domain space, thereby ensuring the consistency of cross-frame statistics. The spatial filtering and frequency bin slicing records output by step S5 are fused with the path gain and stability data to form a comprehensive measurement set of multiple scattering paths. The scattering path set is For each path Stability in time domain, frequency domain and angle domain and gain index are recorded respectively to construct a classification function Comprehensive measurement is performed: Equation (26): ; In the equation, characterizes the comprehensive index of the path , and represents the distribution domain of the path in time t , frequency f and angle ; respectively represent the stability weight and the gain weight; respectively represent the stability value and the gain value of the path at time t , frequency f and angle . The stability adopts the instantaneous SNR variance normalized index in a 200ms (or other value) sliding window. The frequency f is retrieved from the frequency bin slicing record, which is a fixed fragment index table formed after the onboard carrier manager discretizes the continuous bandwidth in OFDM mode; the center frequency point of each scattering path is mapped to the unique slice number in this table and written into a ring log over time, and the log is the frequency bin slicing record.

[0130] A fourth threshold value is set, if is above the fourth threshold value, the path is determined as a high-gain path, and if is below the fourth threshold value, the path is determined as a fluctuation-prone path.

[0131] In some feasible embodiments, the fourth threshold value is set by an adaptive rule: and ,but ;like exist If the above, then the path is determined to be a high-gain path. Below If the threshold is not met, the path is classified as a volatile path. Paths between the two thresholds remain unchanged; they are neither classified as high-gain paths nor volatile paths. Wherein: This represents the mean of the sliding window. N is the size of the sliding window; For sliding window standard deviation, The path classification rules are as follows: .

[0132] S62. To maximize path throughput efficiency, optimize the modulation order and coding rate of each high-gain path respectively.

[0133] Assume the set of high-gain paths is , Indicates the first j A high-gain path, p This represents the total number of high-gain paths. The modulation order of each high-gain path is N. The coding rate is denoted as .

[0134] The modulation order and coding rate are optimized using the following methods: Equation (27): ; In the formula, Indicates a high-gain path The priority weight is calculated based on the signal-to-interference-plus-noise ratio (SINR) of the path: Normalized drying ratio is expressed as , , It is calculated using the above formula (10).

[0135] Representing a path The maximum allowable bit rate under the target bit error rate requirement; This indicates the optimal modulation order and coding rate.

[0136] Modulation order N From the fixed constellation database The corresponding coding rate Then, according to the 3GPP-LDPC curve, the target block error rate is used. The results can be obtained by reverse lookup; both together constitute the optimization variables. . is the Shannon capacity approximation , B is the bandwidth, Take 1.5dB to reserve link margin.

[0137] S63, dynamic adjustment of error correction code rate for volatile paths.

[0138] The set of volatile paths is represented as , represents the i-th volatile path in the set W , j is the total number of volatile paths. q Each volatile path

[0139] is assigned an error correction code rate , between the error rate requirement and the power consumption limit, define the error correction overhead function , the error rate and the delay are jointly constrained: Equation (28): ; In the equation, the error rate comes from CRC monitoring, represents the fading depth threshold of the volatile path , represents the effective gain available in the current state, represents the error rate mapping function under , , , represents the delay mapping function under , , , is the delay penalty coefficient, set in advance.

[0140] According to equation (28), if exceeds the set threshold, that is, if , the adjustment of error correction code rate and power is triggered, and the correction is completed: To ensure that the stability requirement is still met when the fluctuation is frequent ( ), where is the preset error correction code rate fine adjustment step and power fine adjustment step. By implementing optimal scheduling on the error correction code rate , the reliability of the path can be ensured when the scattering environment fluctuates frequently.

[0141] S64, group scheduling of each scattering path with the goal of minimizing communication congestion rate.

[0142] After the configuration of the high-gain path and the volatile path is completed, the two are integrated into the group scheduling module, and the system congestion is reduced through priority allocation and load balancing.

[0143] Definition of data packet set , The i-th data packet in the set D , and N is the total number of data packets. The allocation function r is designed. s The data packet is allocated to the path . .

[0144] The traffic balancing degree is analyzed according to the following method: Equation (29): ; In the equation, represents the throughput contribution of the data packet on the path ; represents the delay of the data packet on the path , and are the weighting coefficients on the throughput contribution and the delay, respectively. In equation (29), is used to measure the contribution of the data queue lag to the throughput, represents the penalty term for the path slice switching cost.

[0145] The optimal group sequence of the high-gain path and the reserved window of the volatile path are obtained by maximizing L through heuristic iterative search on , which ensures that the core data packets obtain better service on the high-gain path while considering the load balancing of the volatile path.

[0146] S65, carrier multiplexing is performed on each scattering path to maximize the communication throughput.

[0147] After the grouping scheduling of the paths is completed, parallel multiplexing and frequency band slicing are performed on multiple paths to improve the overall throughput of the system.

[0148] The high-gain path set H and the volatile path set W are respectively mapped to the carrier frequency set F , and the multiplexing matrix is introduced in the multi-path parallel multiplexing.

[0149] Equation (30): ; In the equation, represents the high-gain path and the volatile path the conflict probability of two paths on the same frequency slice; represents the high-gain path and the volatile path the composite throughput contribution under parallel multiplexing scenario, describing the mergable bandwidth under conflict scenario.

[0150] parallel multiplexing scheduling of each scattering path by maximizing T.

[0151] S66, optimization of the coding parameters of the volatile path aiming at minimizing the error correction overhead.

[0152] combined with the monitoring of parallel leakage and real-time link information, the coding parameters of the volatile path are fine-tuned to balance the error correction overhead and system throughput rate among multiple paths.

[0153] Assume that the set of coding parameters of the volatile path is , each element represents the corresponding coding parameter, and k is the number of coding parameters. Use the dynamic adjustment function to represent the objective function of coding parameter optimization, which is derived from the Lagrange relaxation problem of the multi-dimensional knapsack problem. Where B represents the current overall bandwidth allocation. We have: Equation (31): ; In the formula, , represents the comprehensive index of the error rate BER and the code rate loss of the volatile path under the current bandwidth allocation B and coding parameters , which maps the current code rate and power combination into a comprehensive index containing error code, throughput, and bandwidth, and is represented as: , where , BER(), SNR(), and BW() represent the error rate function, the signal-to-noise ratio function, and the bandwidth function, respectively. represents the importance weight EWMA (Exponentially Weighted Moving-Average) of the volatile path , which is obtained by offline fitting of the service priority and QoS violation rate in the past period (such as one hour) by genetic algorithm, together with the subset compression of the set E, to form a convex quadratic programming, represented as , is a preset weight coefficient. By solving , the fine-tuning of the coding parameters of the volatile path is completed. The process includes: 1) Obtain the mapping results of BER, SNR, and BW on from simulation or LUT; 2) According to update , highlighting the importance of the business in recent; 3) According to formula (31) to solve , the solution method can adopt the interior point method or L-BFGS algorithm, when the number of coding parameters k<5, the single solution time is within 1ms; 4) If ( Convergence threshold), the highest is reduced and recalculated, otherwise the optimal coding parameters .

[0154] Through the above dynamic optimization process, the error correction scheme can be quickly optimized for different degrees of fading and interference, ensuring the robustness and efficiency of multiple concurrent.

[0155] According to the idea of the application, the embodiment of the application further provides a portable satellite communication station signal optimization system, comprising a storage medium and a processor, the storage medium stores a computer program, and the processor runs the computer program to execute the portable satellite communication station signal optimization method described above. In addition, the system can also include an interface circuit, which is used for the connection between the storage medium and the processor, and the connection between the processor and the detection unit.

[0156] The present application is not limited to the foregoing specific embodiments. The present application extends to any new feature or any new combination disclosed in the specification, and any new method or process steps or any new combination disclosed.

Claims

1. A method for optimizing the signal of a portable satellite communication station, characterized in that, include: S1. Verify, time-series slice, and spatial coordinate registration of radar, optical, and infrared tri-mode detection data of the detection area; A surface image is generated by fusing detection data from various modes, and attenuation correction is applied to the surface image based on meteorological observation data; the scattering intensity of the surface image is then categorized and calibrated. S2. Identify high-reflectivity sub-regions based on the surface image and its layered calibration information, and mark the high-reflectivity sub-regions with geometric scatterers; detect weak echo paths in the high-reflectivity sub-regions, and add the weak echo paths to the set of multiple scattering paths; perform discrete distribution mapping on azimuth and elevation angles and secondary time delay correction on each scattering path to generate a multiple scattering path record with gain distribution and interference identification. S3. Based on the gain, interference, and angle information of multiple scattering paths, the antenna array is divided into subarrays according to priority; phase step correction is performed on the antenna elements of each subarray, and the transmission excitation is merged in the coverage area of ​​adjacent paths; through dynamic power scheduling and beamforming algorithms, the directional gain of each subarray is optimized and sidelobe leakage is controlled.

2. The portable satellite communication station signal optimization method as described in claim 1, characterized in that, Also includes: S4. Monitor the signal-to-noise ratio and bit error rate of the output signals of each subarray; compare the monitoring results with the records of the multiple scattering paths, and perform real-time updates of the gain coefficient and arrival time for scattering paths with accumulated deviations or phase drift; isolate sudden interference or invalid reflection signals. S5. Utilize spatial filtering and sidelobe energy scheduling to dynamically schedule the main lobe energy in the high-reflectivity sub-region, and attenuate or correct the phase of the sidelobe energy; perform directional suppression on sudden interference or invalid reflection signals. The multiple scattering paths after spatial filtering and interference suppression are mapped to the frequency domain, and subcarrier slicing is performed based on gain characteristics to identify high-gain carriers and low-gain carriers. S6. Based on the gain and stability of the multiple scattering paths, they are divided into high-gain paths and fluctuating paths; the modulation order and coding rate are optimized for high-gain paths; the error correction code rate is dynamically adjusted for fluctuating paths; and each scattering path is grouped, scheduled, and multiplexed to maximize the overall communication throughput while meeting the requirements for bit error rate and latency.

3. The portable satellite communication station signal optimization method as described in claim 1, characterized in that, Step S1 includes: S11. Verify the consistency of the output of the detection unit for each modal detection data; S12. Based on the expected coverage area, the detection area is divided into sub-regions, and time-series slicing, spatial coordinate registration and scan segment index allocation are performed on each sub-region; S13. Integrate multimodal detection data into a unified coordinate system, and weight and fuse the multimodal detection data of each scan segment into a surface image. S14. Attenuation correction is performed on the surface images of each scan segment based on meteorological observation data; S15. Perform layered aggregation of scattering intensity levels on the surface images of each scan segment.

4. The portable satellite communication station signal optimization method as described in claim 1, characterized in that, Step S2 includes: S21. Calculate the reflection intensity of each sub-region and filter high-reflection sub-regions based on the first threshold. S22. Geometric scattering body marking is performed on the high-reflectivity sub-region; S23. Detect weak echo paths in the high-reflectivity sub-region of the marked geometric scatterer, and add the detected weak echo paths to the set of multiple scattering paths. S24. Map and separate the discrete distribution of each geometric scatterer in azimuth and elevation angles to separate the main reflection path and the weak echo path. S25. Perform secondary time delay verification on the offset of each separated scattering path; S26. Perform gain distribution and interference identification processing on each scattering path to generate a gain-interference mapping table.

5. The portable satellite communication station signal optimization method as described in claim 1, characterized in that, Step S3 includes: S31. Calculate the priority score of each scattering path based on the gain, interference and angle matching information of the multiple scattering paths, and extract a predetermined number / proportion of scattering paths with the highest priority scores. S32. Divide the antenna array into subarrays based on the principle that each subarray focuses on the extracted scattering path; S33. Match the desired phase of the extracted scattering path and correct the phase step of the antenna elements of each subarray. S34. Merge the emission excitations of subarrays that have adjacent path coverage areas; S35. Based on the phase step correction vector and the merged result of the emission excitation, the beamforming algorithm is used to calculate the array response function of each subarray in the three-dimensional angular space, and the scattering path corresponding to the maximum gain is searched in the angular domain so that the main lobe is focused on the direction of the scattering path. S36. With the goal of minimizing the total power of the antenna array, the real-time power of each subarray is dynamically scheduled based on the real-time channel efficiency.

6. The portable satellite communication station signal optimization method as described in claim 2, characterized in that, Step S4 includes: S41. Monitor the output signal of each subarray and collect the signal-to-noise ratio and bit error rate of the output signal of each subarray; S42. Correlate the monitoring results with the scattering path identifiers determined in step S2 and the optimization results of the subarray in step S3 to identify the cumulative deviation between the monitoring results and the configuration values ​​in the multiple scattering paths. S43. Perform multi-point sampling and monitoring of the phase quantity of each scattering path to identify the cumulative phase drift between the monitoring results and the expected reference phase in the multiple scattering paths. S44. Based on the real-time monitoring of gain correction and delay correction, update the gain coefficient and arrival time for scattering paths with accumulated deviation or phase drift. S45. The updated parameters and gain information of the multiple scattering paths are fed back to step S3 to optimize the transmit power and directional gain of each subarray. S46. Identify the sudden interference or invalid reflection signals that appear in steps S44 and S45 according to the predefined interference determination method, and prioritize the suppression of scattering paths with sudden interference or invalid reflection signals to ensure that scattering paths without sudden interference or invalid reflection signals can obtain complete channel resources first.

7. The portable satellite communication station signal optimization method as described in claim 2, characterized in that, Step S5 includes: S51. Based on the updated multiple scattering path record, identify local strong reflection sub-regions based on the second threshold and significant fading sub-regions based on the third threshold. S52. Use spatial filtering to perform dynamic adaptive scheduling of the main lobe energy of the local strong reflection sub-region. S53. Based on the constraint of the interference coefficient between subarrays, the sidelobe energy is attenuated or phase corrected. S54. Perform directional interference suppression on the sudden interference or invalid reflection signals identified in step S4. S55. Map the multiple scattering paths after spatial filtering and interference suppression to the frequency domain, and classify the carrier frequencies into high-gain subcarriers or low-gain subcarriers with the goal of maximizing the carrier gain difference.

8. The portable satellite communication station signal optimization method as described in claim 2, characterized in that, Step S6 includes: S61. The gain and stability of the scattering path are weighted and fused respectively. Based on the relationship between the fusion result and the fourth threshold, each scattering path is classified as a high-gain path or a path prone to fluctuation. S62. To maximize path throughput efficiency, optimize the modulation order and coding rate of each high-gain path respectively. S63. Dynamically adjust the error correction bit rate for paths prone to fluctuation; S64. Group scheduling of each scattering path with the goal of minimizing communication congestion rate; S65. Carrier multiplexing is performed on each scattering path with the goal of maximizing communication throughput; S66. Optimize the encoding parameters of the fluctuating path with the goal of minimizing error correction overhead.

9. A portable satellite communication station signal optimization system, comprising a storage medium and a processor, wherein the storage medium stores a computer program, characterized in that, When the processor runs the computer program, it executes the portable satellite communication station signal optimization method as described in any one of claims 1-8.

10. The portable satellite communication station signal optimization system as described in claim 9, characterized in that, It also includes an interface circuit for connecting the storage medium to the processor, and connecting the processor to the detection unit.

Images