A multi-protocol digital relay adaptive access method and system

Abstract

This invention relates to the field of wireless communication technology and proposes a method and system for adaptive access in multi-protocol digital relays. The method includes: performing cyclic spectrum phase ridge tracking on the received signal in a relay access scenario to obtain the phase offset trajectory of the received signal; based on the phase offset trajectory, performing propagation parameter analysis on the relay access candidate dimensions in the relay access scenario to obtain the path delay difference of the relay access candidate dimensions; based on the path delay difference, performing correlation structure mapping on the relay access candidate dimensions to obtain the mode features of the relay access candidate dimensions; based on the mode features, performing manifold curvature stationary point detection on the relay access candidate dimensions to obtain the extreme point distribution of the relay access candidate dimensions; based on the extreme point distribution, performing neighborhood penetration verification on the relay access adaptation request in the relay access scenario to obtain the access permission credential for the relay access adaptation request. This invention can improve the efficiency of adaptive access in multi-protocol digital relays.

Description

Technical Field

[0001] This invention relates to the field of wireless communication technology, and in particular to a multi-protocol digital relay adaptive access method and system. Background Technology

[0002] Existing multi-protocol digital relay access technologies have significant shortcomings in signal processing and parameter analysis. They lack the accuracy to track the cyclic spectrum phase ridge of the received signal and cannot reliably extract accurate phase offset trajectories. This leads to deviations in the propagation parameter analysis results of relay access candidate dimensions, making it difficult to obtain a true and effective path delay difference set and significantly reducing the reliability of basic relay access analysis.

[0003] Existing technologies suffer from poor performance in the relay access feature modeling and adaptation verification stages. They are unable to efficiently map associated structures to extract accurate pattern features, struggle to accurately identify extreme point distributions in manifold curvature stationary point detection, lack reasonable judgment logic in neighborhood penetration verification, and have weak multi-protocol adaptive access capabilities. Consequently, they cannot guarantee the stability and adaptation efficiency of relay access, making it difficult to meet the actual needs of complex relay access scenarios. Therefore, improving protocol access efficiency has become an urgent problem to be solved. Summary of the Invention

[0004] This invention provides a multi-protocol digital trunk adaptive access method and system to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides a multi-protocol digital trunk adaptive access method, comprising: S1. Perform cyclic spectrum phase ridge tracking on the received signal in the relay access scenario to obtain the phase offset trajectory of the received signal; S2. Based on the phase offset trajectory, perform propagation parameter analysis on the relay access candidate dimensions of the relay access scenario to obtain the path delay difference of the relay access candidate dimensions. S3. Based on the path delay difference, perform an association structure mapping on the relay access candidate dimension to obtain the pattern features of the relay access candidate dimension; S4. Based on the pattern features, perform manifold curvature stationary point detection on the relay access candidate dimension to obtain the extreme point distribution of the relay access candidate dimension; S5. Based on the extreme point distribution, perform neighborhood penetration verification on the relay access adaptation request in the relay access scenario to obtain the access permission credential of the relay access adaptation request.

[0006] In a preferred embodiment, the step of performing cyclic spectrum phase ridge tracing on the received signal in the relay access scenario to obtain the phase offset trajectory of the received signal includes: The received signal in the relay access scenario is accumulated by performing time-domain cyclic cross-correlation to obtain the cyclic cross-correlation tensor of the received signal; Based on the cyclic cross-correlation tensor, the frequency sampling points of the received signal are phase unwrapped and stripped to obtain the cyclic phase ridge skeleton of the received signal. The phase shift trajectory of the received signal is obtained by integrating the interlayer phase difference of the cyclic phase ridge skeleton.

[0007] In a preferred embodiment, the step of performing phase unwrapping and stripping on the cyclic frequency sampling points of the received signal based on the cyclic cross-correlation tensor to obtain the cyclic phase ridge skeleton of the received signal further includes: Median filtering is applied to the cyclic frequency slices of the cyclic cross-correlation tensor to obtain the denoised phase distribution of the cyclic frequency slices; Based on the denoised phase distribution, phase jump detection is performed on the phase values ​​of the cyclic frequency slice to obtain the frequency positions of phase value abrupt changes; By filtering the frequency locations of the phase value abrupt changes, the connection relationships of the candidate points of the cyclic frequency slice are obtained; Based on the connection relationship, candidate points on the cyclic frequency slice are connected to obtain the cyclic phase ridge skeleton of the received signal.

[0008] In a preferred embodiment, the step of performing propagation parameter analysis on the relay access candidate dimensions of the relay access scenario based on the phase offset trajectory to obtain the path delay difference of the relay access candidate dimensions includes: The phase offset trajectory is mapped to obtain the time delay scan interval of the relay access candidate dimension; The time delay scanning interval is truncated by a matched filter sliding window to obtain the filter response energy envelope of the time delay scanning interval; The energy envelope is sorted by local peaks to obtain the local peak sequence of the energy envelope; Sidelobe spurious peaks are removed from the local peak sequence to obtain the multipath arrival time index sequence of the energy envelope; The path delay difference of the relay access candidate dimension is obtained by calibrating the time difference between adjacent time indices in the index sequence.

[0009] In a preferred embodiment, the step of performing association structure mapping on the relay access candidate dimension based on the path delay difference to obtain the pattern features of the relay access candidate dimension includes: The path delay difference is constructed spatially to obtain the delay vector space of the relay access candidate dimension; The association strength is calibrated for every two sample points in the time delay vector space to obtain the pairwise similarity label between the sample points; The pairwise similarity markers are decomposed to obtain a low-dimensional coordinate set of the time delay vector space; The neighborhood assignment of adjacent coordinate points in the low-dimensional coordinate set is determined to obtain the neighborhood partitioning result of the low-dimensional coordinate set. The neighborhood partitioning results are aggregated to obtain the pattern features of the relay access candidate dimensions.

[0010] In a preferred embodiment, the step of detecting manifold curvature stationary points on the relay access candidate dimension based on the pattern features to obtain the extreme point distribution of the relay access candidate dimension includes: Local tangent space estimation is performed on the sample points in the pattern features to obtain the neighborhood tangent coordinate matrix of the sample points; Based on the neighborhood tangent coordinate matrix, the curvature tensor of the local curvature of the manifold where the sample point is located is synthesized to obtain the principal curvature value of the sample point. Extreme point discrimination is performed on the principal curvature values ​​to obtain the peak set and valley set of the principal curvature values; The peak set and valley set of the principal curvature values ​​are screened and merged to obtain the extreme point distribution of the relay access candidate dimension.

[0011] In a preferred embodiment, the step of synthesizing the curvature tensor of the manifold containing the sample point based on the neighborhood tangent coordinate matrix to obtain the principal curvature value of the sample point includes: The covariance of the neighborhood tangent coordinate matrix of the sample point is accumulated to obtain the local covariance descriptor of the sample point. Based on the local covariance descriptor, perturbation analysis is performed on the normal vector direction of the sample point to obtain the normal variation tensor of the sample point. The principal curvature values ​​of the sample points are obtained by symmetric reduction and shrunk of the normal variation tensor.

[0012] In a preferred embodiment, the step of performing neighborhood penetration verification on the relay access adaptation request in the relay access scenario based on the extreme point distribution to obtain the access permission credential for the relay access adaptation request includes: Geodesic distance diffusion is performed on the feature vector of the relay access adaptation request to obtain the geodesic distance distribution of the feature vector on the manifold where the extreme point distribution is located; Based on the geodesic distance distribution, geodesic distance field path finding is performed on each extreme point in the extreme point distribution to obtain the minimum geodesic distance between the feature vector and the extreme point distribution; The minimum geodesic distance is compared with the preset neighborhood penetration distance value to obtain the target penetration indicator of the feature vector; The protocol type is mapped to the penetration establishment flag to obtain the access permission credential for the relay access adaptation request.

[0013] In a preferred embodiment, the step of comparing the minimum geodesic distance with a preset neighborhood penetration distance value to obtain the target penetration indicator of the feature vector includes: By projecting the line connecting the feature vector to the nearest extremum into the manifold tangent space, the approach direction tangent vector of the feature vector is obtained; The curvature gradient at the nearest extreme point is analyzed for direction to obtain the curvature gradient vector at the nearest extreme point; The penetration confidence coefficient of the feature vector is calculated by substituting the cosine of the angle between the approach direction tangent vector and the curvature gradient vector into the penetration confidence coefficient formula. The formula for calculating the penetration confidence coefficient is as follows: ; This is the penetration confidence coefficient. It is a naturally exponentially decaying function. It is a hyperbolic tangent saturation function. The geodesic distance between the feature vector and the nearest extreme point is... The characteristic length of the local neighborhood of the nearest extremum point. Let be the curvature gradient vector at the nearest extreme point. The tangent vector in the approach direction of the feature vector; Based on the penetration confidence coefficient and the preset admission threshold, a threshold comparison is performed to obtain the correction factor for the amplitude comparison; Based on the correction factor, the penetration establishment mark is calibrated to obtain the target penetration establishment mark of the feature vector.

[0014] To address the above problems, the present invention also provides a multi-protocol digital trunk adaptive access system, the system comprising: The cyclic spectrum phase ridge tracking module is used to perform cyclic spectrum phase ridge tracking on the received signal in the relay access scenario to obtain the phase offset trajectory of the received signal. The propagation parameter parsing module is used to perform propagation parameter parsing on the relay access candidate dimensions of the relay access scenario based on the phase offset trajectory, and obtain the path delay difference of the relay access candidate dimensions. The association structure mapping module is used to perform association structure mapping on the relay access candidate dimension based on the path delay difference to obtain the pattern features of the relay access candidate dimension; The manifold curvature stationary point detection module is used to detect manifold curvature stationary points of the relay access candidate dimension based on the pattern features, and obtain the distribution of extreme points of the relay access candidate dimension. The neighborhood penetration verification module is used to perform neighborhood penetration verification on the relay access adaptation request in the relay access scenario based on the extreme point distribution, and obtain the access permission credential of the relay access adaptation request.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention can accurately extract the phase offset trajectory of the received signal by performing cyclic spectrum phase ridge tracing on the received signal. Based on this trajectory, the propagation parameter analysis of the candidate dimensions of relay access is completed. It can stably obtain the path delay difference set, significantly improve the accuracy of basic signal processing and parameter analysis of relay access, and ensure the efficient and stable operation of the early analysis stage of relay access.

[0016] 2. This invention relies on path delay difference sets to complete the mapping of associated structures to extract pattern features, accurately determines the distribution of extreme points through manifold curvature stationary point detection, and efficiently generates access permission certificates by combining neighborhood penetration verification. It can enhance the multi-protocol adaptive access capability, improve the adaptation efficiency and stability of relay access, and efficiently adapt to the actual access needs of various complex relay access scenarios. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating a multi-protocol digital trunk adaptive access method according to an embodiment of the present invention. Figure 2 A functional block diagram of a multi-protocol digital trunk adaptive access system provided in an embodiment of the present invention; The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0018] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0019] This application provides a multi-protocol digital trunk adaptive access method. The execution entity of this multi-protocol digital trunk adaptive access method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application embodiment: a server, a terminal, etc. In other words, the multi-protocol digital trunk adaptive access method can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0020] Reference Figure 1 The diagram shown is a flowchart illustrating a multi-protocol digital trunk adaptive access method according to an embodiment of the present invention. In this embodiment, the multi-protocol digital trunk adaptive access method and system include: S1. Perform cyclic spectrum phase ridge tracking on the received signal in the relay access scenario to obtain the phase offset trajectory of the received signal; In this embodiment of the invention, the step of performing cyclic spectrum phase ridge tracing on the received signal in the relay access scenario to obtain the phase offset trajectory of the received signal includes: The received signal in the relay access scenario is accumulated by performing time-domain cyclic cross-correlation to obtain the cyclic cross-correlation tensor of the received signal; Based on the cyclic cross-correlation tensor, the frequency sampling points of the received signal are phase unwrapped and stripped to obtain the cyclic phase ridge skeleton of the received signal. The phase shift trajectory of the received signal is obtained by integrating the interlayer phase difference of the cyclic phase ridge skeleton.

[0021] The step of performing phase unwrapping and stripping on the cyclic frequency sampling points of the received signal based on the cyclic cross-correlation tensor to obtain the cyclic phase ridge skeleton of the received signal further includes: Median filtering is applied to the cyclic frequency slices of the cyclic cross-correlation tensor to obtain the denoised phase distribution of the cyclic frequency slices; Based on the denoised phase distribution, phase jump detection is performed on the phase values ​​of the cyclic frequency slice to obtain the frequency positions of phase value abrupt changes; By filtering the frequency locations of the phase value abrupt changes, the connection relationships of the candidate points of the cyclic frequency slice are obtained; Based on the connection relationship, candidate points on the cyclic frequency slice are connected to obtain the cyclic phase ridge skeleton of the received signal.

[0022] The time-domain cyclic cross-correlation accumulation of the received signal in the relay access scenario specifically refers to multiplying the received signal with its delayed copy in the time domain point by point and accumulating the results. This accumulation process is repeated for different cyclic frequencies to construct a three-dimensional array. The dimensions of this array correspond to the cyclic frequency, the delay, and the time sampling point, respectively. This three-dimensional array is the cyclic cross-correlation tensor of the received signal.

[0023] Based on this cyclic cross-correlation tensor, the corresponding phase value sequence is extracted for each frequency sampling point of the received signal. Then, the phase values ​​of adjacent frequency sampling points are compared sequentially along the frequency axis. When the phase difference between adjacent points exceeds half a cycle, the phase value is increased or decreased by an integer multiple of the full cycle length to make it continuously unfolded. This process is called phase unwrapping and stripping. Finally, a phase curve that changes continuously with frequency is obtained. This curve is composed of the unwrapped phase values ​​of all frequency sampling points and is called the cyclic phase ridge skeleton of the received signal.

[0024] For the cyclic phase ridge skeleton, the phase difference between each layer and its adjacent upper layer is calculated along the frequency layer direction. These phase differences are accumulated layer by layer starting from the starting frequency layer. The accumulated result is the total phase shift with frequency. The curve formed by the change of this total shift with frequency is the phase shift trajectory of the received signal.

[0025] The cyclic cross-correlation tensor is sliced ​​according to the cyclic frequency dimension. Each slice corresponds to a fixed cyclic frequency value. The slice is a two-dimensional matrix, where the rows correspond to time delays and the columns correspond to time sampling points. The amplitude spectrum is obtained by taking the modulus of each element of the two-dimensional matrix. The amplitude spectrum is then subjected to median filtering, which replaces the original value of each element with the median of the intensity values ​​in the neighborhood of each element. The filtered amplitude spectrum is combined with the original phase to obtain the denoised phase distribution of each cyclic frequency slice.

[0026] For the denoised phase distribution of each cyclic frequency slice, the phase difference between the current frequency point and its adjacent frequency points is calculated point by point along the frequency axis. When the absolute value of the difference exceeds the set threshold, it is determined that a phase jump has occurred at that position. The coordinates of all frequency points that have jumped are recorded. These coordinates constitute the set of frequency positions where the phase value changes abruptly.

[0027] For all frequency locations of phase value abrupt changes, the eight-neighborhood connectivity rule is used for marking. That is, if two abrupt change locations are spatially adjacent (sharing an edge or corner), they are considered to belong to the same connected region. The abrupt change locations in each connected region are arranged in ascending order of frequency, and adjacent locations are connected in sequence to form continuous paths. The connection relationship of these paths is recorded as an adjacency list between candidate points.

[0028] Based on these connections, starting from the candidate points at the lowest frequency layer, the process proceeds layer by layer to higher frequency layers along the connection path. At each layer, candidate points that are connected to the selected candidate points in the previous layer are selected as continuations. If there are multiple selections, all branches are retained. Finally, all selected candidate points are connected in sequence according to the frequency layer to form one or more continuous ridges. The set of these ridges is the cyclic phase ridge skeleton of the received signal.

[0029] The beneficial effects of this invention are that it extracts the phase offset trajectory by tracking the phase ridge of the cyclic spectrum, detects the extreme points of the positioning protocol features by combining the manifold curvature stationary point detection, and uses neighborhood penetration verification for access decision. Under the condition of no need for training samples and threshold adjustment, it significantly improves the adaptive access accuracy and noise robustness of multi-protocol relay.

[0030] S2. Based on the phase offset trajectory, perform propagation parameter analysis on the relay access candidate dimensions of the relay access scenario to obtain the path delay difference of the relay access candidate dimensions. In this embodiment of the invention, the step of performing propagation parameter analysis on the relay access candidate dimensions of the relay access scenario based on the phase offset trajectory to obtain the path delay difference of the relay access candidate dimensions includes: The phase offset trajectory is mapped to obtain the time delay scan interval of the relay access candidate dimension; The time delay scanning interval is truncated by a matched filter sliding window to obtain the filter response energy envelope of the time delay scanning interval; The energy envelope is sorted by local peaks to obtain the local peak sequence of the energy envelope; Sidelobe spurious peaks are removed from the local peak sequence to obtain the multipath arrival time index sequence of the energy envelope; The path delay difference of the relay access candidate dimension is obtained by calibrating the time difference between adjacent time indices in the index sequence.

[0031] Mapping the phase offset trajectory specifically refers to establishing a correspondence between each phase offset value in the phase offset trajectory and a preset time delay search range. That is, the phase offset is converted into the start and end points of the time delay according to a fixed conversion factor. The continuous interval covered by these start and end points is the time delay scan interval of the relay access candidate dimension.

[0032] For this time delay scanning interval, a matched filter window function of known shape is used. Starting from the beginning position of the time delay scanning interval, the window function is gradually moved to the right with a fixed sliding step size. After each movement, the signal energy within the interval covered by the window function is calculated and the energy value is recorded. The curve formed by connecting the energy values ​​corresponding to all sliding positions is the filter response energy envelope of the time delay scanning interval.

[0033] For this energy envelope, the energy values ​​of three adjacent points are compared sequentially from left to right along the time delay axis. If the energy value of a point is greater than the energy values ​​of its left and right neighboring points, then the point is marked as a local peak. All the marked local peaks are arranged in the order in which they appear on the time delay axis, and the resulting sequence is the local peak sequence of the energy envelope.

[0034] For this local peak sequence, the peak with the largest energy value in the sequence is identified as the main peak. Then, the amplitude of each of the remaining peaks is checked in turn. If the amplitude of a peak is less than half of the amplitude of the main peak, the peak is identified as a sidelobe pseudo-peak and removed from the sequence. After removing all sidelobe pseudo-peaks, the remaining peaks are arranged in the original order. The sequence formed by their corresponding time delay positions is the multipath arrival time index sequence of the energy envelope.

[0035] For the multipath arrival time index sequence, starting from the second element, calculate the difference between the current element and the previous element in sequence. This difference represents the interval between the arrival times of two different propagation paths. Arrange all the calculated interval values ​​in the order of the corresponding paths. The resulting numerical sequence is the path delay difference of the relay access candidate dimension.

[0036] The beneficial effects of this invention are that it extracts path delay differences through cyclic spectrum phase ridge tracking and propagation parameter analysis, combines manifold curvature stationary point detection to locate extreme point distribution, and finally uses neighborhood penetration verification to complete protocol adaptive access. Under the condition of no prior training and manual parameter adjustment, it significantly improves the identification accuracy and access success rate of multi-protocol digital relays in complex channel environments.

[0037] S3. Based on the path delay difference, perform an association structure mapping on the relay access candidate dimension to obtain the pattern features of the relay access candidate dimension; In this embodiment of the invention, the step of performing association structure mapping on the relay access candidate dimension based on the path delay difference to obtain the pattern features of the relay access candidate dimension includes: The path delay difference is constructed spatially to obtain the delay vector space of the relay access candidate dimension; The association strength is calibrated for every two sample points in the time delay vector space to obtain the pairwise similarity label between the sample points; The pairwise similarity markers are decomposed to obtain a low-dimensional coordinate set of the time delay vector space; The neighborhood assignment of adjacent coordinate points in the low-dimensional coordinate set is determined to obtain the neighborhood partitioning result of the low-dimensional coordinate set. The neighborhood partitioning results are aggregated to obtain the pattern features of the relay access candidate dimensions.

[0038] Constructing the spatial dimension of path delay difference means taking the path delay difference value corresponding to each sample point as a coordinate component of the sample point in a multidimensional space. The coordinate components of all sample points together define a high-dimensional space. Each dimension of this high-dimensional space corresponds to the delay difference value of a propagation path. This space spanned by the coordinates of all sample points is the delay vector space of the relay access candidate dimension.

[0039] For every two sample points in the time delay vector space, calculate the difference between their coordinates in each dimension, square the difference in each dimension and sum them up, then take the square root of the summation. The resulting value represents the Euclidean distance between the two sample points. The reciprocal of this Euclidean distance is taken as the correlation strength between the two sample points. The greater the correlation strength, the more similar the two sample points are. Repeat the above process for all possible pairs of sample points in the space. All the correlation strength values ​​obtained are organized in the order of the sample point pairs, which is the pairwise similarity label between the sample points.

[0040] Feature decomposition of pairwise similarity labels involves constructing a square matrix with the number of rows and columns equal to the number of sample points. The element in the i-th row and j-th column of the square matrix is ​​filled with the pairwise similarity label between the i-th and j-th sample points. Then, the eigenvalues ​​and eigenvectors of the square matrix are solved, and the eigenvectors corresponding to the largest few eigenvalues ​​are extracted. These eigenvectors are used as new coordinate axes, and the original coordinates of each sample point are projected onto these new coordinate axes. The set of projected coordinate values ​​is the low-dimensional coordinate set of the time delay vector space.

[0041] Determining the neighborhood affiliation of adjacent coordinate points in a low-dimensional coordinate set involves calculating the Euclidean distance between each coordinate point and all other coordinate points, considering the nearest points as its neighborhood members, and considering the two points as having a neighborhood affiliation relationship if one point appears in the neighborhood member list of another point. After traversing all coordinate points, each pair of points with a neighborhood affiliation relationship is recorded. All such pairs of points together constitute the neighborhood partitioning result of the low-dimensional coordinate set.

[0042] Class aggregation of neighborhood partitioning results refers to connecting point pairs with direct or indirect neighborhood affiliation to form a connected graph. All coordinate points within each connected component in the graph are considered to belong to the same category. All connected components are extracted from the neighborhood partitioning results. Each connected component contains a set of interconnected coordinate points. These coordinate points are organized according to their original order in the low-dimensional coordinate set. Each connected component corresponds to a pattern class. The set of all pattern classes is the pattern feature of the relay access candidate dimension.

[0043] The beneficial effects of this invention are that by mapping path delay differences to pattern features on a low-dimensional manifold, it effectively eliminates redundant correlations in multipath delay data, enhances the feature separability between different protocols, and provides a highly discriminative geometric basis for subsequent stationary point detection and penetration verification.

[0044] S4. Based on the pattern features, perform manifold curvature stationary point detection on the relay access candidate dimension to obtain the extreme point distribution of the relay access candidate dimension; In this embodiment of the invention, the step of detecting manifold curvature stationary points on the relay access candidate dimension based on the pattern features to obtain the extreme point distribution of the relay access candidate dimension includes: Local tangent space estimation is performed on the sample points in the pattern features to obtain the neighborhood tangent coordinate matrix of the sample points; Based on the neighborhood tangent coordinate matrix, the curvature tensor of the local curvature of the manifold where the sample point is located is synthesized to obtain the principal curvature value of the sample point. Extreme point discrimination is performed on the principal curvature values ​​to obtain the peak set and valley set of the principal curvature values; The peak set and valley set of the principal curvature values ​​are screened and merged to obtain the extreme point distribution of the relay access candidate dimension.

[0045] The step of synthesizing the curvature tensor of the manifold containing the sample point based on the neighborhood tangent coordinate matrix to obtain the principal curvature value of the sample point includes: The covariance of the neighborhood tangent coordinate matrix of the sample point is accumulated to obtain the local covariance descriptor of the sample point. Based on the local covariance descriptor, perturbation analysis is performed on the normal vector direction of the sample point to obtain the normal variation tensor of the sample point. The principal curvature values ​​of the sample points are obtained by symmetric reduction and shrunk of the normal variation tensor.

[0046] For each sample point in the pattern features, select several other sample points that are closest to it in the feature space via Euclidean distance as its neighborhood points. Subtract the coordinates of the sample point from the original coordinates of all neighborhood points to obtain a set of offset vectors. Arrange these offset vectors into a matrix, with each row of the matrix corresponding to the offset vector of a neighborhood point. Then, perform orthogonal triangular decomposition on the matrix and take the first few column vectors of the orthogonal matrix to form a new matrix. This new matrix is ​​the neighborhood tangent coordinate matrix of the sample point.

[0047] Based on the neighborhood tangent coordinate matrix of the sample point, each column of the neighborhood tangent coordinate matrix is ​​first regarded as a tangent vector. The inner product between each pair of these tangent vectors is calculated and filled into a square matrix to obtain the local metric matrix of the sample point. Then, matrix decomposition is performed on the local metric matrix to extract the anisotropic coefficients describing the curvature of the surface. These coefficients are combined into a higher-order tensor according to specific rules. This tensor reflects the curvature amplitude of the manifold at the sample point in each direction, that is, the curvature tensor corresponding to the local curvature of the manifold at the sample point. Eigenvalues ​​are extracted from this curvature tensor to obtain two non-zero eigenvalues. These two eigenvalues ​​are the principal curvature values ​​of the sample point.

[0048] For each sample point, the principal curvature value is compared with the principal curvature values ​​of all other sample points in the feature space. If the principal curvature value of a sample point is greater than the principal curvature values ​​of all its neighboring sample points, then the sample point is marked as a peak of the principal curvature value. If the principal curvature value of a sample point is less than the principal curvature values ​​of all its neighboring sample points, then the sample point is marked as a valley of the principal curvature value. The set of all sample points marked as peaks is the set of peaks of the principal curvature value, and the set of all sample points marked as valleys is the set of valleys of the principal curvature value.

[0049] For the peak set and valley set of principal curvature values, the elements in the two sets are first merged into a temporary set. Then, it is checked whether there are duplicate points that are very close in the feature space in the merged set. If the two points come from the peak set and valley set respectively and their Euclidean distance is less than a fixed threshold, only the point with the larger absolute value of principal curvature is retained and the other point is removed. After such screening, the remaining points are organized according to their original positions in the feature space, and the new set formed is the extreme point distribution of the relay access candidate dimension.

[0050] For a neighborhood tangent coordinate matrix of a sample point, each column of the matrix is ​​regarded as a tangent vector. The mean vector of all tangent vectors is calculated. Then, the mean vector is subtracted from each tangent vector to obtain a centered vector. The centered vector is multiplied by its own transpose to obtain a square matrix. The square matrices corresponding to all centered vectors are summed and divided by the number of neighborhood points. The resulting matrix is ​​the local covariance descriptor of the sample point.

[0051] Based on this local covariance descriptor, we first solve for the eigenvector corresponding to the smallest eigenvalue of the descriptor. This eigenvector indicates the approximate direction of the manifold normal at the sample point. Then, for each neighboring point in the neighborhood, we calculate the projection length of the offset vector of that neighboring point in the normal direction. We arrange the projection lengths of all neighboring points into a vector field according to the spatial distribution of the neighboring points. We take the partial derivative of this vector field along each direction of the tangent plane, and the resulting gradient tensor is the normal variation tensor of the sample point.

[0052] For the normal variation tensor, first transpose it to obtain another tensor, then add the original tensor and transpose the tensor and divide by two to obtain a symmetric tensor. Then orthogonally diagonalize the symmetric tensor and extract the two principal elements on the diagonal. These two principal elements are the principal curvature values ​​of the sample point.

[0053] The beneficial effects of this invention are that it accurately extracts the principal curvature values ​​on the manifold through local tangent space estimation and curvature tensor synthesis, and obtains a stable distribution of extreme points by screening and merging peak and valley points. It achieves adaptive analysis of the manifold geometry without the need for a preset reference template, which significantly improves the accuracy of protocol feature localization and noise resistance.

[0054] S5. Based on the extreme point distribution, perform neighborhood penetration verification on the relay access adaptation request in the relay access scenario to obtain the access permission credential of the relay access adaptation request.

[0055] In this embodiment of the invention, the step of performing neighborhood penetration verification on the relay access adaptation request in the relay access scenario based on the extreme point distribution to obtain the access permission credential for the relay access adaptation request includes: Geodesic distance diffusion is performed on the feature vector of the relay access adaptation request to obtain the geodesic distance distribution of the feature vector on the manifold where the extreme point distribution is located; Based on the geodesic distance distribution, geodesic distance field path finding is performed on each extreme point in the extreme point distribution to obtain the minimum geodesic distance between the feature vector and the extreme point distribution; The minimum geodesic distance is compared with the preset neighborhood penetration distance value to obtain the target penetration indicator of the feature vector; The protocol type is mapped to the penetration establishment flag to obtain the access permission credential for the relay access adaptation request.

[0056] The step of comparing the minimum geodesic distance with the preset neighborhood penetration distance value to obtain the target penetration indicator of the feature vector includes: By projecting the line connecting the feature vector to the nearest extremum into the manifold tangent space, the approach direction tangent vector of the feature vector is obtained; The curvature gradient at the nearest extreme point is analyzed for direction to obtain the curvature gradient vector at the nearest extreme point; The penetration confidence coefficient of the feature vector is calculated by substituting the cosine of the angle between the approach direction tangent vector and the curvature gradient vector into the penetration confidence coefficient formula. The formula for calculating the penetration confidence coefficient is as follows: ; This is the penetration confidence coefficient. It is a naturally exponentially decaying function. It is a hyperbolic tangent saturation function. The geodesic distance between the feature vector and the nearest extreme point is... The characteristic length of the local neighborhood of the nearest extremum point. Let be the curvature gradient vector at the nearest extreme point. The tangent vector in the approach direction of the feature vector; Based on the penetration confidence coefficient and the preset admission threshold, a threshold comparison is performed to obtain the correction factor for the amplitude comparison; Based on the correction factor, the penetration establishment mark is calibrated to obtain the target penetration establishment mark of the feature vector.

[0057] For the feature vector of the relay access adaptation request, the distance information is propagated in all directions from the location of the feature vector on the manifold where the extreme point distribution is located, simulating the thermal diffusion process. Each point on the manifold records the shortest path length from the feature vector to that point. The scalar field formed by the shortest path lengths of all points according to their positions on the manifold is the geodesic distance distribution of the feature vector on the manifold where the extreme point distribution is located.

[0058] Based on this geodesic distance distribution, for each extreme point in the extreme point distribution, start from the extreme point and move step by step along the direction of the fastest descent of the geodesic distance gradient to the nearest point. At each step, choose the direction that reduces the geodesic distance until the minimum point of the geodesic distance distribution, i.e. the location of the feature vector, is reached. Record the maximum change of geodesic distance along the entire path. The minimum value selected from the geodesic distance values ​​corresponding to the shortest path that can reach the feature vector among all extreme points is the minimum geodesic distance between the feature vector and the extreme point distribution.

[0059] The minimum geodesic distance is compared with a pre-stored neighborhood penetration distance value. If the minimum geodesic distance is less than or equal to the neighborhood penetration distance value, it is determined that the feature vector has passed through the neighborhood of the extreme point, and a true value flag is generated. Otherwise, a false value flag is generated. The true value or false value flag is the penetration flag of the feature vector.

[0060] The penetration establishment flag is compared with the predefined protocol type table in the relay access scenario. Based on the protocol category of the extreme point associated with the penetration establishment flag, the corresponding protocol type is selected as the output. The protocol type is encoded into a control command that the device can recognize. This control command is the access permission credential for the relay access adaptation request.

[0061] During the amplitude comparison process, the eigenvector is connected to the extremum point closest to the eigenvector in the extremum point distribution by a straight line segment. This straight line segment is projected onto the tangent plane of the manifold at the extremum point, and the resulting vector is the tangent vector of the approach direction of the eigenvector.

[0062] At the nearest extreme point, calculate the rate of change of curvature value along each direction of the manifold surface. The direction in which the curvature value changes the fastest with direction is the curvature gradient direction. Extract the unit vector in this direction, and this vector is the curvature gradient vector at the nearest extreme point.

[0063] The geodesic distance between the eigenvector and the nearest extremum is derived from the minimum geodesic distance obtained through pathfinding using the geodesic distance field in the steps described above. The eigenvalue length of the local neighborhood of the nearest extremum is derived from the average geodesic distance from all sample points within a certain range around the extremum to that extremum. The curvature gradient vector is derived from solving for the directional derivative of the curvature field at the nearest extremum, i.e., calculating the rate of change of curvature along each direction and taking the unit vector of the direction with the largest rate of change. The approaching direction tangent vector is derived from the projection of the line connecting the eigenvector and the nearest extremum onto the tangent space of the manifold.

[0064] The penetration confidence coefficient is constructed from two physical dimensions. The exponential decay term measures how close the feature vector is to an extremum in geometric space; the smaller the geodesic distance, the closer this term is to one, indicating high spatial proximity. The hyperbolic tangent term measures the consistency between the direction of motion of the feature vector approaching the extremum and the direction of change of the manifold curvature; the more consistent the directions, the closer this term is to two, indicating high directional plausibility. The penetration confidence coefficient, obtained by multiplying these two terms, falls between zero and two and is used to comprehensively determine whether an access request constitutes a genuine protocol switching behavior.

[0065] When the geodesic distance between the feature vector and the nearest extremum is much greater than the feature length, the exponential term approaches zero, causing the overall penetration confidence coefficient to approach zero, indicating that the distance is too great to trigger a penetration decision. When the geodesic distance is much smaller than the feature length, the exponential term approaches one, and the penetration confidence coefficient is mainly determined by directional consistency. If the approaching tangent vector is in the same direction as the curvature gradient vector, the cosine of the angle approaches one, the hyperbolic tangent term approaches two, and the overall coefficient approaches two, indicating the strongest penetration confidence. If the directions are opposite, the cosine of the angle approaches negative one, the hyperbolic tangent term approaches zero, and the overall coefficient approaches zero, indicating that even if the distance is very close, penetration is rejected due to unreasonable direction. If the directions are perpendicular, the cosine of the angle is zero, the hyperbolic tangent term is one, and the overall coefficient is approximately one, indicating a moderate confidence level that needs to be combined with other criteria.

[0066] The penetration confidence coefficient is compared with a preset admission threshold. If the penetration confidence coefficient is greater than the admission threshold, a positive correction factor is generated; otherwise, a negative correction factor is generated. This correction factor is used to adjust the result of the amplitude comparison.

[0067] The penetration validation flag is calibrated using this correction factor. Specifically, the true or false value of the penetration validation flag is combined with the sign of the correction factor. If the correction factor is positive and the penetration validation flag is true, the result is true. If the correction factor is negative and the penetration validation flag is false, the result is true. Otherwise, the result is false. The result obtained after calibration is the target penetration validation flag of the feature vector.

[0068] The beneficial effects of this invention are that by combining geodesic distance field pathfinding and directional consistency decision, and introducing a penetration confidence coefficient that combines exponential decay and hyperbolic tangent, the noise resistance and directional sensitivity of multi-protocol relay access decision are significantly improved without the need for manual threshold setting, and the probability of false alarms and missed alarms is reduced.

[0069] like Figure 2 The diagram shown is a functional block diagram of a multi-protocol digital trunk adaptive access system provided in an embodiment of the present invention.

[0070] The multi-protocol digital trunk adaptive access system described in this invention can be installed in an electronic device. Depending on the functions implemented, the multi-protocol digital trunk adaptive access system may include a cyclic spectrum phase ridge tracking module, a propagation parameter analysis module, an association structure mapping module, a manifold curvature stationary point detection module, and a neighborhood penetration verification module. The modules described in this invention can also be referred to as units, which are a series of computer program segments that can be executed by the processor of an electronic device and perform a fixed function, and are stored in the memory of the electronic device.

[0071] In this embodiment, the functions of each module / unit are as follows: The cyclic spectrum phase ridge tracking module is used to perform cyclic spectrum phase ridge tracking on the received signal in the relay access scenario to obtain the phase offset trajectory of the received signal. The propagation parameter parsing module is used to perform propagation parameter parsing on the relay access candidate dimension of the relay access scenario based on the phase offset trajectory, and obtain the path delay difference of the relay access candidate dimension. The association structure mapping module is used to perform association structure mapping on the relay access candidate dimension based on the path delay difference to obtain the pattern features of the relay access candidate dimension. The manifold curvature stationary point detection module is used to detect manifold curvature stationary points of the relay access candidate dimension based on the pattern features, and obtain the distribution of extreme points of the relay access candidate dimension. The neighborhood penetration verification module is used to perform neighborhood penetration verification on the relay access adaptation request in the relay access scenario based on the extreme point distribution, and obtain the access permission credential of the relay access adaptation request.

[0072] In the several embodiments provided by this invention, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0073] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0074] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0075] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0076] This application embodiment can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0077] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A multi-protocol digital trunk adaptive access method, characterized in that, The method includes: S1. Perform cyclic spectrum phase ridge tracking on the received signal in the relay access scenario to obtain the phase offset trajectory of the received signal; S2. Based on the phase offset trajectory, perform propagation parameter analysis on the relay access candidate dimensions of the relay access scenario to obtain the path delay difference set of the relay access candidate dimensions; S3. Based on the path delay difference set, perform association structure mapping on the relay access candidate dimension to obtain the pattern features of the relay access candidate dimension; S4. Based on the pattern features, perform manifold curvature stationary point detection on the relay access candidate dimension to obtain the extreme point distribution of the relay access candidate dimension; S5. Based on the extreme point distribution, perform neighborhood penetration verification on the relay access adaptation request in the relay access scenario to obtain the access permission credential of the relay access adaptation request.

2. The multi-protocol digital trunk adaptive access method as described in claim 1, characterized in that, The step of performing cyclic spectrum phase ridge tracking on the received signal in the relay access scenario to obtain the phase offset trajectory of the received signal includes: The received signal in the relay access scenario is accumulated by performing time-domain cyclic cross-correlation to obtain the cyclic cross-correlation tensor of the received signal; Based on the cyclic cross-correlation tensor, the frequency sampling points of the received signal are phase unwrapped and stripped to obtain the cyclic phase ridge skeleton of the received signal. The phase shift trajectory of the received signal is obtained by integrating the interlayer phase difference of the cyclic phase ridge skeleton.

3. The multi-protocol digital trunk adaptive access method as described in claim 2, characterized in that, The step of performing phase unwrapping and stripping on the cyclic frequency sampling points of the received signal based on the cyclic cross-correlation tensor to obtain the cyclic phase ridge skeleton of the received signal further includes: Median filtering is applied to the cyclic frequency slices of the cyclic cross-correlation tensor to obtain the denoised phase distribution of the cyclic frequency slices; Based on the denoised phase distribution, phase jump detection is performed on the phase values ​​of the cyclic frequency slice to obtain the frequency positions of phase value abrupt changes; By filtering the frequency locations of the phase value abrupt changes, the connection relationships of the candidate points of the cyclic frequency slice are obtained; Based on the connection relationship, candidate points on the cyclic frequency slice are connected to obtain the cyclic phase ridge skeleton of the received signal.

4. The multi-protocol digital trunk adaptive access method as described in claim 1, characterized in that, The step of performing propagation parameter analysis on the relay access candidate dimensions of the relay access scenario based on the phase offset trajectory to obtain the path delay difference of the relay access candidate dimensions includes: The phase offset trajectory is mapped to obtain the time delay scan interval of the relay access candidate dimension; The time delay scanning interval is truncated by a matched filter sliding window to obtain the filter response energy envelope of the time delay scanning interval; The energy envelope is sorted by local peaks to obtain the local peak sequence of the energy envelope; Sidelobe spurious peaks are removed from the local peak sequence to obtain the multipath arrival time index sequence of the energy envelope; The path delay difference of the relay access candidate dimension is obtained by calibrating the time difference between adjacent time indices in the index sequence.

5. The multi-protocol digital trunk adaptive access method as described in claim 1, characterized in that, The step of performing an association structure mapping on the relay access candidate dimensions based on the path delay difference to obtain the pattern features of the relay access candidate dimensions includes: The path delay difference is constructed spatially to obtain the delay vector space of the relay access candidate dimension; The association strength is calibrated for every two sample points in the time delay vector space to obtain the pairwise similarity label between the sample points; The pairwise similarity markers are decomposed to obtain a low-dimensional coordinate set of the time delay vector space; The neighborhood assignment of adjacent coordinate points in the low-dimensional coordinate set is determined to obtain the neighborhood partitioning result of the low-dimensional coordinate set. The neighborhood partitioning results are aggregated to obtain the pattern features of the relay access candidate dimensions.

6. The multi-protocol digital trunk adaptive access method as described in claim 1, characterized in that, The step of detecting manifold curvature stationary points for the relay access candidate dimension based on the pattern features, and obtaining the extreme point distribution of the relay access candidate dimension, includes: Local tangent space estimation is performed on the sample points in the pattern features to obtain the neighborhood tangent coordinate matrix of the sample points; Based on the neighborhood tangent coordinate matrix, the curvature tensor of the local curvature of the manifold where the sample point is located is synthesized to obtain the principal curvature value of the sample point. Extreme point discrimination is performed on the principal curvature values ​​to obtain the peak set and valley set of the principal curvature values; The peak set and valley set of the principal curvature values ​​are screened and merged to obtain the extreme point distribution of the relay access candidate dimension.

7. The multi-protocol digital trunk adaptive access method as described in claim 6, characterized in that, The step of synthesizing the curvature tensor of the manifold containing the sample point based on the neighborhood tangent coordinate matrix to obtain the principal curvature value of the sample point includes: The covariance of the neighborhood tangent coordinate matrix of the sample point is accumulated to obtain the local covariance descriptor of the sample point. Based on the local covariance descriptor, perturbation analysis is performed on the normal vector direction of the sample point to obtain the normal variation tensor of the sample point. The principal curvature values ​​of the sample points are obtained by symmetric reduction and shrunk of the normal variation tensor.

8. The multi-protocol digital trunk adaptive access method as described in claim 1, characterized in that, The process of performing neighborhood penetration verification on relay access adaptation requests in the relay access scenario based on the extreme point distribution to obtain access permission credentials for the relay access adaptation requests includes: Geodesic distance diffusion is performed on the feature vector of the relay access adaptation request to obtain the geodesic distance distribution of the feature vector on the manifold where the extreme point distribution is located; Based on the geodesic distance distribution, geodesic distance field path finding is performed on each extreme point in the extreme point distribution to obtain the minimum geodesic distance between the feature vector and the extreme point distribution; The minimum geodesic distance is compared with the preset neighborhood penetration distance value to obtain the target penetration indicator of the feature vector; The protocol type is mapped to the penetration establishment flag to obtain the access permission credential for the relay access adaptation request.

9. The multi-protocol digital trunk adaptive access method as described in claim 8, characterized in that, The step of comparing the minimum geodesic distance with the preset neighborhood penetration distance value to obtain the target penetration indicator of the feature vector includes: By projecting the line connecting the feature vector to the nearest extremum into the manifold tangent space, the approach direction tangent vector of the feature vector is obtained; The curvature gradient at the nearest extreme point is analyzed for direction to obtain the curvature gradient vector at the nearest extreme point; The penetration confidence coefficient of the feature vector is calculated by substituting the cosine of the angle between the approach direction tangent vector and the curvature gradient vector into the penetration confidence coefficient formula. The formula for calculating the penetration confidence coefficient is as follows: ； This is the penetration confidence coefficient. It is a naturally exponentially decaying function. It is a hyperbolic tangent saturation function. The geodesic distance between the feature vector and the nearest extreme point is... The characteristic length of the local neighborhood of the nearest extremum point. Let be the curvature gradient vector at the nearest extreme point. The tangent vector in the approach direction of the feature vector; Based on the penetration confidence coefficient and the preset admission threshold, a threshold comparison is performed to obtain the correction factor for the amplitude comparison; Based on the correction factor, the penetration establishment mark is calibrated to obtain the target penetration establishment mark of the feature vector.

10. A multi-protocol digital trunk adaptive access system, characterized in that, For implementing the multi-protocol digital trunk adaptive access method and system as described in claim 1, the system includes: The cyclic spectrum phase ridge tracking module is used to perform cyclic spectrum phase ridge tracking on the received signal in the relay access scenario to obtain the phase offset trajectory of the received signal. The propagation parameter parsing module is used to perform propagation parameter parsing on the relay access candidate dimensions of the relay access scenario based on the phase offset trajectory, and obtain the path delay difference of the relay access candidate dimensions. The association structure mapping module is used to perform association structure mapping on the relay access candidate dimension based on the path delay difference to obtain the pattern features of the relay access candidate dimension; The manifold curvature stationary point detection module is used to detect manifold curvature stationary points of the relay access candidate dimension based on the pattern features, and obtain the distribution of extreme points of the relay access candidate dimension. The neighborhood penetration verification module is used to perform neighborhood penetration verification on the relay access adaptation request in the relay access scenario based on the extreme point distribution, and obtain the access permission credential of the relay access adaptation request.

Images