A hybrid time service method based on Beidou satellite navigation and TSN

Abstract

The application relates to the technical field of time synchronization, and discloses a mixed time service method based on Beidou satellite navigation and TSN, which comprises the following steps: acquiring network topology structure information and historical time service error data of each synchronization node in a target distributed system, and constructing a dynamic time delay state set for jitter identification; based on the dynamic time delay state set, a current global time reference is extracted through a time calibration algorithm, a time difference reference vector sequence is generated, and initial synchronization adjustment is carried out on a TSN master node; according to the response stability of a slave node, a clock convergence function curve between nodes is constructed, whether the synchronization network is in a low jitter interval is judged based on time offset rate and time drift trend change; in combination with the node time state after directional offset suppression, a multi-channel redundant time adjustment mechanism in a time window is applied, a weight adjustment coefficient is extracted, and a non-uniform calibration factor is introduced; and according to a regression correction result, a time service stability index is evaluated in real time. The application has the advantages of improving time service stability.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of time synchronization, in particular to a hybrid time service method based on Beidou satellite navigation and TSN. BACKGROUND

[0002] In modern distributed systems, industrial control networks and intelligent infrastructures, accurate time synchronization is of great significance to ensure the stable operation of the system, data consistency and event coordination. Traditional network time service methods mainly include synchronization mechanisms based on network time protocol and precision time protocol, which usually calculate the message round-trip time delay between the master clock and the slave clock to achieve time adjustment and alignment. This kind of method can provide a certain degree of synchronization effect in simple or local area network environment, but in complex or large-scale network structure, the problem of unstable network delay is common. Specifically, in the time service process, the time service message in the transmission link is easily disturbed by network congestion, queuing delay, link jitter, routing change and other factors, resulting in unpredictable time delay. This time delay uncertainty will directly introduce time drift error, and then affect the time consistency of multiple nodes in the system. When the time service error reaches a certain degree, it is easy to cause data time sequence disorder, abnormal scheduling execution, and even cause the failure of critical task systems, especially in power dispatching, traffic control, high-speed communication and industrial manufacturing, etc. High reliability scene, this problem is particularly prominent. Therefore, it is necessary to design a hybrid time service method based on Beidou satellite navigation and TSN to improve the stability of time service. SUMMARY

[0003] In view of the defects of the prior art, the application provides a hybrid time service method based on Beidou satellite navigation and TSN, which has the advantages of improving the stability of time service and solving the problems in the background art.

[0004] In order to achieve the above purpose of improving the stability of time service, the application provides the following technical scheme: a hybrid time service method based on Beidou satellite navigation and TSN, comprising the following steps:

[0005] Obtain the network topology structure information and historical time service error data of each synchronization node in the target distributed system, combine the link state fluctuation characteristics and message time delay record, and construct a dynamic time delay state set for jitter identification;

[0006] Based on the dynamic time delay state set, the current global time reference is extracted by a time calibration algorithm, a time difference reference vector sequence is generated, the TSN master node is initially synchronized and adjusted, and the response stability of each slave node is monitored;

[0007] According to the response stability of the slave node, a clock convergence function curve between nodes is constructed, whether the synchronization network is in a low jitter interval is judged based on the time offset rate and the time drift trend change, if yes, a clock consistency strategy scheduling process is triggered to suppress the abnormal clock node by directional offset;

[0008] In combination with the node time state after the directional offset suppression, a multi-channel redundant time mechanism within a time window is applied, a weight adjustment coefficient is extracted and a non-uniform calibration factor is introduced, and a cross-node clock is fitted and corrected;

[0009] According to the regression correction result, the time service stability index is evaluated in real time, and a synchronization consistency evaluation report is output.

[0010] Preferably, the dynamic time delay state set process for jitter identification is as follows:

[0011] The link data packet forwarding delay, queue length, routing path switching frequency and time service error statistical information of each synchronization node within multiple time windows are collected;

[0012] The link data is normalized and reorganized according to the physical topology and logical connection relationship between nodes;

[0013] A multi-dimensional time delay fluctuation feature vector group is constructed based on the reorganized data, and the time delay change trend is extracted through a window sliding mechanism;

[0014] An improved time sequence clustering algorithm is used to classify and model the time delay feature vector group, and a node cluster with common jitter behavior is identified;

[0015] The node cluster and the corresponding time delay fluctuation model are fused with the historical error record to construct an instability quantitative index set describing the current synchronization state as a dynamic time delay state set.

[0016] Preferably, the current global time reference extraction process through the time calibration algorithm is as follows:

[0017] Multiple Beidou satellite navigation signals are received, and high-precision timestamp information contained therein is extracted;

[0018] According to the position, signal pseudo-range and signal-to-noise ratio of the time service receiving node, a pseudo-range joint least squares algorithm is applied for preliminary time solution;

[0019] In combination with the multi-path offset correction model of the received signal and the atmospheric delay compensation function, the solution result is adjusted for deviation;

[0020] A Kalman filter is introduced to estimate and smooth the time deviation sequence within multiple solution periods to obtain a global time reference.

[0021] Preferably, the process of generating the time difference reference vector sequence is as follows:

[0022] The TSN master node and several typical slave nodes were selected as a synchronization control group.

[0023] Using the current global time base as a reference, record the difference in local timestamps between the master and slave nodes within the same time period;

[0024] The timestamp differences are filtered and smoothed to form a basic timestamp vector.

[0025] A node weighting factor is introduced, and the time difference vector is weighted by combining link transmission stability, historical error fluctuation level and topology location sensitivity.

[0026] The final output is a time difference reference vector sequence.

[0027] Preferably, the process of constructing the clock convergence function curve between nodes is as follows:

[0028] Collect the timestamps of each slave node's response to the master node's synchronization signal within a continuous time period;

[0029] Based on the time difference reference vector sequence, the evolution trajectory of the alignment error of each node is calculated;

[0030] The error evolution trajectory is fitted to a time function curve, and the function trend factor is extracted by combining the convergence speed and error fluctuation range of each node.

[0031] The convergence trajectory is smoothed using the sliding window averaging method, and a clock convergence function curve is constructed to reflect the trend of node synchronization state changes.

[0032] Preferably, the process of determining whether the synchronization network is in the low jitter range is as follows:

[0033] Based on the constructed clock convergence function curve, extract the time offset rate and drift trend change value of each node within the preset time window;

[0034] Calculate the synchronization standard deviation and maximum time deviation among all nodes to form a set of network synchronization jitter indicators;

[0035] Set a threshold range for judgment, including the maximum allowable jitter amplitude and the minimum stable value of the synchronization standard;

[0036] When the jitter index of all nodes in the network is lower than the set threshold, and the synchronization standard deviation remains within a stable range for several consecutive cycles, it is determined to be in the low jitter range.

[0037] Preferably, the process of suppressing directional offsets for abnormal clock nodes is as follows:

[0038] Identify nodes whose offset rate deviates from the network average from the clock convergence function curve;

[0039] By combining the historical stability data of abnormal clock nodes with their topological location in the network, the offset suppression priority of abnormal clock nodes is calculated.

[0040] Based on the master node's timing trajectory and the drift direction of abnormal clock nodes, construct a one-sided or two-sided suppression strategy and generate corresponding time correction instructions.

[0041] Suppression commands are injected into the TSN time synchronization scheduling path to ensure that correction information arrives at the target node within the precise time slot;

[0042] The effect of offset suppression was verified by reviewing the correction results and the changes in alignment error before and after suppression.

[0043] Preferably, the fitting regression correction process for clocks across nodes is as follows:

[0044] Based on the suppressed offset node time state, reconstruct the time synchronization error matrix between master and slave nodes;

[0045] A multi-channel redundancy mechanism is introduced to compare the BeiDou time reference with the time difference within the TSN network in parallel.

[0046] A weighted regression algorithm is used to establish a cross-node time fitting model, fit the error distribution, and extract the fitting residuals;

[0047] An unbalanced calibration factor is introduced during the fitting process, and nodes with poor stability are given adaptively adjusted weights.

[0048] Preferably, the process for outputting the synchronization consistency assessment report is as follows:

[0049] Summarize the alignment error, drift trend, jitter index and clock stability score of all nodes after regression correction;

[0050] The above indicators are compared with the set evaluation threshold library to generate synchronization status labels for each node.

[0051] Construct a network synchronization consistency scoring matrix to mark the trajectories of abnormal nodes and areas with potential synchronization risks;

[0052] Extract the set of nodes with insufficient stability to form a target list for the next round of synchronization optimization and control, and output a synchronization consistency assessment report.

[0053] Compared with existing technologies, this invention provides a hybrid timing method based on BeiDou satellite navigation and TSN, which has the following advantages:

[0054] The application effectively overcomes the time drift problem caused by link fluctuation, asymmetric path transmission and time delay jitter in traditional network time service by fusing the high-precision global time reference provided by Beidou satellite and the deterministic synchronization mechanism under TSN architecture; the dynamic delay state set construction mechanism and multi-cycle error modeling means are adopted to accurately identify the node cluster with synchronization state anomaly in the network; the Beidou and TSN multi-source time reference are fused based on Kalman filtering and weighted regression algorithms, which significantly improves the precision and robustness of clock alignment; by constructing the clock convergence function curve between nodes, the dynamic quantification of node synchronization trend and fluctuation state is realized, which helps to identify and suppress abnormal clock offset path; finally, the synchronization consistency evaluation report is provided to provide the system-level synchronization state diagnosis result and control basis, which has the advantages of high stability, high scalability and high adaptability, and is especially suitable for key scenes such as distributed industrial control system, intelligent manufacturing platform, wide-area sensor network and vehicle network with strict time consistency requirements. BRIEF DESCRIPTION OF DRAWINGS

[0055] Figure 1 The step flowchart of the application. DETAILED DESCRIPTION

[0056] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the protection scope of the application.

[0057] Embodiment 1: please refer to Figure 1 The hybrid time service method based on Beidou satellite navigation and TSN according to the application, as shown in the drawings, comprises the following steps:

[0058] S1: obtaining the network topology structure information and historical time service error data of each synchronization node in the target distributed system, combining the link state fluctuation characteristics and message delay record between nodes, and constructing a dynamic delay state set for jitter identification.

[0059] The process of constructing the dynamic delay state set for jitter identification in S1 is:

[0060] Collect the link data packet forwarding delay, queue length, routing path switching frequency and timing error statistical information of each synchronization node in multiple time windows; set a unified time reference benchmark, and divide the entire timing system into multiple sliding time windows, each window period being 1 to 10 seconds, which is dynamically adjusted according to the network rate; in each time window, the time stamps of the data packet sending and receiving of each node are grabbed in real time through the TSN synchronization communication protocol, and the single-hop and multi-hop link forwarding delay is calculated; at the same time, the queue length and queue waiting time in the queue of the switching node are collected and extracted as the characteristics reflecting the link congestion state; the routing path switching events occurring in the network are monitored, and the path switching times and triggering time are recorded, which are used to characterize the path fluctuation frequency; the timing deviation records of each node in the time window are summarized to form an error statistical sequence, and a basic link behavior data set is formed;

[0061] The link data is normalized and reorganized according to the physical topology and logical connection relationship between nodes; for the numerical features collected, such as forwarding delay, queue length, error deviation, etc., Z-score standardization method is used for normalization processing, so that the mean value is 0 and the standard deviation is 1, so as to eliminate the influence of dimension; for non-numeric features, such as path switching state, One-Hot encoding or event frequency statistics method is used to convert them into numerical input; combined with the topology structure diagram and link connection table configured in the system, the normalized features are rearranged according to the node pair dimension to form a multi-dimensional feature vector with each pair of adjacent nodes as the unit; the rearranged structure is unified into a three-dimensional tensor with time window sequence, node pair identifier and feature dimension as the dimensions;

[0062] Based on the reorganized data, a multi-dimensional delay fluctuation feature vector group is constructed, and the delay change trend is extracted through the window sliding mechanism; in the above three-dimensional data tensor, a sliding sub-sequence is formed by selecting a plurality of continuous time windows, and the change rate of the feature values in each sub-sequence is calculated; for the forwarding delay, queue length and other indicators, the mean, variance, maximum and minimum values and first derivative in the window are calculated to capture the change trend; a delay fluctuation feature vector group of each node pair in a given time period is formed, and the structure of the vector group retains the time sequence characteristics; at the same time, the window position and corresponding time stamp are recorded to ensure that the fluctuation trend can be bound with a specific time period;

[0063] The improved time series clustering algorithm is used for classifying and modeling the time delay feature vector group, and identifying the node cluster with common jitter behavior; the dynamic time warping distance measurement method is used for similarity calculation of the feature change curves of different node pairs; the improved K-shape clustering algorithm is introduced for clustering analysis of all fluctuation feature vector groups; the clustering quality is evaluated by indicators such as the contour coefficient and the Davies-Bouldin index, and the optimal cluster number is determined; after clustering, the node pairs belonging to the same cluster are merged into a time delay behavior similar cluster, which is marked as a node cluster with common jitter characteristics;

[0064] The node cluster and the corresponding time delay fluctuation model are fused with the historical error record to construct a set of instability quantitative indicators describing the current synchronization state as a dynamic time delay state set.

[0065] S2: Based on the dynamic time delay state set, the current global time reference is extracted by a time calibration algorithm, a time difference reference vector sequence is generated, the TSN master node is initially synchronized and adjusted, and the response stability of each slave node is monitored.

[0066] The process of extracting the current global time reference in S2 by the time calibration algorithm is as follows:

[0067] Receive multiple Beidou satellite navigation signals and extract high-precision timestamp information contained therein; configure a GNSS receiving module on the target time transfer node to support simultaneous reception of multiple frequency signals such as Beidou system B1I / B3I; each Beidou satellite signal frame contains UTC reference time, high-precision intra-week time and millisecond-level time transfer labels; the receiving module extracts the navigation message field containing the timestamp by synchronously detecting the satellite signal main carrier and the navigation text; based on ephemeris data, the current visible satellite list is solved, and low-elevation and low-SNR inferior satellite signals are removed; the time transfer information of high-quality satellites is cached in the time transfer data buffer;

[0068] According to the position, signal pseudo-range and signal-to-noise ratio of the time transfer receiving node, a pseudo-range joint least squares algorithm is applied for preliminary time solution; the pseudo-range calculation formula is:

[0069] ;

[0070] In the formula, is the pseudo-range between the receiving node and the i-th satellite, c is the speed of light, is the receiving time; is the transmission time of the i-th satellite, is the path offset term, is the error term;

[0071] The position (x, y, z) of the receiver and the time deviation of the receiver are jointly estimated by using at least 4 Beidou satellite signals, and the weight coefficient is adjusted according to the SNR index of each satellite, and a higher SNR value is given a greater weight to reduce the interference of weak signals; the output of the solution result includes the time deviation estimation value of the current node relative to UTC;

[0072] The deviation adjustment is performed on the solution result in combination with the multi-path offset correction model of the received signal and the atmospheric delay compensation function; the local scene terrain model is introduced to judge the building and terrain reflection conditions, and the relative time delay of the reflection path and the main path is calculated; the code phase difference model is constructed by using the carrier phase difference measured by the dual-frequency receiver to eliminate the pseudo-range delay caused by the building; the maximum expectation-based pseudo-range residual error model is applied to correct the errors of the satellite signals affected by the multi-path effect;

[0073] The Kalman filter is introduced to estimate and smooth the time deviation sequence in multiple solution periods to obtain a global time reference.

[0074] The process of generating the time difference reference vector sequence in S2 is as follows:

[0075] The TSN master node and a plurality of typical slave nodes are selected as a synchronization control group; in the distributed TSN time service network, the current master clock node is identified through the network management module; according to the network topology structure and the node communication frequency, a plurality of slave nodes representative of the master node are preferentially selected; the selected nodes are marked with a synchronization monitoring mark, and a master-slave node control relationship table is established, which contains the unique identification, physical connection path and historical error range of each pair of synchronization nodes; the number of nodes in the synchronization control group should be not less than 3 to ensure that the reference value has a certain statistical representativeness and anti-interference ability;

[0076] The local time stamp difference of the master-slave nodes in the same time period is recorded with the current global time reference as the reference; a unified sampling window and an alignment period are set; in each synchronization period, based on the global time reference extracted from the Beidou system, the time service request of the master node and all control slave nodes is triggered; the current local system clock value of each node is collected to form the master node time and the corresponding item of the slave node time The original time difference value between the master and the slave is calculated, and the formula is as follows: Where i is the slave node number; the sequence and the sampling time stamp are recorded to form a group of original time difference samples;

[0077] The time stamp difference value is error filtered and smoothed to form a basic time difference vector;

[0078] Introducing node weight factor, combining link transmission stability, historical error fluctuation level and topology position sensitivity to weight the time difference vector; assigning a weight factor to each slave node, considering the following indicators:

[0079] Link stability coefficient: calculated based on the link transmission delay fluctuation range in the last several periods;

[0080] Historical error stability coefficient: evaluated by the historical time service deviation standard deviation;

[0081] Topology sensitivity coefficient: depends on the hierarchical depth and path dependence of the node in the topology structure;

[0082] Finally output the time difference reference vector sequence.

[0083] S3: According to the response stability of the slave node, construct the clock convergence function curve between nodes, judge whether the synchronization network is in the low jitter interval based on the time offset rate and time drift trend change, if yes, trigger the clock consistency strategy scheduling process, and suppress the directional offset of the abnormal clock node.

[0084] The process of constructing the clock convergence function curve between nodes in S3 is:

[0085] Collect the response time stamp of each slave node to the master node synchronization signal in continuous time period; establish a synchronization communication mechanism based on IEEE 802.1AS between the master node and each slave node, the master node broadcasts synchronization frame regularly, during each synchronization frame transmission process, the slave node records the receiving time stamp, the master node records the sending time stamp, each slave node calculates the deviation value between its local time and the master node time through frame comparison, records the deviation value sequence of each slave node in multiple continuous synchronization periods, forms the time stamp time sequence;

[0086] Based on the time difference reference vector sequence, calculate the evolution trajectory of the alignment error of each node; obtain the time difference sequence between the master node and the slave node at the corresponding moment from the generated time difference reference vector sequence, for each node, compare the actual response time stamp of the node with the weighted reference time difference of the node, get the instantaneous alignment error, sequence the instantaneous alignment error according to time dimension, form the error evolution trajectory sequence of the node;

[0087] Fit the error evolution trajectory into a time function curve, and extract the function trend factor combined with the convergence speed and error fluctuation range of each node;

[0088] Smooth the convergence trajectory by using the sliding window average method, and construct the clock convergence function curve to reflect the trend of the change of the node synchronization state.

[0089] The process of judging whether the synchronization network is in the low jitter interval in S3 is:

[0090] According to the constructed clock convergence function curve, the time offset rate and the drift trend change value of each node in the preset time window are extracted; the size of the sliding time window, for example, 5 consecutive time periods, is set to capture the dynamic trend of the clock offset change; for each synchronization node, according to the constructed clock convergence function curve, the time alignment error at each moment in the window is extracted; the time offset rate of the node in the window, that is, the average growth rate of the time offset, is calculated, which is used to evaluate the deviation speed of the node from the master node; and the drift trend change value, that is, the first-order difference average of the time offset, is calculated, which is used to quantify the fluctuation intensity of the offset value in the window, and reflects the synchronization fluctuation behavior of the node.

[0091] The synchronization standard deviation and the maximum time deviation value between all nodes are calculated to form a network synchronization jitter index set; in each time period, the current offset of all nodes forms an offset data set; the standard deviation of the offset data set is calculated to reflect the dispersion degree of the alignment error of each node in the entire network, which is used as a quantitative index of synchronization consistency; at the same time, the difference between the maximum offset and the minimum offset is extracted to obtain the maximum time deviation value, which represents the most extreme time error phenomenon in the network; the time offset rate, the drift trend, the synchronization standard deviation and the maximum deviation value together form the network synchronization jitter index set, which is used to comprehensively evaluate the stability of the synchronization network.

[0092] The determination threshold interval is set, including the maximum allowed jitter amplitude and the minimum synchronization standard stable value; a set of standard thresholds for judging the synchronization state is preset, including:

[0093] The maximum allowed offset rate threshold, for example, not more than 50 nanoseconds per second;

[0094] The maximum drift trend change threshold, for example, the average fluctuation amplitude is not more than 20 nanoseconds;

[0095] The synchronization standard deviation tolerance upper limit, for example, not more than 100 nanoseconds;

[0096] The maximum time deviation upper limit, for example, the time difference between any two nodes is not more than 200 nanoseconds;

[0097] When the jitter indicators of all nodes in the network are lower than the set threshold, and the synchronization standard deviation remains in the stable interval for multiple periods, it is determined to be in the low jitter interval.

[0098] The directional offset suppression process of the abnormal clock node in S3 is:

[0099] Identify the nodes with offset rate deviating from the average value of the whole network from the clock convergence function curve; analyze the clock convergence function curve of all synchronization nodes, and extract the time offset rate of each node; calculate the average value and standard deviation of the time offset rate of the whole network, and construct a reference alignment reference interval; compare the offset rate of each node with the average value of the whole network, and if the absolute offset amplitude exceeds the set multiple threshold, mark the node as an abnormal offset node; the identified node set is used as the candidate target for subsequent offset suppression;

[0100] Combine the historical stability data of the abnormal clock node and the topological position in the network to calculate the abnormal clock node offset suppression priority; extract the historical time service deviation record of each abnormal node, and calculate the historical stability score by counting the synchronization standard deviation, maximum offset and average drift direction in the past several periods; evaluate the criticality of the node in the network topology, such as whether it is at the edge, whether it is a bridging forwarding node, whether it has a multi-hop forwarding path, etc., and generate a topological sensitivity coefficient; combine the historical stability score and the topological sensitivity coefficient, and use a weighted ranking model to calculate the suppression priority index of each abnormal node, which is used to determine the processing order and strategy strength;

[0101] According to the main node time service trajectory and the drift direction of the abnormal clock node, construct a one-sided or two-sided suppression strategy, and generate the corresponding time correction instruction; compare the relative change trend of the current offset direction of the abnormal node and the main node time service trajectory; if the node only offsets in one direction and has a stable drift slope, a one-sided suppression strategy is adopted, that is, a one-time correction timestamp is sent to the node; if the offset direction of the node fluctuates greatly or the offset amount presents periodic oscillation, a two-sided suppression strategy is adopted, that is, upper and lower limits are set in the synchronization window, and the node convergence interval is adjusted through continuous suppression; according to the suppression strategy, construct the time correction instruction, which includes the suppression target node, the correction value, the correction direction and the execution timestamp;

[0102] Inject the suppression instruction into the TSN time synchronization scheduling path to ensure that the correction information reaches the target node within the accurate time slot; analyze the scheduling path and time slot configuration of the TSN network, and select appropriate idle time slots or management time slots in the GCL cycle to inject the suppression instruction; encapsulate the time correction command using the management frame format supported by IEEE 802.1AS or 802.1Qbv protocol to ensure that it can be identified as a high-priority synchronization control message; issue the correction frame in the network control plane through the controller or scheduler to ensure that it reaches the target node along the optimal path within the specified synchronization cycle;

[0103] Verify the offset suppression effect by checking the correction result and the change of alignment error before and after suppression.

[0104] Embodiment 2: as Figure 1As shown, a hybrid time service method based on Beidou satellite navigation and TSN, further comprising the following steps:

[0105] S4: In combination with the suppressed node time state of the directional offset, the multi-channel redundant time synchronization mechanism within the time window is applied to extract the weight adjustment coefficient and introduce the non-uniform calibration factor, and the clock across nodes is fitted and corrected.

[0106] The fitting and correction process of the clock across nodes in S4 is as follows:

[0107] Based on the node time state after the offset suppression, the time synchronization error matrix between the master and slave nodes is reconstructed; after completing the offset suppression of the abnormal node, the time alignment difference between all master and slave nodes at the current time is collected; taking the master node as the reference, the local timestamp difference between it and each slave node is recorded to construct a preliminary error matrix, and the error matrix is processed to remove abnormal data points such as signal packet loss and timeout;

[0108] The multi-channel redundancy mechanism is introduced to compare the Beidou time reference and the internal time difference of the TSN network in parallel; while obtaining the internal error matrix of the TSN network, the high-precision UTC timestamp corresponding to the current time of the master node is obtained through the Beidou time service module; the current Beidou timestamp is also bound to all slave nodes, and a double-channel time difference list is constructed through the TSN time and Beidou time corresponding to the master and slave nodes respectively; the consistency and deviation of the TSN time service result and the Beidou time service data are identified to form a cross-channel time difference reference comparison table;

[0109] A cross-node time fitting model is established using a weighted regression algorithm to fit the error distribution and extract the fitting residual;

[0110] In the fitting process, a non-uniform calibration factor is introduced to give adaptive adjustment weight to nodes with poor stability; for each node, the error fluctuation range, offset rate change trend and historical error stability coefficient are counted within multiple cycles to form a comprehensive stability score, and a non-uniform calibration factor is introduced, the worse the stability, the larger the calibration factor.

[0111] S5: According to the regression correction result, the time service stability index is evaluated in real time, and a synchronization consistency evaluation report is output.

[0112] The process of outputting the synchronization consistency evaluation report in S5 is as follows:

[0113] Collecting the alignment error, drift trend, jitter index and clock stability score of each node after regression correction; after completing the regression correction step, for each synchronization node, extracting the residual error sequence of the latest time, that is, the difference between the fitting value and the actual alignment time; statistics of residual error, calculating the average alignment error, maximum offset and error fluctuation range of each node; at the same time, based on the error change slope in the continuous time period, extract the drift trend parameters of the node, such as the offset direction and change speed; using the standard deviation of error and drift rate and other data, generate synchronization jitter index to reflect the stability of the node time synchronization; collect the above items, combine the node historical performance and current response behavior, and calculate the clock stability score of the node through the preset weight model;

[0114] Compare the above indexes with the set evaluation threshold library to generate the synchronization state label of each node;

[0115] Construct a network synchronization consistency score matrix to mark the abnormal node trajectory and potential synchronization hazard area; according to the synchronization state label and stability score of each node, construct a node-level synchronization score table; combine the network topology relationship between nodes to combine the scores of each node into a two-dimensional matrix, and the row and column represent the node number and the key synchronization attribute respectively; visually mark the nodes with significantly low scores in the matrix, and track the time sequence score change trajectory to identify the error evolution path; if multiple problem nodes exist in regional aggregation or link correlation, mark them as potential synchronization hazard areas, and prompt that there is a topological structural risk;

[0116] Extract the node set with insufficient stability to form the target list of the next round of synchronization optimization and control, and output the synchronization consistency evaluation report.

[0117] It should be noted that in this paper, relationship terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Moreover, the term "include", "contain" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment.

[0118] Although the embodiments of the present application have been shown and described, it can be understood by those of ordinary skill in the art that various changes, modifications, replacements and variations can be made to the embodiments without departing from the principles and spirits of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A hybrid time service method based on Beidou satellite navigation and TSN, characterized in that, The method comprises the following steps: Obtain network topology information and historical time service error data of each synchronization node in the target distributed system, combine link state fluctuation characteristics and message delay records to construct a dynamic delay state set for jitter identification; Based on the dynamic delay state set, extract the current global time reference through a time calibration algorithm, generate a time difference reference vector sequence, perform initial synchronization adjustment on the TSN master node, and monitor the response stability of each slave node; According to the response stability of the slave nodes, construct a clock convergence function curve between the nodes, judge whether the synchronization network is in a low jitter interval based on the time offset rate and time drift trend changes, and if so, trigger a clock consistency strategy scheduling process to suppress the directional offset of the abnormal clock node; Combine the time state of the node after directional offset suppression, apply a multi-channel redundant time mechanism within a time window, extract a weight adjustment coefficient and introduce an unbalanced calibration factor to perform fitting and regression correction on the clock between nodes; According to the regression correction result, real-time evaluate the time service stability index, and output a synchronization consistency evaluation report.

2. The hybrid time service method based on Beidou satellite navigation and TSN according to claim 1, characterized in that, The process of constructing a dynamic delay state set for jitter identification is as follows: Collect link data packet forwarding delay, queue length, routing path switching frequency and time service error statistical information of each synchronization node within multiple time windows; Normalize the link data, and reorganize the data according to the physical topology and logical connection relationship between nodes; Based on the reorganized data, construct a multi-dimensional delay fluctuation feature vector group, and extract the delay change trend through a window sliding mechanism; Use an improved time sequence clustering algorithm to classify and model the delay feature vector group, and identify node clusters with common jitter behavior; Fuse the node clusters and the corresponding delay fluctuation model with the historical error records to construct a set of instability quantitative indicators describing the current synchronization state as a dynamic delay state set.

3. The hybrid time service method based on Beidou satellite navigation and TSN according to claim 1, characterized in that, The process of extracting the current global time reference through a time calibration algorithm is as follows: Receive multiple Beidou satellite navigation signals and extract high-precision timestamp information contained therein; According to the position, signal pseudorange and signal-to-noise ratio of the time service receiving node, apply a pseudorange joint least squares algorithm for preliminary time solution; Combine the multi-path offset correction model of the received signal and the atmospheric delay compensation function to adjust the solution result; Introduce Kalman filtering to estimate and smooth the time deviation sequence within multiple solution periods to obtain the global time reference.

4. The hybrid time service method based on Beidou satellite navigation and TSN according to claim 1, characterized in that, The process of generating a time difference reference vector sequence is as follows: Select the TSN master node and several typical slave nodes as a synchronization control group; Record the local timestamp difference of the master and slave nodes within the same time period with the current global time reference as the reference; Error filter and smooth the timestamp difference to form a basic time difference vector; Introduce a node weight factor, combine link transmission stability, historical error fluctuation level and topological position sensitivity to weight the time difference vector; Finally output the time difference reference vector sequence.

5. The hybrid time service method based on Beidou satellite navigation and TSN according to claim 1, characterized in that, The process of constructing a clock convergence function curve between nodes is as follows: Collect the response timestamp of each slave node to the master node synchronization signal within a continuous time period; Based on the time difference reference vector sequence, the evolution trajectory of the alignment error of each node is calculated; The error evolution trajectory is fitted as a time function curve, and the function trend factor is extracted by combining the convergence speed and error fluctuation range of each node; The sliding window average method is used to smooth the convergence trajectory, and the clock convergence function curve is constructed to reflect the trend of the change in the synchronization state of the node.

6. The hybrid time service method based on Beidou satellite navigation and TSN according to claim 1, characterized in that, The process of determining whether the synchronization network is in the low jitter interval is as follows: According to the constructed clock convergence function curve, the time offset rate and drift trend change value of each node in the preset time window are extracted; The synchronization standard deviation and the maximum time deviation value between all nodes are calculated to form a network synchronization jitter index set; Set the threshold interval, including the maximum allowed jitter amplitude and the minimum synchronization standard stable value; When the jitter index of all nodes in the network is lower than the set threshold, and the synchronization standard deviation remains in the stable interval for multiple periods, it is determined to be in the low jitter interval.

7. The hybrid time service method based on Beidou satellite navigation and TSN according to claim 1, characterized in that, The process of directional offset suppression of abnormal clock nodes is as follows: From the clock convergence function curve, identify the nodes whose offset rate deviates from the average value of the whole network; Combine the historical stability data of the abnormal clock node and its topological position in the network to calculate the offset suppression priority of the abnormal clock node; According to the time tracking trajectory of the master node and the drift direction of the abnormal clock node, construct a one-sided or two-sided suppression strategy, and generate the corresponding time correction instruction; Inject the suppression instruction into the TSN time synchronization scheduling path to ensure that the correction information reaches the target node within the accurate time slot; Verify the offset suppression effect by checking the correction results and the change in alignment error before and after suppression.

8. The hybrid time service method based on Beidou satellite navigation and TSN according to claim 1, characterized in that, The process of fitting and regression correction of cross-node clocks is as follows: Based on the time state of the nodes after the offset suppression, reconstruct the time error matrix between the master and slave nodes; Introduce a multi-channel redundancy mechanism to compare the Beidou time reference and the internal time difference value of the TSN network in parallel; Use a weighted regression algorithm to establish a cross-node time fitting model to fit the error distribution and extract the fitting residual; Introduce an unbalanced calibration factor in the fitting process to give adaptive adjustment weight to nodes with poor stability.

9. The hybrid time service method based on Beidou satellite navigation and TSN according to claim 1, characterized in that, The process of outputting the synchronization consistency evaluation report is as follows: Summarize the corresponding alignment error, drift trend, jitter index and clock stability score of all nodes after regression correction; Compare the above indexes with the set evaluation threshold library to generate the synchronization state label of each node; Construct a network synchronization consistency score matrix to mark the abnormal node trajectory and potential synchronization hidden danger area; Extract the node set with insufficient stability to form the target list of the next round of synchronization optimization control, and output the synchronization consistency evaluation report.

Images