A dangerous goods waybill verification method and device based on space-time trajectory matching and a medium

Abstract

The application discloses a dangerous goods transport order verification method and device based on space-time trajectory matching, and a medium, relates to the technical field of big data processing, and comprises the following steps: boundary positioning is performed on a trajectory missing interval in real-time trajectory verification information, boundary states before and after the trajectory missing interval are extracted, a candidate path is deduced in combination with road space-time constraints, unreachable paths are filtered out according to the continuous accessibility of the boundary states before and after the trajectory missing interval, and possible trajectories are obtained; the current state of each vehicle in the possible trajectories is extracted, mapped to a dangerous goods transport order task constraint graph, an initial executable domain of remaining tasks is constructed, and the initial executable domain of the remaining tasks is executed layer by layer according to the current state of each vehicle as a starting point to perform reachable propagation and constraint contraction, and a branch task executable domain is obtained. The application realizes joint judgment of the task fulfillment situation of the occurred tasks and the continuous executability of the remaining tasks, and improves the accuracy of the dangerous goods transport order full-link verification in the trajectory discontinuous scene.

Description

Technical Field

[0001] This invention relates to the field of big data processing technology, and in particular to a method, equipment and medium for verifying dangerous goods cargo manifests based on spatiotemporal trajectory matching. Background Technology

[0002] Dangerous goods waybill verification is a crucial part of the safety supervision of dangerous goods transportation. Currently, the industry mainly relies on vehicle-mounted positioning terminals to collect transportation spatiotemporal data. With the large-scale application of big data processing capabilities, dangerous goods waybill verification is gradually integrating vehicle positioning records, road network data, and waybill execution information. Conventional methods typically involve first matching vehicle positioning records with roads to obtain vehicle travel trajectories, and then performing spatiotemporal comparisons and batch verifications of loading, in-transit transportation, arrival unloading, and delivery based on the areas traversed by the trajectory, the time of stay, and the waybill plan. This has become a routine regulatory process, providing data support for transportation risk prevention and control.

[0003] However, conventional verification methods mostly rely on a single path or existing trajectory as the basis for verification, and do not conduct multi-branch spatiotemporal reachability extrapolation for the trajectory gap caused by positioning interruption. This makes it difficult to fully identify and distinguish possible trajectories within the gap. They focus on compliance judgment of transportation behavior that has already occurred, and do not simultaneously verify the continuous executability of the remaining tasks under different possible trajectories, making it difficult to adapt to the end-to-end verification requirements in scenarios with discontinuous trajectories. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a dangerous goods shipment verification method based on spatiotemporal trajectory matching to solve the problems of insufficient trajectory identification in multiple branches within a trajectory missing interval and difficulty in synchronously verifying the feasibility of fulfilling the remaining tasks across the entire chain in scenarios with discontinuous trajectories.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, the present invention provides a dangerous goods transport manifest verification method based on spatiotemporal trajectory matching, comprising: performing task-oriented parsing on dangerous goods transport verification data, sequentially dividing the manifest execution information into task nodes, configuring spatiotemporal and business constraints for each task node and the transfer between nodes, and constructing a dangerous goods transport manifest task constraint graph; based on the dangerous goods transport manifest task constraint graph, performing road matching on vehicle positioning records, obtaining road matching trajectory sequences, and dividing reliable trajectory segments and trajectory missing intervals according to the spatiotemporal continuity relationship of adjacent trajectory points, while determining completed task nodes according to the spatiotemporal correspondence between reliable trajectory segments and task nodes, generating real-time trajectory verification information; and processing the trajectory missing intervals in the real-time trajectory verification information. Boundary localization is performed within the lost interval, and the states of the preceding and following boundaries are extracted. Candidate paths are deduced in conjunction with road spatiotemporal constraints. At the same time, inaccessible paths are filtered out based on the continuous reachability of the preceding and following boundary states to obtain possible trajectories. The current state of the vehicles on each possible trajectory is extracted and mapped to the hazardous goods transport order task constraint graph to construct the initial executable domain of the remaining tasks. Starting from the current state of each vehicle, the initial executable domain of the remaining tasks is subjected to layer-by-layer reachability propagation and constraint contraction to obtain the executable domain of the branch tasks. Based on the hazardous goods transport order task constraint graph, the possible trajectories and the corresponding branch task executable domains are dynamically verified at the set level. The satisfaction of each branch with the constraints of the already occurred tasks and the remaining tasks is compared to determine the verification status of the hazardous goods transport order.

[0007] As a preferred embodiment of the dangerous goods shipment manifest verification method based on spatiotemporal trajectory matching described in this invention, the steps for constructing the dangerous goods shipment manifest task constraint graph are as follows: Arrange the waybill execution information in the dangerous goods transportation verification data according to the task occurrence time sequence, identify the time points when the transportation location, cargo handover status and vehicle execution status change, use the adjacent time points of status change as the start and end boundaries of the task stage, divide the waybill execution information into the task stages of transportation preparation, loading and departure, in-transit transportation, arrival and unloading and delivery completion, and convert each task stage into task nodes to generate a dangerous goods waybill task node sequence. Extract the effective time interval and applicable spatial region of each task node to form the spatiotemporal allowable range of each task node, and determine the transfer direction and transfer conditions based on the state acceptance requirements between adjacent task nodes to form transfer constraints between nodes. Based on the sequence of task nodes in the dangerous goods shipment, establish connection edges between task nodes, and configure the spatiotemporal allowable range of each task node and the transfer constraints between nodes to generate a dangerous goods shipment task constraint graph.

[0008] As a preferred embodiment of the dangerous goods cargo manifest verification method based on spatiotemporal trajectory matching described in this invention, the steps for obtaining the road matching trajectory sequence are as follows: Based on the task constraint graph of dangerous goods transport orders, vehicle location records bound to dangerous goods transport orders to be verified are extracted from the dangerous goods transport verification data and arranged in order of location time to generate a vehicle location sequence. Candidate roads are retrieved from the road network centered on each location point in the vehicle location sequence, and the spatial matching degree between the location point and the candidate road is calculated. At the same time, the road transfer matching degree is calculated by combining the topological connection relationship between adjacent location points and corresponding candidate roads. Based on spatial matching degree and road transfer matching degree, the comprehensive matching degree of each candidate road is calculated. Candidate roads with a comprehensive matching degree not lower than the preset comprehensive matching threshold are selected. The candidate road with the highest comprehensive matching degree among the selected candidate roads is determined as the road affiliation of the corresponding positioning point. The road positions of each positioning point are sequentially connected according to the time sequence of the vehicle positioning sequence to generate a road matching trajectory sequence.

[0009] As a preferred embodiment of the dangerous goods shipment document verification method based on spatiotemporal trajectory matching described in this invention, the steps for generating real-time trajectory verification information are as follows: Based on the road matching trajectory sequence, the time interval between adjacent trajectory points and the road network travel distance are calculated to determine the travel speed between trajectory points. Combined with the road connection relationship, trajectory points that meet the spatiotemporal continuity condition are marked as continuous trajectory points, and trajectory points that do not meet the spatiotemporal continuity condition are marked as abnormal trajectory points. According to the positioning time sequence, continuous trajectory points that belong to the road and satisfy the road reachability relationship are sequentially aggregated into a reliable trajectory segment, and the positioning interruption and road jump that occur between the reliable trajectory segments are used as the interval boundary to divide the trajectory missing interval. The spatiotemporal coverage of the reliable trajectory segment is matched with the spatiotemporal allowable range of the task nodes in the dangerous goods shipment task constraint diagram. The execution order between task nodes is matched based on the transfer constraints between nodes. Task nodes that meet both the spatiotemporal conditions and the transfer conditions are identified as completed task nodes, and real-time trajectory verification information is generated.

[0010] As a preferred embodiment of the dangerous goods shipment document verification method based on spatiotemporal trajectory matching described in this invention, the steps for obtaining possible trajectories are as follows: Extract the adjacent reliable trajectory segments on both sides of the missing interval of each trajectory from the real-time trajectory verification information, set the end road position and end time of the previous reliable trajectory segment as the front boundary state, set the start road position and start time of the next reliable trajectory segment as the back boundary state, and calculate the missing duration based on the time difference between the front and back boundary states. Starting from the previous boundary state, the subsequent road states are expanded layer by layer according to the connection direction of the road network and the spatiotemporal constraints of the road. The reachable time range of each subsequent road state is calculated, and the road states that can be reached within the missing duration are selected to generate a forward reachable road set. Using the subsequent boundary state as the endpoint, the preceding road states are deduced layer by layer according to the reverse connection relationship of the road network. The travel time required for each preceding road state to connect to the subsequent boundary state is calculated, and the road states that can connect to the subsequent boundary state within the missing duration are selected to generate a set of backward reachable roads. Cross-matching is performed on the forward reachable road set and the backward reachable road set to extract the bidirectional reachable road states that exist in both the forward reachable road set and the backward reachable road set, and candidate paths are formed by combining the road states according to the connection relationship between them. Calculate the estimated travel time for each candidate path and verify the connection relationship between the first and last roads of the candidate path and the state of the front and rear boundaries. Filter out unreachable paths that do not meet the requirements of time continuity and road accessibility. At the same time, according to the positioning time order, connect the candidate paths retained in the missing interval of each trajectory with the corresponding reliable trajectory segments in sequence to form possible trajectories.

[0011] As a preferred embodiment of the dangerous goods shipment verification method based on spatiotemporal trajectory matching described in this invention, the steps for constructing the initial executable domain for the remaining tasks are as follows: Extract the vehicle's spatiotemporal state at the latest positioning time from each possible trajectory and match it with the task nodes in the hazardous goods transport order task constraint diagram. When the vehicle's spatiotemporal state is within the spatiotemporal allowable range of a certain task node and all preceding task nodes have been completed, the task node is determined as the currently executing task node. When the vehicle's spatiotemporal state has left the spatiotemporal allowable range of the currently executing task node and satisfies the transfer constraint that the currently executing task node points to the next task node, the currently executing task node is determined as the completed task node and the next task node is determined as the task node to be executed. Determine the task progress based on the task order position of the currently executing task node and the task node to be executed in the hazardous goods transport order task constraint diagram, and bind the state according to the possible trajectory identifier to generate the current state branch of the vehicle. Map the current state branches of each vehicle to the task constraint graph of the dangerous goods transport order, and extract the unfinished task nodes from the task nodes where the corresponding task progress is located to generate the remaining task subgraph. The hierarchical spatiotemporal expansion is performed according to the task execution order in the remaining task subgraph, and the reachable state range of each task node is calculated based on the spatiotemporal allowable range of the task nodes and the transition constraints between nodes. The initial executable domain of the remaining tasks is then generated by summarizing these results.

[0012] As a preferred embodiment of the dangerous goods shipment document verification method based on spatiotemporal trajectory matching described in this invention, the steps for obtaining the executable domain of the branch task are as follows: Starting from the current state branch of each vehicle, according to the task execution level in the remaining task subgraph, perform layer-by-layer reachability propagation in the initial executable domain of the remaining tasks, calculate the transition time required from the current execution state to the next task node and determine the reachable path, and at the same time eliminate execution states that do not meet the spatiotemporal allowable range of task nodes and the transition constraints between nodes to obtain valid execution states; Based on the remaining task subgraph, the effective execution states of adjacent levels are continuously reachable, and the effective execution states that can be continuously propagated from the current state of the vehicle to the termination task node are retained to obtain the branch task executable domain of each possible trajectory.

[0013] As a preferred embodiment of the dangerous goods shipment manifest verification method based on spatiotemporal trajectory matching described in this invention, the steps for determining the verification status of the dangerous goods shipment manifest are as follows: Based on the possible trajectory identifiers, each possible trajectory is bound to the corresponding branch task executable domain and mapped to the dangerous goods freight bill task constraint graph. The task progress corresponding to the current state branch of the vehicle is used as the dividing boundary. The task nodes and connecting edges that have been completed before the task progress are determined as the already occurred task paths, and the unfinished task nodes, subsequent task nodes and connecting edges corresponding to the task progress are determined as the remaining task paths. Each possible trajectory is spatiotemporally matched with the corresponding completed task path. The performance status of the completed task path is determined based on the spatiotemporal allowable range of the task node and the transfer constraints between nodes of the possible trajectory. The continuous reachable execution chain from the current state of the vehicle to the termination task node is retrieved along the remaining task path to generate branch task chain performance information. A set-level comparison is performed on the fulfillment information of the branch task chains. When all branches meet the constraints of the tasks that have occurred and there are continuous reachable execution chains for the remaining task paths corresponding to each branch, the verification status of the dangerous goods shipment order is determined to be verified as verified. When there are no branches that simultaneously meet the constraints of the tasks that have occurred and there are continuous reachable execution chains for the corresponding remaining task paths, the verification status of the dangerous goods shipment order is determined to be verified as ...

[0014] In a second aspect, the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the dangerous goods shipment verification method based on spatiotemporal trajectory matching as described in the first aspect of the present invention.

[0015] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the dangerous goods shipment verification method based on spatiotemporal trajectory matching as described in the first aspect of the present invention.

[0016] The beneficial effects of this invention are as follows: by combining road spatiotemporal constraints to deduce candidate paths and filtering out unreachable paths based on the continuous reachability of the preceding and following boundary states, possible trajectories are recovered, improving the reliability of possible trajectory identification; by performing set-level dynamic verification on possible trajectories and the executable domains of corresponding branch tasks, the joint judgment of the performance status of already occurred tasks and the continuous executableness of remaining tasks is realized, improving the accuracy of the full-link verification of dangerous goods cargo manifests in scenarios with discontinuous trajectories. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of a dangerous goods shipment verification method based on spatiotemporal trajectory matching.

[0019] Figure 2 A flowchart for constructing a task constraint graph for hazardous goods transport orders and generating real-time trajectory verification information.

[0020] Figure 3 A comparison chart of F1 scores for possible trajectory identification under different trajectory missing durations.

[0021] Figure 4 Comparative data chart showing the improvement in accuracy of verification status determination under different possible numbers of trajectory branches and remaining task nodes. Detailed Implementation

[0022] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0023] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0024] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0025] Reference Figures 1-4 As one embodiment of the present invention, this embodiment provides a dangerous goods shipment document verification method based on spatiotemporal trajectory matching, including the following steps: S1. Perform task-based parsing on the dangerous goods transportation verification data, divide the waybill execution information into task nodes in sequence, and configure spatiotemporal and business constraints for each task node and the transfer between nodes to construct a dangerous goods waybill task constraint graph.

[0026] S1.1: Arrange the waybill execution information in the dangerous goods transportation verification data according to the task occurrence sequence, identify the time points when the transportation location, cargo handover status and vehicle execution status change, use the adjacent time points of status change as the start and end boundaries of the task stage, divide the waybill execution information into the task stages of transportation preparation, loading and departure, in-transit transportation, arrival and unloading and delivery completion, and convert each task stage into task nodes to generate a dangerous goods waybill task node sequence.

[0027] Specifically, the task occurrence time of the waybill execution information is extracted from the dangerous goods transportation verification data, and the waybill execution information is arranged from first to last according to the task occurrence time. The transportation location, cargo handover status (such as not handed over, loading handover in progress, loading handover completed, unloading handover in progress, unloading handover completed, and delivery confirmation completed) and vehicle execution status (such as standby, heading to loading location, loading operation, transportation en route, arriving at unloading location, unloading operation, and transportation task completed) in adjacent waybill execution information are compared in turn. When any status changes, the corresponding task occurrence time is determined as the status change time point. Using the time points of adjacent state changes as the start and end boundaries of the task stages, the process of the vehicle completing transportation preparation but not yet loading dangerous goods is divided into the transportation preparation task stage; the process from the start of loading dangerous goods to the vehicle leaving the loading point is divided into the loading and departure task stage; the process from the vehicle leaving the loading point to arriving at the unloading point is divided into the in-transit transportation task stage; the process from the vehicle arriving at the unloading point to the completion of unloading dangerous goods is divided into the arrival and unloading task stage; and the process from the completion of unloading dangerous goods to the completion of delivery confirmation is divided into the delivery completion task stage. Each task stage is converted into task nodes according to the sequence of task occurrence, generating a dangerous goods shipment task node sequence.

[0028] S1.2: Extract the effective time interval and applicable spatial region of each task node to form the spatiotemporal allowable range of each task node, and determine the transfer direction and transfer conditions based on the state acceptance requirements between adjacent task nodes to form transfer constraints between nodes.

[0029] Specifically, the start and end times of the plan are extracted from the waybill execution information associated with each task node, and the time span between the two is calculated to obtain the effective time interval; the operation location and the permitted transportation range corresponding to each task node are extracted, the area covered by the operation location is connected with the permitted transportation section to form the applicable spatial area, and the effective time interval and the applicable spatial area are associated according to the task node identifier to form the spatiotemporal allowable range of each task node. According to the sequence of task nodes in the dangerous goods shipment, the direction of transfer is set from the previous task node to the next task node. When the previous task node reaches the completed state, the next task node reaches the started state, the end time of the previous task node is not later than the start time of the next task node, and the end position of the previous task node can reach the start position of the next task node within the time interval between the two task nodes, the previous task node is allowed to transfer to the next task node. The corresponding business status acceptance conditions, time connection conditions, and spatial reachability conditions are configured between the previous task node and the next task node to form inter-node transfer constraints.

[0030] S1.3: Establish connection edges between task nodes based on the sequence of task nodes in the dangerous goods shipment, and associate and configure the spatiotemporal allowable range of each task node and the transfer constraints between nodes to generate a dangerous goods shipment task constraint graph.

[0031] Specifically, according to the arrangement order in the dangerous goods shipment task node sequence, a unique task node identifier is configured for each task node, and a connection edge is established between adjacent task nodes according to the transfer direction from the previous task node to the next task node. The spatiotemporal allowable range of each task node is written into the corresponding task node, and the transfer constraints between nodes are written into the corresponding connection edge. All task nodes are connected according to the execution order of the dangerous goods shipment task node sequence to generate a dangerous goods shipment task constraint graph.

[0032] S2. Based on the task constraint graph of dangerous goods freight bill, perform road matching on vehicle positioning records to obtain road matching trajectory sequences. Divide reliable trajectory segments and trajectory missing intervals according to the spatiotemporal continuity of adjacent trajectory points. At the same time, determine the completed task nodes according to the spatiotemporal correspondence between reliable trajectory segments and task nodes, and generate real-time trajectory verification information.

[0033] S2.1: Based on the dangerous goods transport order task constraint graph, extract the vehicle location records bound to the dangerous goods transport order to be verified from the dangerous goods transport verification data, and arrange them in the order of location time to generate a vehicle location sequence.

[0034] Specifically, the vehicle identifiers associated with the dangerous goods freight bills to be verified and the execution time range of the bills are read from the dangerous goods freight bill task constraint diagram. Using the vehicle identifier as the search condition, vehicle location records with the same vehicle identifier and a location time within the execution time range of the bill are selected from the dangerous goods transportation verification data. The location time and location in the vehicle location records are extracted, and vehicle location records with missing location time, invalid location, and duplicate location time are removed. The retained vehicle location records are arranged from earliest to latest according to the location time to generate a vehicle location sequence.

[0035] S2.2: Retrieve candidate roads from the road network centered on each location point in the vehicle location sequence, calculate the spatial matching degree between the location point and the candidate road, and calculate the road transfer matching degree by combining the topological connection relationship between adjacent location points and corresponding candidate roads.

[0036] Specifically, for each location point in the vehicle location sequence, with the location of the location point as the center and a specific search radius (e.g., 50 meters) as the range, roads in the road network that are less than the center distance from the search radius are retrieved as candidate roads corresponding to the location point. For each candidate road corresponding to a location point, the location point is vertically projected onto the candidate road. When the vertically projected point is within the road segment of the candidate road, the vertically projected distance is taken as the spatial distance between the location point and the candidate road. When the vertically projected point is not within the road segment of the candidate road, the straight-line distance between the two endpoints of the location point and the candidate road is calculated separately, and the smaller value is taken as the spatial distance between the location point and the candidate road. The expression for calculating spatial matching degree is:

[0037] ; in, Indicates the first position in the vehicle location sequence The positioning point and the first Spatial matching degree between candidate roads; Indicates the first The positioning point and the first Spatial distance between candidate roads; Indicates the search radius; Indicates the location point number in the vehicle's location sequence; Indicates the first The candidate road number corresponding to each location point; According to the order of adjacent positioning points in the vehicle positioning sequence, the candidate roads corresponding to the previous positioning point and the candidate roads corresponding to the next positioning point are combined. Directed connection paths between candidate roads are searched along the road network. When no directed connection path exists between candidate roads, the road transfer matching degree is set to the minimum value (e.g., 0). When directed connection paths exist between candidate roads, the road transfer matching degree is calculated based on the matching degree between the road network travel distance and the linear displacement between adjacent positioning points, and the matching degree between the estimated travel time and the time interval between adjacent positioning points. The expression is:

[0038] ; in, Indicates the first The corresponding location point is the th Candidate roads and the first +1 positioning point corresponding to the first Road transfer matching degree between candidate roads; Weights representing the degree of consistency between distances; Weights representing time consistency; Indicates the first The positioning point is at the _ ... Projected positions on candidate roads up to the first +1 positioning point at the The road network travel distance between the projected locations on the candidate roads; Indicates the maximum distance deviation; This represents the linear displacement between adjacent positioning points; Indicates the maximum time deviation; Indicates the path along the directed connection from the first... The candidate roads were driven to the first The estimated travel time required for each candidate road; Indicates the time interval between adjacent positioning points; It should be noted that the weight of distance consistency... Weight of consistency with time Calibration was performed using historically confirmed road matching samples, and changes were made. The value of is used to compare the consistency between the road matching result and the actual road assignment, and the value corresponding to the highest consistency is selected. As a weight for distance consistency, As a weight for time consistency; The maximum distance deviation is set according to the upper limit of the deviation between the road network travel distance and the linear displacement between adjacent positioning points in the normal road matching sample, and the maximum time deviation is set according to the upper limit of the deviation between the expected travel time and the time interval between adjacent positioning points in the normal road matching sample.

[0039] S2.3: Based on the spatial matching degree and road transfer matching degree, calculate the comprehensive matching degree of each candidate road, select candidate roads with a comprehensive matching degree not lower than the preset comprehensive matching threshold, and determine the candidate road with the highest comprehensive matching degree among the selected candidate roads as the road affiliation of the corresponding positioning point. Connect the road positions of each positioning point in time sequence according to the time order of the vehicle positioning sequence to generate a road matching trajectory sequence.

[0040] Specifically, for each positioning point in the vehicle positioning sequence except the first positioning point, each candidate road of the positioning point is combined with each candidate road retained from the previous positioning point, and the maximum value of the corresponding road transfer matching degree is determined as the road transfer matching degree of the candidate road; the spatial matching degree and the road transfer matching degree of the candidate road are weighted and summed to obtain the comprehensive matching degree of the corresponding candidate road, wherein the comprehensive matching degree of the candidate road corresponding to the first positioning point in the vehicle positioning sequence adopts the spatial matching degree; The comprehensive matching degree of each candidate road is compared with the comprehensive matching threshold one by one. Candidate roads with a comprehensive matching degree lower than the comprehensive matching threshold are eliminated. The remaining candidate roads are arranged from high to low according to the comprehensive matching degree. The candidate road with the highest score is selected as the road affiliation of the corresponding positioning point. Each positioning point is mapped to the road position of the corresponding road affiliation. The road positions of adjacent positioning points are connected in order from early to late according to the positioning time to generate a road matching trajectory sequence.

[0041] It should be noted that the comprehensive road matching threshold is set by statistically analyzing the comprehensive matching degree corresponding to the correct road attribution in the normal road matching samples, and taking the lowest percentile (such as the 5th percentile) after arranging the comprehensive matching degree from low to high. An exemplary value range is 0.6 to 0.8. The weights of spatial matching degree and road transfer matching degree are calibrated using historically confirmed road matching samples. The values ​​of the two are changed respectively, and the consistency between the road matching results and the actual road affiliation is compared. The weight combination corresponding to the highest consistency is determined as the weight for calculating the comprehensive matching degree.

[0042] S2.4: Based on the road matching trajectory sequence, calculate the time interval between adjacent trajectory points and the road network travel distance, determine the travel speed between trajectory points, and combine the road connection relationship to mark trajectory points that meet the spatiotemporal continuity conditions as continuous trajectory points, and mark trajectory points that do not meet the spatiotemporal continuity conditions as abnormal trajectory points.

[0043] Specifically, the positioning time and road position of adjacent trajectory points are extracted, the time difference between the positioning times of the preceding and following trajectory points is calculated as the time interval between adjacent trajectory points, and the connecting path between the road position of the preceding trajectory point and the road position of the following trajectory point is retrieved along the road network. The road lengths included in the connecting path are summarized as the road network travel distance between adjacent trajectory points, and the ratio of the road network travel distance to the time interval is calculated to obtain the travel speed between trajectory points. When the time interval between adjacent trajectory points does not exceed the positioning continuity threshold, there is a connecting path between the corresponding roads, and the driving speed is within the allowable speed range corresponding to the connecting path, the adjacent trajectory points are marked as continuous trajectory points. When the time interval between adjacent trajectory points exceeds the positioning continuity threshold, there is no connecting path between the corresponding roads, or the driving speed exceeds the allowable speed range corresponding to the connecting path, the adjacent trajectory points are marked as abnormal trajectory points. It should be noted that the location continuity threshold is set according to the normal reporting cycle of the vehicle positioning device and the number of consecutive missing positioning records allowed, so as to cover short-term communication delays but be able to identify substantial positioning interruptions. An exemplary value range is 60 seconds to 180 seconds.

[0044] S2.5: According to the positioning time sequence, the continuous trajectory points that belong to the road and satisfy the road reachability relationship are sequentially aggregated into a reliable trajectory segment, and the positioning interruption and road jump that occur between reliable trajectory segments are used as the interval boundary to divide the trajectory missing interval.

[0045] Specifically, when adjacent consecutive trajectory points are located on the same road or the road where the previous consecutive trajectory point is located can be reached along the road network, the adjacent consecutive trajectory points are grouped into the same trajectory point set, and the road locations in the same trajectory point set are connected according to the positioning time order to form a reliable trajectory segment. When the time interval between adjacent trajectory points exceeds the positioning continuity threshold, the corresponding time period between adjacent reliable trajectory segments is identified as a positioning interruption. When there is no reachable path between the roads where adjacent trajectory points are located, the change in road position between adjacent reliable trajectory segments is identified as a road jump. The starting and ending positions and times corresponding to the positioning interruption and road jump are used as interval boundaries to divide the trajectory missing interval.

[0046] S2.6: Match the spatiotemporal coverage of the trusted trajectory segment with the spatiotemporal allowable range of the task nodes in the hazardous goods shipment task constraint diagram, and verify the execution order between task nodes based on the inter-node transfer constraint verification. Determine the task nodes that meet both the spatiotemporal conditions and the transfer conditions as completed task nodes, and generate real-time trajectory verification information.

[0047] Specifically, the start and end times of each reliable trajectory segment and the road location coverage are extracted. The location times of the reliable trajectory segments are compared with the effective time intervals of each task node in the dangerous goods shipment task constraint diagram. The road location coverage of the reliable trajectory segments is compared with the applicable spatial area of ​​each task node. When a reliable trajectory segment enters and passes through the applicable spatial area of ​​a task node within its effective time interval, a matching relationship is established between the reliable trajectory segment and the task node. The matching task nodes are arranged according to the positioning time sequence of the reliable trajectory segment, and compared one by one with the task node sequence in the dangerous goods cargo order task node sequence. When the execution order of adjacent matching task nodes is consistent with the dangerous goods cargo order task node sequence, and the completion status, end time and end position of the previous task node satisfy the node transfer constraint between the previous task node and the next task node, the transfer between adjacent matching task nodes is determined to be valid. Task nodes whose spatiotemporal allowable range is covered by a reliable trajectory segment and whose preceding node transfers are all valid are identified as completed task nodes. Reliable trajectory segments, missing trajectory intervals, and completed task nodes are associated to generate real-time trajectory verification information.

[0048] S3. Locate the missing trajectory intervals in the real-time trajectory verification information, extract the front and rear boundary states, and deduce candidate paths in combination with road spatiotemporal constraints. At the same time, filter out inaccessible paths based on the continuous reachability of the front and rear boundary states to obtain possible trajectories.

[0049] S3.1: Extract the adjacent reliable trajectory segments on both sides of the missing interval of each trajectory from the real-time trajectory verification information, set the end road position and end time of the previous reliable trajectory segment as the front boundary state, set the start road position and start time of the next reliable trajectory segment as the back boundary state, and calculate the missing duration based on the time difference between the front and back boundary states.

[0050] Specifically, the reliable trajectory segment whose end time is closest to the start time of the trajectory missing interval and the reliable trajectory segment whose start time is closest to the end time of the trajectory missing interval are retrieved respectively. The reliable trajectory segment located before the trajectory missing interval is taken as the previous reliable trajectory segment, and the reliable trajectory segment located after the trajectory missing interval is taken as the next reliable trajectory segment. Extract the road location and location time corresponding to the trajectory point with the latest location time in the previous reliable trajectory segment, and set it as the front boundary state; extract the road location and location time corresponding to the trajectory point with the earliest location time in the next reliable trajectory segment, and set it as the back boundary state. The time difference between the positioning time corresponding to the rear boundary state and the positioning time corresponding to the front boundary state is calculated to obtain the missing duration.

[0051] S3.2: Starting from the previous boundary state, expand the subsequent road states layer by layer according to the connection direction of the road network and the spatiotemporal constraints of the road. Calculate the reachable time range of each subsequent road state, and filter the road states that can be reached within the missing duration to generate a forward reachable road set.

[0052] Specifically, starting from the end road position and end time in the previous boundary state, the subsequent roads directly connected to the current road are searched layer by layer along the allowed driving direction of the road network. For each subsequent road, the cumulative shortest travel time and the cumulative longest travel time are calculated based on the cumulative road length from the previous boundary state to the subsequent road and the corresponding road's allowed speed range. These are then correlated with the end time in the previous boundary state to obtain the reachable time range of the subsequent road state. Subsequent road states whose cumulative shortest travel time does not exceed the missing duration and whose reachable time range overlaps with the permitted travel period are retained and expanded layer by layer. Subsequent road states whose cumulative shortest travel time exceeds the missing duration or whose reachable time range does not overlap with the permitted travel period are removed. The retained road states are summarized to generate a forward reachable road set.

[0053] It should be noted that the spatiotemporal constraints of roads include the permitted driving direction, road length, permitted speed range, and permitted time periods for vehicles transporting dangerous goods.

[0054] S3.3: Using the subsequent boundary state as the endpoint, the preceding road states are deduced layer by layer according to the reverse connection relationship of the road network. The travel time required for each preceding road state to connect to the subsequent boundary state is calculated, and the road states that can connect to the subsequent boundary state within the missing duration are selected to generate the backward reachable road set.

[0055] Specifically, the starting road position and starting time in the subsequent boundary state are used as the endpoint of the reverse deduction. The preceding roads that can enter the current road along the road's allowed driving direction are searched layer by layer according to the reverse connection relationship of the road network. The ratio of the road length of each preceding road to the upper limit of the road's allowed speed range is used as the shortest travel time for a single segment, and the ratio of the road length of each preceding road to the lower limit of the road's allowed speed range is used as the longest travel time for a single segment. The shortest travel time and the longest travel time for each single segment are summarized according to the road sequence from the preceding road to the subsequent boundary state. The cumulative shortest travel time and the cumulative longest travel time required for each preceding road state to connect to the subsequent boundary state are obtained respectively. By tracing back the start time of each backward boundary state to accumulate the shortest and longest travel times, the possible departure time ranges for each preceding road state are obtained. The preceding road states whose cumulative shortest travel time does not exceed the missing duration and whose possible departure time range overlaps with the allowed travel period are retained and the process continues to evolve layer by layer. The preceding road states whose cumulative shortest travel time exceeds the missing duration or whose possible departure time range does not overlap with the allowed travel period are removed. The retained road states are then summarized to generate a backward reachable road set.

[0056] S3.4: Perform cross-matching on the forward reachable road set and the backward reachable road set, extract the bidirectional reachable road states that exist in both the forward reachable road set and the backward reachable road set, and combine them according to the connection relationship between the road states to form candidate paths.

[0057] Specifically, the road status of the forward-reachable road set and the backward-reachable road set are matched one by one according to road signs and road locations. The reachable time range and departure time range corresponding to the same road location are compared. Road status with the same road location and overlapping time range are extracted as bidirectional reachable road status. Using each bidirectional accessible road state as a connection point, the forward road sequence from the front boundary state to the bidirectional accessible road state and the backward road sequence from the bidirectional accessible road state to the rear boundary state are extracted respectively. The forward road sequence and the backward road sequence are then connected according to the road's permitted driving direction to form candidate paths.

[0058] S3.5: Calculate the estimated travel time for each candidate path, and verify the connection relationship between the first and last roads of the candidate path and the state of the front and rear boundaries. Filter out unreachable paths that do not meet the requirements of time continuity and road accessibility. At the same time, according to the positioning time order, connect the candidate paths retained in the missing interval of each trajectory with the corresponding reliable trajectory segments in sequence to form possible trajectories.

[0059] Specifically, the travel time of each road is calculated based on the travel length and permissible speed of each road in the candidate path, and the travel time of each road is summarized into the expected travel time of the candidate path. The sum of the end time in the front boundary state and the expected travel time is taken as the expected arrival time. Simultaneously, according to the road connection order in the candidate path, the shortest and longest travel times for each road are calculated based on the travel length and permissible speed range of each road. These times are then accumulated segment by segment, starting from the end time in the previous boundary state, to obtain the expected arrival time range for each road state in the candidate path. The road identifier, road location, and expected arrival time range are then associated according to the road connection order to form a temporal road state sequence corresponding to the candidate path. If the first road of a candidate path is the same as the last road in the previous boundary state or directly connected along the allowed driving direction of the road, and the last road of the candidate path is the same as the first road in the subsequent boundary state or directly connected along the allowed driving direction of the road, and the expected arrival time is not later than the start time in the subsequent boundary state, the candidate path is retained; otherwise, the candidate path is eliminated. According to the positioning time sequence, the time-series road state sequences corresponding to the candidate paths retained in each trajectory missing interval are connected between the corresponding reliable trajectory segments, and the candidate paths retained in different trajectory missing intervals are combined to form possible trajectories containing road location and corresponding time information. When the number of possible trajectories is zero, the verification status can be directly determined as verification abnormal.

[0060] Figure 3 This paper presents the overall trend of the F1 score for potential trajectory recognition (used to comprehensively measure the balance between precision and recall in potential trajectory recognition; a higher value indicates better potential trajectory recognition performance) as the trajectory loss duration changes under different trajectory recovery methods. The upper part is an overview curve covering the entire trajectory loss duration, used to present the characteristics and differences in the F1 score of potential trajectory recognition under three conditions: one-way reachability deduction method, two-way reachability deduction method, and the method of this invention, as the trajectory loss duration increases. The one-way reachability deduction method is a method that deduces reachable paths based only on the boundary state of one side of the trajectory loss interval, while the two-way reachability deduction method is based on the boundary state of the previous boundary. The method described above generates reachable road states from the initial and subsequent boundary states and performs cross-matching. This invention, based on bidirectional reachability deduction, calculates the estimated travel time of candidate paths, verifies the connection relationship between the first and last roads of the candidate paths and the preceding and following boundary states, and filters out unreachable paths based on temporal continuity, road reachability, and road spatiotemporal constraints. The lower part of the invention provides a magnified view of the duration intervals of missing trajectories and marks the locations of the greatest local differences, reflecting that under specific trajectory missing duration conditions, this invention's method has the most significant comprehensive advantage over unidirectional and bidirectional reachability deduction methods in retaining actual feasible trajectories and eliminating unreachable candidate paths. Therefore... Figure 3 This demonstrates that by combining the continuous accessibility of the preceding and following boundary states with the spatiotemporal constraints of the road to screen candidate paths, the accuracy and stability of potential trajectory identification can be improved even when trajectory positioning records are interrupted, providing a more reliable trajectory basis for subsequent determination of the current vehicle status and verification of dangerous goods cargo manifests.

[0061] S4. Extract the current state of the vehicle for each possible trajectory, map it to the task constraint graph of the dangerous goods transport order, construct the initial executable domain of the remaining tasks, and perform layer-by-layer reachability propagation and constraint contraction on the initial executable domain of the remaining tasks starting from the current state of each vehicle to obtain the executable domain of the branch tasks.

[0062] S4.1: Extract the vehicle's spatiotemporal state at the latest positioning time from each possible trajectory and match it with the task nodes in the hazardous goods transport order task constraint diagram. When the vehicle's spatiotemporal state is within the spatiotemporal allowable range of a certain task node and all preceding task nodes have been completed, the task node is determined as the currently executing task node. When the vehicle's spatiotemporal state has left the spatiotemporal allowable range of the currently executing task node and satisfies the transfer constraint from the currently executing task node to the next task node, the currently executing task node is determined as the completed task node and the next task node is determined as the task node to be executed. The task progress is determined according to the task order position of the currently executing task node and the task node to be executed in the hazardous goods transport order task constraint diagram, and the state is bound according to the possible trajectory identifier to generate the current state branch of the vehicle.

[0063] Specifically, for each possible trajectory, the spatiotemporal coverage of the possible trajectory is matched with the spatiotemporal allowable range of each task node in the dangerous goods shipment task constraint diagram according to the order of the road locations and corresponding time information in the possible trajectory. The execution order between the matching task nodes is verified according to the transfer constraints between nodes. The task node that satisfies both the spatiotemporal conditions and the transfer conditions is determined as the branch completed task node corresponding to the possible trajectory. For each possible trajectory, the road location and positioning time corresponding to the trajectory point with the latest positioning time are extracted to form the vehicle spatiotemporal state at the latest positioning time. The vehicle spatiotemporal state at the latest positioning time is then matched one by one with the spatiotemporal allowable range of each task node in the hazardous goods freight bill task constraint diagram: when the positioning time is within the effective time interval of the task node and the road location is within the applicable spatial area of ​​the task node, and all task nodes before the task node belong to the completed task nodes of the branch corresponding to the possible trajectory, the task node that meets the conditions and is the last in the hazardous goods freight bill task node sequence is taken as the current execution task node; when the vehicle spatiotemporal state at the latest positioning time is no longer within the spatiotemporal allowable range of the current execution task node, and the current execution task node has reached the completed state and meets the business state inheritance conditions, time connection conditions, and spatial reachability conditions for the current execution task node to point to the next task node, the current execution task node is updated to a completed task node, and the next task node is taken as a task node to be executed. When the vehicle is still within the spatiotemporal allowable range of the current task node, the task progress is represented by the task sequence position of the current task node in the hazardous goods cargo manifest task constraint graph. When the vehicle has completed the current task node and satisfies the inter-node transfer constraint for transferring to the next task node, the task progress is represented by the task sequence position of the task node to be executed in the hazardous goods cargo manifest task constraint graph. The vehicle's spatiotemporal state, the current task node, the task node to be executed, and the task progress at the latest positioning time are bound to the possible trajectory identifiers to generate the vehicle's current state branch.

[0064] S4.2: Map the current state branches of each vehicle to the task constraint graph of the hazardous goods transport order, and extract the unfinished task nodes from the task nodes where the corresponding task progress is located, and generate the remaining task subgraph.

[0065] Specifically, the task progress in the current state branch of each vehicle is read, and the task node corresponding to the task progress is located in the dangerous goods cargo order task constraint graph according to the task node identifier. Taking the task node where the task progress is located as the starting point, the task is traversed backward along the connecting edges in the dangerous goods cargo order task constraint graph according to the task execution order. The unfinished task content in the task node where the task progress is located and all unfinished task nodes after the task node where the task progress is located are retained. At the same time, the connecting edges between unfinished task nodes, the spatiotemporal allowable range of each unfinished task node and the corresponding node transition constraints are retained. The completed task nodes and their corresponding connecting edges before the task node where the task progress is located are removed, and the remaining task subgraph is generated.

[0066] S4.3: Perform hierarchical spatiotemporal expansion according to the task execution order in the remaining task subgraph, and calculate the reachable state range of each task node based on the spatiotemporal allowable range of the task nodes and the transition constraints between nodes, and summarize to generate the initial executable domain of the remaining tasks.

[0067] Specifically, according to the task execution order in the remaining task subgraph, the task node with the current task progress is taken as the first layer, and the subsequent unfinished task nodes are sequentially assigned to the corresponding layers. Taking the vehicle's spatiotemporal state in the current state branch as the starting point, the time range and reachable road location for reaching the first-layer task node are calculated along the connecting edges in the remaining task subgraph. The intersection with the spatiotemporal allowable range of the first-layer task node is then used for filtering. The arrival time range and reachable road location that satisfy the inter-node transition constraints and have a non-empty intersection are taken as the reachable state range of the first-layer task node. The reachable state range of the task node in the previous layer is used as the propagation starting point of the task node in the next layer. The travel time and reachable road location between adjacent task nodes are calculated layer by layer. The states that cannot be connected are filtered out according to the spatiotemporal allowable range of the task node in the next layer and the transition constraints between nodes. The reachable state range of each task node is obtained and the initial executable domain of the remaining tasks is generated.

[0068] It should be noted that the initial executable domain of the remaining tasks unifies the spatiotemporal allowable range and node transfer constraints of the current state branches of each vehicle and the remaining task subgraph, thus forming the reachable state boundary of each unfinished task node in advance. This allows the verification of dangerous goods cargo manifests to no longer be limited to static compliance judgment of the already occurred trajectory, but to further judge whether the remaining tasks still have the conditions for continuous execution under different possible trajectories, thereby improving the accuracy of the judgment of the continuous performance capability of dangerous goods cargo manifests in the case of missing trajectories.

[0069] S4.4: Starting from the current state branch of each vehicle, according to the task execution level in the remaining task subgraph, perform layer-by-layer reachability propagation in the initial executable domain of the remaining tasks, calculate the transition time required from the current execution state to the next task node and determine the reachable path, and at the same time eliminate execution states that do not meet the spatiotemporal allowable range of task nodes and the transition constraints between nodes to obtain valid execution states.

[0070] Specifically, the system reads the vehicle's spatiotemporal state and task progress from each vehicle's current state branch, takes the vehicle's spatiotemporal state as the first-level current execution state, and propagates it layer by layer to the next task node according to the task execution hierarchy in the remaining task subgraph. Starting from the road position corresponding to the current execution state, the system searches adjacent roads layer by layer along the allowed driving direction of the road network until the searched road enters the applicable space area of ​​the next task node. The search sequence of roads that are connected continuously is taken as the reachable path, and the transfer time is calculated based on the road length and the allowable speed range of the road. The sum of the arrival time and transfer time of the current execution state is used as the expected arrival time of the next task node. Execution states whose expected arrival time is within the effective time interval of the next task node, whose reachable paths can enter the applicable space area of ​​the next task node, and whose inter-node transfer constraints are satisfied are retained. Execution states that do not meet the corresponding conditions are eliminated to obtain the effective execution states.

[0071] S4.5: Based on the remaining task subgraph, perform continuous reachability filtering on the effective execution states of adjacent levels, retain the effective execution states that can be continuously propagated from the current state of the vehicle to the termination task node, and obtain the branch task executable domain of each possible trajectory.

[0072] Specifically, according to the task execution order in the remaining task subgraph, the effective execution states corresponding to adjacent task nodes are compared layer by layer. When the effective execution state of the previous layer can reach the road position of the effective execution state of the next layer along the corresponding reachable path, the expected arrival time is within the effective time interval of the next task node, and the task state satisfies the transition constraints between nodes, a continuous reachable relationship is established between the two effective execution states. Starting from the valid execution state corresponding to the terminated task node, the system filters layer by layer in reverse order of task execution, retaining valid execution states that can be continuously connected to the next level of valid execution states, and eliminating valid execution states that cannot be connected to the next level of valid execution states, until the current state of the vehicle is traced back. The system also summarizes the road location range, arrival time range, task status range, and reachable paths between adjacent task nodes for each task node according to the possible trajectory identifier, and obtains the branch task executable domain of each possible trajectory.

[0073] S5. Based on the dangerous goods shipment task constraint diagram, perform set-level dynamic verification of possible trajectories and corresponding branch task executable domains, compare the satisfaction of each branch with the constraints of the already occurred tasks and the remaining tasks, and determine the verification status of the dangerous goods shipment.

[0074] S5.1: Based on the possible trajectory identifier, bind each possible trajectory to the corresponding branch task executable domain and map it to the dangerous goods freight bill task constraint graph. Use the task progress corresponding to the current state branch of the vehicle as the dividing boundary. The task nodes and connecting edges that have been completed before the task progress are determined as the already occurred task path, and the unfinished task nodes, subsequent task nodes and connecting edges corresponding to the task progress are determined as the remaining task path.

[0075] Specifically, possible trajectories with the same possible trajectory identifier, vehicle current state branches and branch task executable domains are bound together, and task nodes in the hazardous goods cargo order task constraint diagram are located according to the task progress in the vehicle current state branch. The task nodes corresponding to the task progress are used as the dividing boundary. Following the connecting edges in the task constraint graph of the dangerous goods shipment, we backtrack forward according to the task execution order and connect the task nodes that have been completed before the task progress in sequence to determine the task path that has occurred. We retain the task nodes that have not been completed according to the task progress and traverse backward along the connecting edges according to the task execution order to the termination task node. We connect the task nodes that have not been completed, the subsequent task nodes, and the connecting edges that connect the task nodes in sequence to determine the remaining task path.

[0076] S5.2: Perform spatiotemporal matching of each possible trajectory with the corresponding completed task path, determine the performance status of the completed task path based on the spatiotemporal allowable range of the task node and the transfer constraints between nodes of the possible trajectory, and retrieve the continuous reachable execution chain from the current state of the vehicle to the termination task node along the remaining task path to generate branch task chain performance information.

[0077] Specifically, the road location and positioning time of each possible trajectory are matched one by one with the applicable spatial area and effective time interval of each task node. When a possible trajectory enters and passes through the applicable spatial area within the effective time interval of a task node, and the execution order, time connection, and spatial accessibility of the corresponding trajectory segments of adjacent task nodes meet the transfer constraints between nodes, the task node is marked as fulfilled. When any task node fails to meet the spatiotemporal allowable range or the transfer between adjacent task nodes is not established, the task path that has occurred is marked as unfulfilled. Starting from the current state of the vehicle, the valid execution states in the executable domain of the branch tasks are searched layer by layer along the remaining task path. When each subsequent task node has a valid execution state that can be continuously connected with the previous task node and finally reaches the termination task node, a continuous reachable execution chain is formed. The fulfillment status of the task path that has occurred, the continuous reachable execution chain, and the possible trajectory identifiers are associated to generate branch task chain fulfillment information.

[0078] S5.3: Perform a set-level comparison of the branch task chain performance information. When all branches meet the constraints of the occurred task and the remaining task paths corresponding to each branch have a continuous reachable execution chain, the dangerous goods shipment verification status is determined to be verified as verified. When there are no branches that simultaneously meet the constraints of the occurred task and the corresponding remaining task paths have a continuous reachable execution chain, the dangerous goods shipment verification status is determined to be verified as verified as abnormal. When there are branches that simultaneously meet the constraints of the occurred task and the corresponding remaining task paths have a continuous reachable execution chain, but not all branches simultaneously meet the constraints of the occurred task and the corresponding remaining task paths have a continuous reachable execution chain, the dangerous goods shipment verification status is determined to be pending continuous verification.

[0079] Specifically, the task chain performance information of each branch is read one by one according to the possible trajectory identifier. The occurrence of the task path corresponding to each branch is checked to see if it meets the constraints of the occurrence of the task. The remaining task path corresponding to each branch is checked to see if there is a continuous reachable execution chain from the current state of the vehicle to the termination task node. The branch that meets the constraints of the occurrence of the task and has a continuous reachable execution chain in the remaining task path is marked as a complete performance branch. Otherwise, it is marked as an incomplete performance branch. A set-level statistical comparison is performed on complete and incomplete fulfillment branches. When all branches are complete fulfillment branches, it indicates that all possible trajectories can support the compliance of the already occurred tasks and the continuous execution of the remaining tasks, and the dangerous goods cargo manifest verification status is determined to be verification passed. When all branches are incomplete fulfillment branches, it indicates that there is no possible trajectory that can simultaneously satisfy the constraints of the already occurred tasks and the corresponding remaining task paths have a continuously reachable execution chain, and the dangerous goods cargo manifest verification status is determined to be verification abnormal. When complete and incomplete fulfillment branches exist simultaneously, it indicates that only some possible trajectories can support complete fulfillment, and the dangerous goods cargo manifest verification status is determined to be pending continuous verification.

[0080] It should be noted that the constraints of the tasks that have occurred include the spatiotemporal allowable range of each task node in the path of the tasks that have occurred, as well as the inter-node transfer constraints between adjacent task nodes.

[0081] Figure 4The horizontal axis represents the number of possible trajectory branches, and the vertical axis represents the number of remaining task nodes. Color changes demonstrate the improvement in accuracy of hazardous goods shipment verification status determination after adopting set-level dynamic verification, compared to the method of only verifying already occurred task paths. The method of only verifying already occurred task paths determines the verification status based on the satisfaction of constraints related to completed task nodes by possible trajectories, without further determining whether the remaining tasks can be executed continuously. In contrast, the method of this invention associates each possible trajectory with the executable domain of its corresponding branch tasks. Based on verifying the fulfillment of already occurred tasks, it further determines whether there is a continuously reachable execution chain for the remaining task paths and determines the hazardous goods shipment verification status through set-level comparison. The overall color distribution shows that as the number of remaining task nodes increases, judging solely based on already occurred tasks is more likely to miss potential anomalies where subsequent tasks cannot be executed continuously. The present invention, by simultaneously verifying the continuous executability of remaining tasks, significantly improves the accuracy of verification status determination, especially when the number of possible trajectory branches is small to medium and the remaining task chains are long. Therefore, Figure 4 This demonstrates that by jointly assessing the performance of completed tasks and the continued executability of remaining tasks, misjudgments can be reduced in scenarios with discontinuous trajectories that rely solely on historical performance, thereby improving the accuracy and reliability of the entire verification process for dangerous goods shipments.

[0082] This embodiment also provides a computer device applicable to the dangerous goods shipment verification method based on spatiotemporal trajectory matching, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the dangerous goods shipment verification method based on spatiotemporal trajectory matching as proposed in the above embodiment.

[0083] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0084] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the dangerous goods shipment verification method based on spatiotemporal trajectory matching as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0085] In summary, this invention achieves the recovery of possible trajectories and improves the reliability of possible trajectory identification by combining road spatiotemporal constraints to deduce candidate paths and filtering out unreachable paths based on the continuous reachability of the preceding and following boundary states; and by performing set-level dynamic verification of possible trajectories and the executable domains of corresponding branch tasks, it achieves the joint judgment of the fulfillment status of already occurred tasks and the continuous executableness of remaining tasks, thereby improving the accuracy of the full-link verification of dangerous goods cargo manifests in scenarios with discontinuous trajectories.

[0086] 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, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for verifying dangerous goods cargo manifests based on spatiotemporal trajectory matching, characterized in that, include: Task-based parsing is performed on the verification data for dangerous goods transportation, and the waybill execution information is sequentially divided into task nodes, including: Arrange the waybill execution information in the dangerous goods transportation verification data according to the task occurrence time sequence, identify the time points when the transportation location, cargo handover status and vehicle execution status change, use the adjacent time points of status change as the start and end boundaries of the task stage, divide the waybill execution information into the task stages of transportation preparation, loading and departure, in-transit transportation, arrival and unloading and delivery completion, and convert each task stage into task nodes to generate a dangerous goods waybill task node sequence. And configure spatiotemporal and business constraints for each task node and the transfer between nodes, and construct a dangerous goods cargo order task constraint graph; Based on the task constraint graph of dangerous goods freight bills, road matching is performed on vehicle positioning records to obtain road matching trajectory sequences. Based on the spatiotemporal continuity of adjacent trajectory points, reliable trajectory segments and trajectory missing intervals are divided. At the same time, based on the spatiotemporal correspondence between reliable trajectory segments and task nodes, completed task nodes are determined, and real-time trajectory verification information is generated. The missing trajectory intervals in the real-time trajectory verification information are located at their boundaries, and the preceding and following boundary states are extracted, including: Extract the adjacent reliable trajectory segments on both sides of the missing interval of each trajectory from the real-time trajectory verification information, set the end road position and end time of the previous reliable trajectory segment as the front boundary state, set the start road position and start time of the next reliable trajectory segment as the back boundary state, and calculate the missing duration based on the time difference between the front and back boundary states. In addition, candidate paths are deduced by combining road spatiotemporal constraints, and inaccessible paths are eliminated based on the continuous reachability of the preceding and following boundary states to obtain possible trajectories. Extract the current state of the vehicle for each possible trajectory, map it to the task constraint graph of the dangerous goods transport order, construct the initial executable domain of the remaining tasks, and perform layer-by-layer reachability propagation and constraint contraction on the initial executable domain of the remaining tasks starting from the current state of each vehicle to obtain the executable domain of the branch tasks. Based on the hazardous goods cargo manifest task constraint diagram, a set-level dynamic verification is performed on the possible trajectories and the executable domains of the corresponding branch tasks. The satisfaction of each branch with the constraints of the already occurred and remaining tasks is compared to determine the hazardous goods cargo manifest verification status, including: A set-level comparison is performed on the fulfillment information of the branch task chains. When all branches meet the constraints of the tasks that have occurred and there are continuous reachable execution chains for the remaining task paths corresponding to each branch, the verification status of the dangerous goods shipment order is determined to be verified as verified. When there are no branches that simultaneously meet the constraints of the tasks that have occurred and there are continuous reachable execution chains for the corresponding remaining task paths, the verification status of the dangerous goods shipment order is determined to be verified as ...

2. The dangerous goods cargo manifest verification method based on spatiotemporal trajectory matching as described in claim 1, characterized in that, The steps for constructing the task constraint graph for dangerous goods cargo manifests are as follows: Extract the effective time interval and applicable spatial region of each task node to form the spatiotemporal allowable range of each task node, and determine the transfer direction and transfer conditions based on the state acceptance requirements between adjacent task nodes to form transfer constraints between nodes. Based on the sequence of task nodes in the dangerous goods shipment, establish connection edges between task nodes, and configure the spatiotemporal allowable range of each task node and the transfer constraints between nodes to generate a dangerous goods shipment task constraint graph.

3. The dangerous goods cargo manifest verification method based on spatiotemporal trajectory matching as described in claim 1, characterized in that, The steps for obtaining the road matching trajectory sequence are as follows: Based on the task constraint graph of dangerous goods transport orders, vehicle location records bound to dangerous goods transport orders to be verified are extracted from the dangerous goods transport verification data and arranged in order of location time to generate a vehicle location sequence. Candidate roads are retrieved from the road network centered on each location point in the vehicle location sequence, and the spatial matching degree between the location point and the candidate road is calculated. At the same time, the road transfer matching degree is calculated by combining the topological connection relationship between adjacent location points and corresponding candidate roads. Based on spatial matching degree and road transfer matching degree, the comprehensive matching degree of each candidate road is calculated. Candidate roads with a comprehensive matching degree not lower than the preset comprehensive matching threshold are selected. The candidate road with the highest comprehensive matching degree among the selected candidate roads is determined as the road affiliation of the corresponding positioning point. The road positions of each positioning point are sequentially connected according to the time sequence of the vehicle positioning sequence to generate a road matching trajectory sequence.

4. The dangerous goods shipment verification method based on spatiotemporal trajectory matching as described in claim 3, characterized in that, The steps for generating real-time trajectory verification information are as follows: Based on the road matching trajectory sequence, the time interval between adjacent trajectory points and the road network travel distance are calculated to determine the travel speed between trajectory points. Combined with the road connection relationship, trajectory points that meet the spatiotemporal continuity condition are marked as continuous trajectory points, and trajectory points that do not meet the spatiotemporal continuity condition are marked as abnormal trajectory points. According to the positioning time sequence, continuous trajectory points that belong to the road and satisfy the road reachability relationship are sequentially aggregated into a reliable trajectory segment, and the positioning interruption and road jump that occur between the reliable trajectory segments are used as the interval boundary to divide the trajectory missing interval. The spatiotemporal coverage of the reliable trajectory segment is matched with the spatiotemporal allowable range of the task nodes in the dangerous goods shipment task constraint diagram. The execution order between task nodes is matched based on the transfer constraints between nodes. Task nodes that meet both the spatiotemporal conditions and the transfer conditions are identified as completed task nodes, and real-time trajectory verification information is generated.

5. The dangerous goods cargo manifest verification method based on spatiotemporal trajectory matching as described in claim 4, characterized in that, The steps for obtaining possible trajectories are as follows: Starting from the previous boundary state, the subsequent road states are expanded layer by layer according to the connection direction of the road network and the spatiotemporal constraints of the road. The reachable time range of each subsequent road state is calculated, and the road states that can be reached within the missing duration are selected to generate a forward reachable road set. Using the subsequent boundary state as the endpoint, the preceding road states are deduced layer by layer according to the reverse connection relationship of the road network. The travel time required for each preceding road state to connect to the subsequent boundary state is calculated, and the road states that can connect to the subsequent boundary state within the missing duration are selected to generate a set of backward reachable roads. Cross-matching is performed on the forward reachable road set and the backward reachable road set to extract the bidirectional reachable road states that exist in both the forward reachable road set and the backward reachable road set, and candidate paths are formed by combining the road states according to the connection relationship between them. Calculate the estimated travel time for each candidate path and verify the connection relationship between the first and last roads of the candidate path and the state of the front and rear boundaries. Filter out unreachable paths that do not meet the requirements of time continuity and road accessibility. At the same time, according to the positioning time order, connect the candidate paths retained in the missing interval of each trajectory with the corresponding reliable trajectory segments in sequence to form possible trajectories.

6. The dangerous goods cargo manifest verification method based on spatiotemporal trajectory matching as described in claim 5, characterized in that, The steps for constructing the initial executable domain for the remaining tasks are as follows: Extract the vehicle's spatiotemporal state at the latest positioning time from each possible trajectory and match it with the task nodes in the hazardous goods transport order task constraint diagram. When the vehicle's spatiotemporal state is within the spatiotemporal allowable range of a certain task node and all preceding task nodes have been completed, the task node is determined as the currently executing task node. When the vehicle's spatiotemporal state has left the spatiotemporal allowable range of the currently executing task node and satisfies the transfer constraint that the currently executing task node points to the next task node, the currently executing task node is determined as the completed task node and the next task node is determined as the task node to be executed. Determine the task progress based on the task order position of the currently executing task node and the task node to be executed in the hazardous goods transport order task constraint diagram, and bind the state according to the possible trajectory identifier to generate the current state branch of the vehicle. Map the current state branches of each vehicle to the task constraint graph of the dangerous goods transport order, and extract the unfinished task nodes from the task nodes where the corresponding task progress is located to generate the remaining task subgraph. The hierarchical spatiotemporal expansion is performed according to the task execution order in the remaining task subgraph, and the reachable state range of each task node is calculated based on the spatiotemporal allowable range of the task nodes and the transition constraints between nodes. The initial executable domain of the remaining tasks is then generated by summarizing these results.

7. The dangerous goods cargo manifest verification method based on spatiotemporal trajectory matching as described in claim 6, characterized in that, The steps for obtaining the executable domain of the branch task are as follows: Starting from the current state branch of each vehicle, according to the task execution level in the remaining task subgraph, perform layer-by-layer reachability propagation in the initial executable domain of the remaining tasks, calculate the transition time required from the current execution state to the next task node and determine the reachable path, and at the same time eliminate execution states that do not meet the spatiotemporal allowable range of task nodes and the transition constraints between nodes to obtain valid execution states; Based on the remaining task subgraph, the effective execution states of adjacent levels are continuously reachable, and the effective execution states that can be continuously propagated from the current state of the vehicle to the termination task node are retained to obtain the branch task executable domain of each possible trajectory.

8. The dangerous goods shipment verification method based on spatiotemporal trajectory matching as described in claim 7, characterized in that, The steps for determining the verification status of dangerous goods transport documents are as follows: Based on the possible trajectory identifiers, each possible trajectory is bound to the corresponding branch task executable domain and mapped to the dangerous goods freight bill task constraint graph. The task progress corresponding to the current state branch of the vehicle is used as the dividing boundary. The task nodes and connecting edges that have been completed before the task progress are determined as the already occurred task paths, and the unfinished task nodes, subsequent task nodes and connecting edges corresponding to the task progress are determined as the remaining task paths. Each possible trajectory is spatiotemporally matched with the corresponding completed task path. The performance status of the completed task path is determined based on whether the possible trajectory satisfies the spatiotemporal allowable range of the task node and the transfer constraints between nodes. The continuous reachable execution chain from the current state of the vehicle to the termination task node is retrieved along the remaining task path to generate branch task chain performance information.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the dangerous goods shipment verification method based on spatiotemporal trajectory matching as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the dangerous goods shipment verification method based on spatiotemporal trajectory matching as described in any one of claims 1 to 8.

Images