Logistics whole-process scheduling management system based on positioning and multi-platform data

Abstract

This invention relates to the field of logistics scheduling data management technology, specifically a logistics end-to-end scheduling and management system based on location and multi-platform data. The system includes: a data aggregation module that collects waybill master data, cargo status snapshots, and real-time vehicle location trajectories to form a multi-source heterogeneous logistics dataset; a fusion modeling module that performs cross-system format parsing and spatiotemporal calibration on the data to generate a standardized fusion data model; a graph construction module that binds vehicles and waybills, mapping the cargo status chain to the vehicle's operational path time sequence to construct a logistics spatiotemporal knowledge graph; a dynamic update module that injects real-time location points to update paths and triggers predictive cargo status shifts; and an instruction generation module that generates scheduling instructions based on the updated graph. This solution eliminates the spatiotemporal misalignment problem of multi-system data, connects warehousing and transportation end-to-end data, enables dynamic iteration of logistics status, and ensures that scheduling instructions are synchronized with the actual logistics operation status.

Description

Technical Field

[0001] This invention relates to the field of logistics scheduling data management technology, and in particular to a logistics end-to-end scheduling management system based on location and multi-platform data. Background Technology

[0002] Currently, in the logistics industry's dispatch and management processes, waybill information is only stored within the logistics carrier's system, cargo storage status is recorded separately by the warehouse management system, and vehicle location data is uploaded independently by the vehicle terminal. Various types of logistics data are only collected and stored within their respective systems without cross-platform integration. Data format standards differ across systems, and the recording time of cargo status and the spatiotemporal information of vehicle location are not uniformly calibrated. Logistics data in the warehousing and transportation stages remain independent, failing to form a coherent logistics data system.

[0003] Traditional logistics scheduling models cannot match changes in cargo status with vehicle trajectories in a timely manner, making it difficult to build a data framework covering the entire process from warehousing to transportation. New location data reported by vehicle terminals can only update the vehicle's path and cannot drive corresponding temporal deductions of cargo status. Scheduling instructions lack a unified and dynamic spatiotemporal data foundation. Therefore, it is necessary to parse and spatiotemporally calibrate multi-source heterogeneous logistics data, build a standardized fusion data model, bind vehicles and waybills, synchronously map cargo status and vehicle paths, and construct a correlation graph. Based on real-time location data, path updates and sequential changes in cargo status can be completed to generate corresponding scheduling decision instructions. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a logistics end-to-end scheduling and management system based on location and multi-platform data.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a logistics end-to-end scheduling and management system based on positioning and multi-platform data, comprising: The data aggregation module obtains waybill master data from at least one logistics carrier system, cargo status snapshots from at least one warehouse management system, and continuous real-time positioning trajectories from positioning terminals deployed on vehicles, forming a multi-source heterogeneous logistics dataset. The fusion modeling module performs cross-system format parsing and spatiotemporal calibration on the multi-source heterogeneous logistics dataset to generate a fusion data model containing a waybill object with unified coding, a cargo status chain with timestamp sequences, and a vehicle operation path with geographic coordinates and time alignment. The graph construction module, based on the fused data model, binds the vehicle identification code to the waybill object, synchronously maps the cargo status chain and the vehicle operation path on the timeline, and constructs a complete logistics spatiotemporal knowledge graph with waybill as the index and running through warehouse nodes and transportation paths; The dynamic update module injects new positioning points reported by the positioning terminal into the complete logistics spatiotemporal knowledge graph in real time, updates the vehicle's running path, and triggers the predictive shift of the cargo state chain. The instruction generation module generates a set of scheduling decision instructions based on the updated complete logistics spatiotemporal knowledge graph.

[0006] As a further aspect of the present invention, cross-system format parsing and spatiotemporal calibration are performed on the multi-source heterogeneous logistics dataset to generate a fusion data model containing a unified-coded waybill object, a cargo status chain with a timestamp sequence, and a vehicle operation path aligned with geographic coordinates and time, including: The field structure of the waybill master data is parsed, and the waybill identifier, cargo description, sender and receiver addresses, and planned time window information are extracted and converted into the internally unified waybill object; The status code and description text of the cargo status snapshot are parsed, the status code and description text are mapped to entries in the standard status set, and the cargo status chain with timestamp sequence is generated based on the snapshot time. The continuous real-time positioning trajectory is subjected to coordinate system transformation, drift point filtering and time synchronization compensation to generate a vehicle running path that is aligned with the geographic coordinates and time. Establish an association index based on a globally unique logistics event identifier for the waybill object, the cargo status chain, and the vehicle operation path, and complete the construction of the fused data model.

[0007] As a further aspect of the present invention, the cargo status chain and the vehicle operation path are synchronously mapped on a timeline to construct a complete logistics spatiotemporal knowledge graph indexed by waybills and spanning warehouse nodes and transportation paths, including: Using the timeline as the main axis, the timestamp of each state in the cargo state chain is marked on the timeline as a state node. The timestamps of each key trajectory point in the vehicle's running path are marked on the time axis as spatial nodes. The key trajectory points include the starting point, intersections along the route, highway entrances and exits, and stops. A bidirectional connection edge is established between the state node and the corresponding spatial node based on the matching degree of their timestamps. The bidirectional connection edge represents the consistency relationship between the cargo status and the vehicle spatial position at a specific point in time. By taking the waybill object as the root node, and connecting all its corresponding state nodes and spatial nodes, a complete logistics spatiotemporal knowledge graph with time as the dimension and spanning state and space is formed.

[0008] As a further aspect of the present invention, newly added positioning points reported by the positioning terminal are injected into the complete logistics spatiotemporal knowledge graph in real time, the vehicle's operating path is updated, and the predictive shift of the cargo state chain is triggered, including: The system receives the newly added positioning point and converts it into a standard format trajectory point, including coordinates, timestamp, vehicle speed and direction information. Insert the trajectory points in the standard format into the corresponding vehicle operation path sequence in the complete logistics spatiotemporal knowledge graph to form the updated vehicle operation path; Based on the updated vehicle's destination coordinates, speed, and direction, and combined with preset digital map road network data, the predicted path of the vehicle within a preset time period is calculated. Based on the predicted path, it is determined whether the vehicle will enter a logical region different from the current cargo state in the future. If so, a predictive cargo state node corresponding to the logical region is added to the end of the cargo state chain to complete the predictive shift of the cargo state chain.

[0009] As a further aspect of the present invention, the step of generating a set of scheduling decision instructions based on the updated complete logistics spatiotemporal knowledge graph includes: The updated complete logistics spatiotemporal knowledge graph is analyzed to identify spatial transformation events that will occur to the waybill object on the vehicle's running path. These spatial transformation events include arriving at the transfer center, leaving the city boundary, and approaching the destination fence. The identified spatial transformation events are matched with a preset logistics scheduling rule base to generate a set of scheduling decision instructions for a specific waybill object or a specific vehicle. Based on the set of scheduling decision instructions, a differentiated task message is generated and distributed to at least one warehouse management system, at least one logistics carrier system, or driver terminal.

[0010] As a further aspect of the present invention, the analysis of the updated complete logistics spatiotemporal knowledge graph to identify the spatial transformation events that will soon occur on the vehicle's operating path for the waybill object includes: Load a preset set of geofences, which includes electronic fences for multiple logistics nodes, administrative boundaries, and special road sections; In the updated complete logistics spatiotemporal knowledge graph, the real-time spatial relationship between the updated vehicle operation path and the predicted path, and the geofence set is calculated. When the calculation results show that the real-time location of the updated vehicle's operating path or the predicted path meets the conditions for entering, leaving, or staying at a specific geofence, a candidate event is generated. The candidate events are subjected to context verification, which includes verifying whether the current cargo status allows for conversion and whether the vehicle's historical trajectory supports inference. The candidate events that pass the verification are determined as the spatial conversion events.

[0011] As a further aspect of the present invention, the step of matching the identified spatial transformation event with a preset logistics scheduling rule base to generate a set of scheduling decision instructions for a specific waybill object or a specific vehicle includes: The logistics scheduling rule base contains multiple rules, each of which consists of an event pattern, context conditions, and execution actions; The type of the spatial transformation event, the logistics nodes involved, the current time, and the cargo attributes are used as inputs to traverse and match the rules in the logistics scheduling rule base. When the spatial transformation event satisfies the event pattern of a certain rule, and the current state of the complete logistics spatiotemporal knowledge graph satisfies the context condition of the rule, the rule is activated; The execution actions of all activated rules are extracted, merged and deduplicated to form the scheduling decision instruction set, which includes instruction content, execution object, and expected execution time window.

[0012] As a further aspect of the present invention, based on the set of scheduling decision instructions, a differentiated task message is generated, including: For each instruction in the set of scheduling decision instructions, the type of its execution object is identified. The types of execution objects include warehouse management system, logistics carrier dispatcher, and transport vehicle driver. Based on the different types of execution objects, an appropriate message template is selected, wherein the message template defines the structure, level of detail, and interaction method of the information; Contextual information related to the current instruction is extracted from the complete logistics spatiotemporal knowledge graph. The contextual information includes waybill details, real-time vehicle location, estimated arrival time, and associated cargo status. The instruction content and the extracted context information are filled into the selected message template to generate the task message that can be understood and processed by the corresponding execution object system or personnel.

[0013] As a further aspect of the present invention, the continuous real-time positioning trajectory is subjected to coordinate system-one transformation, drift point filtering, and time synchronization compensation to generate a vehicle operation path aligned with the geographic coordinates and time, including: Obtain the original positioning trajectory point sequence reported by the positioning terminal, the original positioning trajectory point sequence includes coordinate data and timestamp data based on different spatial references; Identify the original coordinate system used by each trajectory point in the original positioning trajectory point sequence, and uniformly transform the coordinate data of all trajectory points to the preset global standard coordinate system; In the converted trajectory point sequence, the instantaneous velocity and displacement between consecutive trajectory points are calculated. When the instantaneous velocity exceeds the preset physical velocity threshold or the displacement direction physically conflicts with the preset road network allowed direction, the corresponding trajectory point is marked as a drift point and removed from the sequence. For the trajectory point sequence after removing drift points, the timestamp of each trajectory point is compensated and calibrated according to the deviation between the built-in clock of the positioning terminal and the system standard time, so that the timestamps of all trajectory points are aligned with the standard time reference. The time-calibrated trajectory points are sorted and smoothed according to the timestamp order to form the vehicle running path aligned with the geographic coordinates and time.

[0014] As a further aspect of the present invention, based on the updated vehicle's endpoint coordinates, speed, and direction, and combined with preset digital map road network data, the predicted path of the vehicle within a preset future time period is calculated, including: From the updated vehicle running path, extract the latest trajectory point as the path endpoint, and obtain its coordinates, instantaneous velocity value and instantaneous direction of motion; Load the preset digital map road network data and map the coordinates of the path endpoint to the nearest road segment in the digital map road network data; Using the coordinates of the path endpoint, instantaneous velocity value, and instantaneous direction of motion as the initial motion state, the motion state is recursively deduced in the topological road network defined by the digital map road network data, according to road connectivity rules and traffic flow constraints. The motion state is recursively calculated within each recursion step, based on the current speed, road speed limit and historical average speed, to determine the position at the next moment, and the direction of motion is only allowed to be changed at road intersections according to a preset turning probability model. The motion state is continuously recursively calculated until the recursion time reaches the preset future time period. All position points generated during the recursion process are then connected in chronological order to form the predicted path of the vehicle within the preset future time period.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: Cross-system format parsing and spatiotemporal calibration are performed on multi-source heterogeneous logistics datasets to generate waybill objects with unified codes, cargo status chains with timestamp sequences, and vehicle operation paths with geographic coordinates aligned with time, forming a corresponding fused data model. Format differences between logistics data from different sources are eliminated, waybill data is standardized through unified coding, the temporal changes in cargo status are fully presented, the geographic coordinates and time nodes of vehicle operation are accurately matched, the spatiotemporal attributes of various types of logistics data are standardized and unified, the corresponding relationships between waybills, cargo, and vehicles are clearly defined, spatiotemporal misalignment between data is eliminated, the structured characteristics of logistics data are fully preserved, and the logical relationships between data in each dimension remain clear.

[0016] By binding vehicle identification codes to waybill objects, the cargo status chain and vehicle operation path are synchronously mapped on the timeline, constructing a logistics spatiotemporal knowledge graph indexed by waybills and connecting warehouse nodes and transportation paths. New location points reported by positioning terminals are injected into the graph in real time, updating vehicle operation paths and triggering predictive shifts in the cargo status chain. Waybills become the index carrier connecting the entire logistics process, linking logistics links between warehousing nodes and transportation paths into a complete whole. Real-time changes in vehicle location directly drive the dynamic iteration of the spatiotemporal knowledge graph. Cargo status can follow the vehicle's trajectory to achieve sequential changes in time. The set of scheduling decision instructions is generated based on real-time updated spatiotemporal correlation data, ensuring that the logistics links corresponding to the instructions are consistent with the actual spatiotemporal state of operation. Data correlation and dynamic iteration throughout the entire logistics process remain synchronized. Attached Figure Description

[0017] Figure 1 This is a sequence diagram of the logistics end-to-end scheduling and management system based on location and multi-platform data described in this invention. Figure 2 A flowchart for constructing a logistics spatiotemporal knowledge graph by synchronously mapping the cargo state chain with the vehicle operation path timeline. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0019] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0020] See Figure 1 The system includes a data aggregation module, which is responsible for acquiring master waybill data from at least one logistics carrier system, snapshots of cargo status from at least one warehouse management system, and continuous real-time location trajectories from positioning terminals deployed on transport vehicles. This data, from different sources and in varying formats, is integrated to form a multi-source heterogeneous logistics dataset. The fusion modeling module processes this dataset, performing cross-system format parsing and spatiotemporal calibration. Its output is a fused data model containing uniformly coded waybill objects, cargo status chains with timestamp sequences, and vehicle travel paths with strictly aligned geographic coordinates and time. The graph construction module utilizes this fused data model, binding waybill objects to specific vehicles through vehicle identification codes and synchronously mapping cargo status chains to vehicle travel paths on a unified timeline. This constructs a complete spatiotemporal knowledge graph of logistics, indexed by waybills and spanning warehouse nodes and transport paths. The dynamic update module continuously operates on this knowledge graph. It receives new location points reported by the positioning terminals in real time and uses these points to update the vehicle operation paths in the knowledge graph. This update process further triggers predictive shifts in the cargo state chain. The instruction generation module analyzes and calculates based on the updated complete logistics spatiotemporal knowledge graph to generate a set of executable scheduling decision instructions.

[0021] In one embodiment of the present invention, cross-system format parsing and spatiotemporal calibration are performed on multi-source heterogeneous logistics datasets to generate a fused data model containing a waybill object with unified coding, a cargo status chain with timestamp sequences, and vehicle operation paths aligned with geographic coordinates and time. This process includes the following steps: parsing the field structure of the waybill master data obtained from the logistics carrier system, extracting waybill identifier, cargo description, dispatch and receipt addresses, and planned time window information from each field, and converting and reorganizing this extracted information according to predefined internal data specifications to generate a waybill object with an internally unified format. Simultaneously, parsing cargo status snapshots obtained from the warehouse management system, identifying the status codes and descriptive text contained in the snapshots, mapping these codes and texts to specific entries in a predefined standard status set of the system, and arranging these standard status entries in chronological order according to the timestamp information of each snapshot to form a cargo status chain with a clear timestamp sequence. For continuous real-time positioning trajectories obtained from positioning terminals, coordinate system transformation, drift point filtering, and time synchronization compensation are required. Specifically, this involves acquiring the original positioning trajectory point sequence reported by the positioning terminal. This sequence contains coordinate data and timestamp data based on different spatial reference systems. The original coordinate system used by each trajectory point is identified, and a corresponding coordinate transformation algorithm is used to uniformly transform the coordinate data of all points to a preset global standard coordinate system. In the transformed trajectory point sequence, the instantaneous velocity and displacement between consecutive points are calculated. When the calculated instantaneous velocity exceeds a preset physical velocity threshold, or according to... When the direction of motion determined by the displacement vector physically conflicts with the permitted traffic direction of roads in the preset digital road network, the corresponding trajectory point is identified as a drift point and removed from the sequence. For the trajectory point sequence after removing drift points, the timestamp carried by each trajectory point is compensated and calibrated based on an estimation model of the deviation between the built-in clock of the positioning terminal and the system standard time, aligning the timestamps of all trajectory points with the standard time reference used by the system. Finally, the time-calibrated trajectory points are sorted according to their timestamps and processed using a trajectory smoothing algorithm to form a vehicle operation path with precise geographical and time alignment. A globally unique logistics event identifier is assigned or associated with the waybill object, cargo status chain, and vehicle operation path generated in the above process, and an association index is established between the three based on this identifier, thereby completing the construction of the fused data model.

[0022] In practical implementation, processing continuous real-time positioning trajectories involves starting with the original positioning trajectory point sequence reported by the positioning terminal. This original sequence may contain points like [{"lng": 114.123456, "lat": 22.543210, "coordType": "GCJ-02", "time": "2025-08-01 09:05:20"}, ...]. The original coordinate system used by each trajectory point in the original positioning trajectory point sequence is identified, and the coordinate data of all trajectory points are uniformly converted to the preset global standard coordinate system "WGS-84". The conversion process is completed by calling coordinate conversion library functions. In the converted trajectory point sequence, the instantaneous velocity and displacement between consecutive trajectory points are calculated to filter drift points. The instantaneous velocity between consecutive trajectory points is calculated using the following formula: in: Indicates instantaneous velocity. This represents the great circle distance between adjacent trajectory points on the Earth's surface. and These represent the timestamps of the subsequent trajectory point and the previous trajectory point, respectively. When the instantaneous velocity... When a trajectory point exceeds a preset physical speed threshold or its displacement direction is opposite to the road's permissible direction obtained from preset road network data, the corresponding trajectory point is marked as a drift point and removed from the sequence. In some embodiments, after drift point filtering, time synchronization compensation is performed on the trajectory point sequence. Since the built-in clock of the positioning terminal may have a fixed deviation from the system standard time (e.g., 2 seconds slower), the timestamp of each trajectory point is compensated and calibrated, changing the original timestamp "2025-08-01 09:05:20" to "2025-08-01 09:05:22" to align all trajectory point timestamps with the standard time reference. After time calibration, the trajectory points are sorted according to timestamp order and smoothed using a Kalman filter algorithm to form a vehicle operation path aligned with geographic coordinates and time. This path is represented as a sequence of [("2025-08-01 09:05:22", 114.123456, 22.543210), ("2025-08-01 09:05:25", 114.123460, 22.543212), ...]. Essentially, this establishes a globally unique logistics event identifier-based association index for the waybill object, the cargo status chain with timestamp sequences, and the vehicle operation path aligned with geographic coordinates and time. This globally unique logistics event identifier can be "EVENT_ABC123_20250801". The fused data model uses this identifier to associate the three parts of data, completing the construction of the fused data model.

[0023] In one embodiment of the present invention, the cargo state chain and the vehicle operation path are synchronously mapped on a timeline to construct a complete logistics spatiotemporal knowledge graph indexed by waybills and spanning warehouse nodes and transportation paths. This construction process includes: (See reference...) Figure 2 Within the system, a virtual timeline is constructed as the backbone of the knowledge graph. On this timeline, the timestamp corresponding to each standard state recorded in the cargo state chain is marked as a state node, recording the state type and the time of occurrence. On the timeline, a series of key trajectory points are extracted from the vehicle's operating path. These key trajectory points include the path's starting point, important intersections, highway entrances and exits, and stops. The timestamp of each key trajectory point is marked as a spatial node, recording its precise geographical coordinates and corresponding time. Connections are established between state nodes and spatial nodes. Specifically, the timestamps of state nodes and spatial nodes are matched. When the timestamps meet a preset matching requirement, a bidirectional connection edge is established between the state node representing the same or adjacent timestamps and the spatial node. This bidirectional connection edge represents the consistent association between a specific state of the cargo and the specific spatial location of the vehicle at a specific point in time. The waybill object, representing a specific logistics task, is taken as the root node of the knowledge graph. Through data relationships, the waybill object is connected to all its corresponding state nodes and all spatial nodes, thus forming a network knowledge structure with time as the core dimension, which can connect the changes in the state of goods with the spatial movement trajectory of the vehicle, namely, a complete logistics spatiotemporal knowledge graph.

[0024] In practice, the fusion data model of "Waybill ABC123" is used as input, and a virtual timeline is built internally as the backbone of the knowledge graph. This virtual timeline is a continuous timeline from the waybill creation time to the estimated delivery time. On the virtual timeline, the timestamp of each state in the cargo status chain is marked as a status node. For the cargo status chain of "Waybill ABC123" [("2025-08-01 08:30:00", "S01", "Picked up"), ("2025-08-01 09:05:20", "S02", "In transit")], the first status node is created at the position of "2025-08-01 08:30:00" on the virtual timeline. This status node records the status type "S01" and the status name "Picked up". The second status node is created at the position of "2025-08-01 09:05:20" on the virtual timeline. This status node records the status type "S02" and the status name "In transit". In practical implementation, the timestamps of each key trajectory point in the vehicle's operating path are marked as spatial nodes on a virtual timeline. For the vehicle operating path of "Waybill ABC123" with its geographic coordinates aligned with time [("2025-08-01 09:05:22", 114.123456, 22.543210), ... ("2025-08-01 09:30:00", 114.135000, 22.550000)], the system extracts data from the vehicle operating path with its geographic coordinates aligned with time. Key trajectory points are selected, including the path start point, highway entrances / exits, and stops. For example, the starting point with timestamp "2025-08-01 09:05:22" and coordinates (114.123456, 22.543210), and a highway exit with timestamp "2025-08-01 09:30:00" and coordinates (114.135000, 22.550000) are each used to create spatial nodes at their corresponding positions on the virtual timeline. These spatial nodes record the precise geographic coordinates under their respective timestamps. In some embodiments, bidirectional connections are established between state nodes and their corresponding spatial nodes. The basis for establishing bidirectional connections is the matching degree between the timestamps of the state nodes and the spatial nodes. The matching degree can be calculated using the following formula: in: A boolean result indicating whether or not alignment is enabled. The timestamp representing the state node. A timestamp representing a spatial node. This indicates the preset time difference threshold. When... When true, a bidirectional connection is established between the state node and the spatial node closest to the corresponding time. This bidirectional connection represents the consistency between the cargo status and the vehicle's spatial location at a specific point in time. For example, a connection is established between the "in transit" state node and the spatial node "2025-08-01 09:05:22". It can be understood that the waybill object is taken as the root node, connecting all its corresponding state nodes and spatial nodes. The waybill object "Waybill ABC123" is taken as the root node of the knowledge graph, and is connected to the state nodes "2025-08-01 08:30:00", "2025-08-01 09:05:20", and all related spatial nodes such as "2025-08-01 09:05:22" and "2025-08-01 09:30:00" through data relationship pointers, forming a complete logistics spatiotemporal knowledge graph with time as the dimension, spanning state and space. In some embodiments, the complete logistics spatiotemporal knowledge graph is stored in memory as an attribute graph structure, where nodes contain attributes such as "type," "timestamp," and "details," and bidirectional connecting edges contain a "relationship type" attribute. Optionally, the scale precision of the virtual timeline can be set to the second or millisecond level according to business needs. Optionally, the key trajectory point extraction algorithm can automatically identify based on path curvature changes, in addition to fixed rules. The construction of the complete logistics spatiotemporal knowledge graph is the foundational data format for subsequent dynamic updates and scheduling decisions.

[0025] In one embodiment of the present invention, newly reported positioning points are injected in real time into the complete logistics spatiotemporal knowledge graph, updating the vehicle's operating path and triggering predictive shifts in the cargo state chain. Specifically, the dynamic update module receives the newly reported positioning points from the positioning terminal, standardizes the data format of these points, and converts them into standard format trajectory points containing coordinates in a standard coordinate system, a calibrated timestamp, and calculated instantaneous vehicle speed and direction information. These standard format trajectory points are then inserted into the end of the vehicle's corresponding operating path sequence in the complete logistics spatiotemporal knowledge graph, allowing the path sequence to continue and thus forming the updated vehicle operating path. Based on the updated vehicle's operating path, the latest trajectory point at its end is extracted as the current path endpoint. The coordinates, instantaneous velocity, and instantaneous direction of motion of this point are obtained. Combined with the system's preset digital map road network data, the predicted path of the vehicle within a preset future time period is calculated. This calculation process includes mapping the coordinates of the path endpoint to the nearest road segment in the digital map road network data. Using these coordinates, instantaneous velocity, and direction of motion as the initial motion state, the motion state is recursively simulated within the topological road network structure defined by the digital map road network data, according to the connectivity rules between roads and traffic flow constraints. Within each set recursion step, the possible position at the next moment is calculated based on the current motion state, road speed limit information, and the vehicle's historical average speed. The direction of motion is only changed at road intersections according to a preset turning probability model. This recursion process continues until the simulation time reaches the length of the preset future time period. All position points generated during the recursion process are connected in chronological order to form the predicted path. Based on the calculated predicted path, it is determined whether the vehicle will enter a new logical region on its future trajectory that is different from the logical region associated with the current state of the cargo. For example, it may enter the "arrival at destination city" region from the "transportation in transit" region. If the determination result is yes, a predictive cargo state node is added at the end of the cargo state chain according to the type of the new logical region. This node is marked as the predicted state and has an estimated occurrence time, thereby completing the predictive shift of the cargo state chain.

[0026] In practical implementation, the dynamic updating of the complete logistics spatiotemporal knowledge graph and the predictive progression of the cargo state chain are illustrated through a continuous scenario. The dynamic update module receives new location points reported by the positioning terminal. The new location points may be reported in the original format {"deviceId":"Truck-001","lng":114.145000,"lat":22.560000,"speed":65.2,"heading":35.5,"time":"2025-08-0110:05:30"}. After receiving a new positioning point, the dynamic update module standardizes the data format of the new positioning point and converts it into an internally defined standard format trajectory point. The conversion operation includes converting the coordinates (114.145000, 22.560000) to the WGS-84 coordinate system, aligning the time "2025-08-01 10:05:30" with the system's standard time reference, and directly recording the speed value of 65.2 (unit: km / h) and the heading angle of 35.5 (unit: degrees). The generated standard format trajectory point is represented as (114.145000, 22.560000, 65.2, 35.5, 2025-08-01 10:05:30). In practice, standard format trajectory points are inserted into the corresponding vehicle operation path sequence in the complete logistics spatiotemporal knowledge graph. The vehicle operation path bound to the newly added positioning point device identifier "Truck-001" is found. The current end point of this vehicle operation path is (114.142000,22.558000,63.0,34.0,2025-08-0110:05:00). The new standard format trajectory points are appended to the end of this sequence according to their timestamp order to form the updated vehicle operation path. The updated vehicle operation path sequence is updated in memory as a new array containing the new points.

[0027] In some embodiments, based on the updated vehicle's endpoint coordinates, speed, and direction, and combined with preset digital map road network data, the predicted path of the vehicle within a preset future time period is calculated. The latest trajectory point in the updated vehicle's path is extracted as the path endpoint, for example, the point (114.145000, 22.560000, 65.2, 35.5, 2025-08-01 10:05:30), and its coordinates (114.145000, 22.560000), instantaneous speed value of 65.2, and instantaneous direction of motion of 35.5 are obtained. Preset digital map road network data is loaded. This data includes road segments, connectivity relationships, and attributes. The coordinates (114.145000, 22.560000) of the path endpoint are matched to the nearest road segment in the digital map road network data using a spatial mapping algorithm. This road segment has a unique identifier "Road_Segment_123". Using the coordinates of the path endpoint (114.145000, 22.560000), instantaneous velocity of 65.2, and instantaneous direction of motion of 35.5 as the initial motion state, the motion state is recursively calculated within the topological road network defined by the digital map road network data, according to road connectivity rules and traffic flow constraints. Within each recursive step, the position for the next moment is calculated based on the current velocity, road speed limit, and historical average velocity. The recursive process is constrained by the following formula: in: This represents the coordinate vector of the predicted position at the next moment. The coordinate vector representing the current position. This indicates the instantaneous velocity scalar value used in the recursion. This indicates the preset recursion step size time interval. This represents the unit direction vector of the current road segment. The direction of movement only changes at road intersections according to a preset turning probability model, which may define a 70% probability of going straight, a 20% probability of turning left, and a 10% probability of turning right. The movement state is continuously recursively extrapolated until the total recursion time reaches a preset future time period. Each location point generated during the extrapolation process is connected in chronological order to form the predicted path of the vehicle within the preset future time period. The predicted path can be represented as a series of coordinate sequences of future time points.

[0028] In practice, the predicted path is used to determine whether the vehicle will enter a logical region different from its current cargo status in the future. The system maintains a logical region definition library, mapping geographical coordinates to logical regions. For example, coordinates (114.145000, 22.560000) belong to the logical region "Shenzhen Intra-city Transportation," while the predicted coordinates (114.200000, 22.600000) that the vehicle is expected to reach in 15 minutes are mapped to the logical region "Leaving the Shenzhen Boundary." The cargo status node at the end of the current cargo status chain indicates the logical region "Shenzhen Intra-city Transportation," while the future coordinates shown by the predicted path indicate the logical region "Leaving the Shenzhen Boundary." Since "Leaving the Shenzhen Boundary" and "Shenzhen Intra-city Transportation" are different logical regions, the determination result is "yes." At the end of the cargo state chain, a predictive cargo state node corresponding to the logical region of "leaving the Shenzhen city boundary" is added. This new node contains the state information "about to leave Shenzhen city" and is marked as a predicted state. Its estimated occurrence time is set to the current time plus the predicted time offset for reaching the boundary of this logical region, such as "2025-08-01 10:20:30", thus completing the predictive shift of the cargo state chain. It can be understood that the predictive cargo state node is identified by a special attribute in the knowledge graph to distinguish it from the actual cargo state node. In some embodiments, the determination of the logical region can be based on geofencing or polygon range matching. Optionally, the triggering condition for the predictive shift can be configured, for example, adding a node only when a specific type of logical region is predicted to be entered. Optionally, the calculation of the predicted path can consider real-time traffic flow information to improve accuracy. The predictive shift of the cargo state chain provides a state basis for forward-looking scheduling.

[0029] In one embodiment of the present invention, a set of scheduling decision instructions is generated based on an updated complete spatiotemporal knowledge graph of logistics. This generation process includes real-time analysis of the updated complete spatiotemporal knowledge graph to identify spatial transformation events that are about to occur on the associated vehicle's path for each waybill object. These spatial transformation events include the vehicle's expected arrival at a transfer center, its imminent departure from a city boundary, or its approach to a destination's electronic fence. The specific steps for identifying spatial transformation events are as follows: The system loads a preset set of geofences, which includes the definitions of geofences for multiple logistics nodes, administrative boundaries, and special road sections. The system calculates the real-time spatial relationship between the latest real-time location of the vehicle's operating path and its predicted path in the updated complete logistics spatiotemporal knowledge graph and all geofences in the geofence set. The system determines the positional relationship between the location point or path segment and the geofence polygon. When the calculation results show that the real-time location or predicted path of the vehicle meets the conditions for entering, leaving, or staying in a specific geofence, a candidate record for this event is generated. The candidate record is then subjected to contextual verification, which includes verifying whether the current state of the goods allows this spatial transformation to occur and whether the historical trajectory characteristics of the vehicle support the inference of this event. The candidate record that passes all verifications is finally determined as a valid spatial transformation event. The identified spatial transformation events are matched against a pre-defined logistics scheduling rule base. This rule base contains multiple rules consisting of event patterns, context conditions, and execution actions. The matching process takes the type of spatial transformation event, the logistics nodes involved, the current time, and cargo attributes as input and searches through the rule base. When a spatial transformation event satisfies the event pattern defined by a rule, and the overall state of the current complete logistics spatiotemporal knowledge graph satisfies the context conditions of that rule, the rule is activated. All execution actions defined in the activated rules are extracted, merged, and deduplicated to form a set of scheduling decision instructions. This set contains the specific instruction content, the instruction execution object, and the expected execution time window. Based on the set of dispatch decision instructions, differentiated task messages are generated. For each instruction in the set, the type of execution object is identified, including warehouse management systems, logistics carrier dispatchers, or transport vehicle drivers. According to different execution object types, a template with suitable structure, information detail, and interaction method is selected from the message template library. Contextual information related to the instruction is extracted from the complete logistics spatiotemporal knowledge graph. The contextual information includes the relevant waybill details, real-time vehicle location, estimated arrival time, and associated cargo status. The instruction content and extracted contextual information are filled into the selected message template to generate a task message that can be understood and processed by the corresponding execution object system or personnel. These task messages are then distributed to the corresponding warehouse management system, logistics carrier system, or driver terminal.

[0030] In practical implementation, the process of generating a set of scheduling decision instructions based on the updated complete logistics spatiotemporal knowledge graph involves real-time analysis of the knowledge graph to identify specific events. To illustrate with a specific scenario, assuming the latest location of the updated vehicle's running path associated with "Waybill ABC123" in the updated complete logistics spatiotemporal knowledge graph is (114.200000, 22.600000), and its predicted path shows the vehicle moving in the forward direction (114.205000, 22.605000), the core step in analyzing the updated complete logistics spatiotemporal knowledge graph is to identify the spatial transformation events that will occur on the updated vehicle's running path for the waybill object. These spatial transformation events include the vehicle's expected arrival at the transfer center, its imminent departure from the city boundary, or its approach to the destination fence. The specific implementation of identifying spatial transformation events includes loading a preset set of geofences. This set of geofences contains the spatial definitions of electronic fences for multiple logistics nodes, administrative boundaries, and special road sections. The geofence set is stored in the system in the form of a data table; some definitions are shown in Table 1. Table 1: Geofencing Definition Table In practical implementation, the real-time spatial relationship between the updated vehicle operation path's real-time location and the predicted path in the updated complete logistics spatiotemporal knowledge graph and the geofence set is calculated in real time. The relative position of the real-time location point (114.200000, 22.600000) of the updated vehicle operation path and each geofence in the geofence set is calculated to determine whether the location point enters, leaves, or remains at a specific geofence. For the predicted path, it is calculated whether the path segment intersects with the geofence polygon within a future time period. The relationship between a point and the polygonal fence can be determined using the ray method, while the distance from a point to the fence boundary can be calculated using the spherical distance formula. in: This represents the great circle distance between two points. Represents the Earth's average radius. Indicates the latitude and longitude of the location point. This represents the latitude and longitude of the nearest point on the fence boundary. A candidate event is generated when the calculation results show that the real-time position or predicted path of the updated vehicle's operating path meets the conditions for entering, leaving, or staying at a specific geofence. Contextual validation is performed on the candidate event, including verifying whether the current cargo status allows for the transition and whether the vehicle's historical trajectory supports the inference. Taking the above candidate event as an example, the validation logic checks whether the current cargo status of "Waybill ABC123" is "in transit" and the logical region is "transportation within Shenzhen," and checks whether the vehicle's historical trajectory shows a movement trend from within Shenzhen outwards. Candidate events that pass the validation are determined as the formal spatial transition event "leaving the Shenzhen boundary."

[0031] In some embodiments, differentiated task messages are generated based on a set of scheduling decision instructions. For each instruction in the set of scheduling decision instructions, the type of the execution object is identified, including warehouse management systems, logistics carrier dispatchers, and transport vehicle drivers. An appropriate message template is selected based on the type of execution object. The message template defines the structure, level of detail, and interaction method of the information. For example, for the "warehouse management system" type, a machine-readable template with a JSON structure is selected; for the "transport vehicle driver" type, a concise text message template is selected. Contextual information related to the current instruction is extracted from a complete logistics spatiotemporal knowledge graph. This contextual information includes waybill details, real-time vehicle location, estimated arrival time, and associated cargo status. For example, details of "waybill ABC123," the current vehicle location (114.200000, 22.600000), the estimated arrival time at the Beijing transit center "2025-08-02 16:00:00," and the current cargo status "in transit" are extracted. The instruction content and extracted context information are populated into the selected message template to generate a task message that can be understood and processed by the corresponding execution target system or personnel. For example, the generated task message for the warehouse management system is a JSON message containing fields such as instruction type, waybill number, and estimated arrival time, while the message for the driver might be a text message: "Your waybill ABC123 has left Shenzhen. Please proceed as planned. It is expected to arrive at the Beijing transfer center at 16:00 tomorrow." It can be understood that the generated task message is distributed through the system interface to at least one corresponding warehouse management system, at least one logistics carrier system, or driver terminal. Optionally, the message template can be customized according to the interface specifications of different carriers or warehouses. Optionally, the expected execution time window can be used to set the priority or delay the message sending. The generation of differentiated task messages ensures that instruction information is conveyed to different executors in the most appropriate form.

[0032] In one embodiment of the present invention, the logistics scheduling rule base contains multiple rules, each consisting of three parts: an event pattern, context conditions, and execution actions. The event pattern describes the characteristics of a type of spatial transformation event that can trigger the rule; the context conditions define the state of the knowledge graph when the rule takes effect; and the execution actions define the specific scheduling instructions that should be generated after the rule is activated. The type of spatial transformation event, the identifiers of the logistics nodes involved in the event, the current time of the event, and the attribute information of the goods associated with the event are used as input parameters. All rules in the logistics scheduling rule base are traversed, and each rule is matched against the input parameters sequentially. The matching calculation first checks whether the input event characteristics conform to the event pattern defined by the rule, and then checks whether the real-time system state reflected by the current complete logistics spatiotemporal knowledge graph satisfies the context conditions of the rule. Context conditions may include the current state of the goods, the historical operating mode of the vehicle, time window limitations, etc. When a spatial transformation event simultaneously satisfies the event pattern and context conditions of a rule, that rule is marked as active. The system collects the execution actions of all activated rules, merges and deduplicates these actions to form a structured set of scheduling decision instructions. This set clearly lists the specific content of each instruction, the execution object to which the instruction points, and the time window information for when the instruction is expected to be executed.

[0033] Based on the set of scheduling decision instructions, differentiated task messages are generated. For each instruction in the set, its execution object field is parsed to identify the target type to which the instruction needs to be delivered. The execution object types are mainly divided into three categories: warehouse management system, logistics carrier dispatcher, and transport vehicle driver. The system maintains a message template library, in which different message templates are predefined for different types of execution objects. These templates differ in information structure, content detail, and interaction methods. For example, the message template sent to the warehouse management system may be a structured data interface message, while the one sent to the driver may be a concise text or voice prompt. Based on the identified execution object type, a suitable message template is selected from the template library. From the current complete logistics spatiotemporal knowledge graph, contextual information closely related to the instruction being processed is extracted. This information typically includes the details of the waybill involved, the latest real-time location of the vehicle, the estimated time of arrival at the relevant node, and the current status of the goods. The specific instruction content in the instruction and the relevant contextual information extracted from the knowledge graph are filled and assembled according to the format and fields defined by the selected message template to generate a task message that can be directly received, understood, and processed by the corresponding execution object system or personnel.

[0034] In some embodiments, multiple rules may be activated by the same event. The system needs to handle potential conflicts between rules or determine the execution order, which can be achieved by introducing rule priority weights. To resolve this, the formula is used to calculate the final priority: in: This represents the final priority weight value of the rule. This represents the basic priority score of the rule. The score represents the degree of closeness between the rule's contextual conditions and the current state of the knowledge graph. This represents the accuracy score of matching the event type with the rule event pattern. The execution actions of all activated rules are extracted, merged, and deduplicated to form a scheduling decision instruction set. This set includes the instruction content, the execution object, and the expected execution time window. For example, for the "leaving the Shenzhen city boundary" event, two rules may be activated simultaneously. One rule's execution action is "generate an instruction to send a pre-arrival notification to the destination city's transit center," and the other rule, due to the "high timeliness" attribute of the goods, has the execution action "generate an instruction to send a priority processing reminder to the destination city's dispatcher." The system merges these two instructions and removes any possible duplicates or complexities, forming a scheduling decision instruction set containing the two specific instructions.

[0035] In practical implementation, differentiated task messages are generated based on the set of scheduling decision instructions. For each instruction in the set, the type of the execution object is identified. These types include warehouse management systems, logistics carrier dispatchers, and transport vehicle drivers. For example, the instruction "Send a pre-arrival notification to the destination city transfer center" is executed by the "warehouse management system," while the instruction "Send a priority processing reminder to the destination city dispatcher" is executed by the "logistics carrier dispatcher." Appropriate message templates are selected based on the different execution object types. The message templates define the information structure, level of detail, and interaction method. For the "warehouse management system" execution object type, a structured JSON interface message template is selected, with fixed fields and machine readability. For the "logistics carrier dispatcher" execution object type, a message card template from an internal enterprise communication software is selected, containing key information summaries and operation buttons. For the "transport vehicle driver" execution object type, a concise text message template or voice broadcast template is selected. Contextual information related to the current instruction is extracted from the complete logistics spatiotemporal knowledge graph. This contextual information includes waybill details, real-time vehicle location, estimated arrival time, and associated cargo status. For example, for the two instructions mentioned above, the details of "Waybill ABC123" extracted from the graph include consignor and consignee information and cargo description; the real-time vehicle location is extracted as (114.200000, 22.600000); the estimated arrival time at the Beijing transit center is extracted as "2025-08-02 16:00:00"; and the associated cargo status is extracted as "in transit". It can be understood that the instruction content and extracted contextual information are filled into a selected message template to generate a task message that can be understood and processed by the corresponding execution object system or personnel. For messages to the warehouse management system, the instruction "send pre-arrival notification" and contextual information are filled into a JSON template to generate a message. For messages to the logistics carrier dispatcher, a graphic message card is generated with the title "Time-sensitive goods are about to arrive," and the content includes the waybill number and estimated time. In some embodiments, the message generation process may include filtering or de-identifying sensitive information. Optionally, for multiple instructions targeting the same task, the system can merge messages to reduce information overload. The generation of differentiated task messages ensures that different roles and systems can receive and process scheduling instructions in the most appropriate manner.

[0036] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A logistics end-to-end scheduling and management system based on location and multi-platform data, characterized in that: The system includes: The data aggregation module obtains waybill master data from at least one logistics carrier system, cargo status snapshots from at least one warehouse management system, and continuous real-time positioning trajectories from positioning terminals deployed on vehicles, forming a multi-source heterogeneous logistics dataset. The fusion modeling module performs cross-system format parsing and spatiotemporal calibration on the multi-source heterogeneous logistics dataset to generate a fusion data model containing a waybill object with unified coding, a cargo status chain with timestamp sequences, and a vehicle operation path with geographic coordinates and time alignment. The graph construction module, based on the fused data model, binds the vehicle identification code to the waybill object, synchronously maps the cargo status chain and the vehicle operation path on the timeline, and constructs a complete logistics spatiotemporal knowledge graph with waybill as the index and running through warehouse nodes and transportation paths; The dynamic update module injects new positioning points reported by the positioning terminal into the complete logistics spatiotemporal knowledge graph in real time, updates the vehicle's running path, and triggers the predictive shift of the cargo state chain. The instruction generation module generates a set of scheduling decision instructions based on the updated complete logistics spatiotemporal knowledge graph.

2. The logistics end-to-end scheduling and management system based on positioning and multi-platform data as described in claim 1, characterized in that, The multi-source heterogeneous logistics dataset undergoes cross-system format parsing and spatiotemporal calibration to generate a fusion data model containing a unified-coded waybill object, a cargo status chain with timestamp sequences, and vehicle operation paths aligned with geographic coordinates and time. This model includes: The field structure of the waybill master data is parsed, and the waybill identifier, cargo description, sender and receiver addresses, and planned time window information are extracted and converted into the internally unified waybill object; The status code and description text of the cargo status snapshot are parsed, the status code and description text are mapped to entries in the standard status set, and the cargo status chain with timestamp sequence is generated based on the snapshot time. The continuous real-time positioning trajectory is subjected to coordinate system transformation, drift point filtering and time synchronization compensation to generate a vehicle running path that is aligned with the geographic coordinates and time. Establish an association index based on a globally unique logistics event identifier for the waybill object, the cargo status chain, and the vehicle operation path, and complete the construction of the fused data model.

3. The logistics end-to-end scheduling and management system based on positioning and multi-platform data as described in claim 1, characterized in that, The cargo status chain and the vehicle operation path are synchronously mapped on the timeline to construct a complete logistics spatiotemporal knowledge graph indexed by waybills and spanning warehouse nodes and transportation paths, including: Using the timeline as the main axis, the timestamp of each state in the cargo state chain is marked on the timeline as a state node. The timestamps of each key trajectory point in the vehicle's running path are marked on the time axis as spatial nodes. The key trajectory points include the starting point, intersections along the route, highway entrances and exits, and stops. A bidirectional connection edge is established between the state node and the corresponding spatial node based on the matching degree of their timestamps. The bidirectional connection edge represents the consistency relationship between the cargo status and the vehicle spatial position at a specific point in time. By taking the waybill object as the root node, and connecting all its corresponding state nodes and spatial nodes, a complete logistics spatiotemporal knowledge graph with time as the dimension and spanning state and space is formed.

4. The logistics end-to-end scheduling and management system based on positioning and multi-platform data as described in claim 1, characterized in that, The system injects new location points reported by the positioning terminal into the complete logistics spatiotemporal knowledge graph in real time, updates the vehicle's operating path, and triggers predictive transitions in the cargo state chain, including: The system receives the newly added positioning point and converts it into a standard format trajectory point, including coordinates, timestamp, vehicle speed and direction information. Insert the trajectory points in the standard format into the corresponding vehicle operation path sequence in the complete logistics spatiotemporal knowledge graph to form the updated vehicle operation path; Based on the updated vehicle's destination coordinates, speed, and direction, and combined with preset digital map road network data, the predicted path of the vehicle within a preset time period is calculated. Based on the predicted path, it is determined whether the vehicle will enter a logical region different from the current cargo state in the future. If so, a predictive cargo state node corresponding to the logical region is added to the end of the cargo state chain to complete the predictive shift of the cargo state chain.

5. The logistics end-to-end scheduling and management system based on positioning and multi-platform data as described in claim 1, characterized in that, The process of generating a set of scheduling decision instructions based on the updated complete logistics spatiotemporal knowledge graph includes: The updated complete logistics spatiotemporal knowledge graph is analyzed to identify spatial transformation events that will occur to the waybill object on the vehicle's running path. These spatial transformation events include arriving at the transfer center, leaving the city boundary, and approaching the destination fence. The identified spatial transformation events are matched with a preset logistics scheduling rule base to generate a set of scheduling decision instructions for a specific waybill object or a specific vehicle. Based on the set of scheduling decision instructions, a differentiated task message is generated and distributed to at least one warehouse management system, at least one logistics carrier system, or driver terminal.

6. The logistics end-to-end scheduling and management system based on positioning and multi-platform data according to claim 1, characterized in that, The analysis of the updated complete logistics spatiotemporal knowledge graph identifies upcoming spatial transformation events for waybill objects along the vehicle's path, including: Load a preset set of geofences, which includes electronic fences for multiple logistics nodes, administrative boundaries, and special road sections; In the updated complete logistics spatiotemporal knowledge graph, the real-time spatial relationship between the updated vehicle operation path and the predicted path, and the geofence set is calculated. When the calculation results show that the real-time location of the updated vehicle's operating path or the predicted path meets the conditions for entering, leaving, or staying at a specific geofence, a candidate event is generated. The candidate events are subjected to context verification, which includes verifying whether the current cargo status allows for conversion and whether the vehicle's historical trajectory supports inference. The candidate events that pass the verification are determined as the spatial conversion events.

7. The logistics end-to-end scheduling and management system based on positioning and multi-platform data as described in claim 5, characterized in that, The process of matching the identified spatial transformation events with a preset logistics scheduling rule base to generate a set of scheduling decision instructions for a specific waybill object or a specific vehicle includes: The logistics scheduling rule base contains multiple rules, each of which consists of an event pattern, context conditions, and execution actions; The type of the spatial transformation event, the logistics nodes involved, the current time, and the cargo attributes are used as inputs to traverse and match the logistics scheduling rule base; When the spatial transformation event satisfies the event pattern of a certain rule, and the current state of the complete logistics spatiotemporal knowledge graph satisfies the context condition of the rule, the rule is activated; The execution actions of all activated rules are extracted, merged and deduplicated to form the scheduling decision instruction set, which includes instruction content, execution object, and expected execution time window.

8. The logistics end-to-end scheduling and management system based on positioning and multi-platform data according to claim 5, characterized in that, Based on the set of scheduling decision instructions, differentiated task messages are generated, including: For each instruction in the set of scheduling decision instructions, the type of its execution object is identified. The types of execution objects include warehouse management system, logistics carrier dispatcher, and transport vehicle driver. Based on the different types of the execution objects, an appropriate message template is selected, wherein the message template defines the structure, level of detail, and interaction method of the information; Contextual information related to the current instruction is extracted from the complete logistics spatiotemporal knowledge graph. The contextual information includes waybill details, real-time vehicle location, estimated arrival time, and associated cargo status. The instruction content and the extracted context information are filled into the selected message template to generate the task message that can be understood and processed by the corresponding execution object system or personnel.

9. The logistics end-to-end scheduling and management system based on positioning and multi-platform data according to claim 2, characterized in that, The continuous real-time positioning trajectory undergoes coordinate system transformation, drift point filtering, and time synchronization compensation to generate a vehicle operation path aligned with the geographic coordinates and time, including: Obtain the original positioning trajectory point sequence reported by the positioning terminal, the original positioning trajectory point sequence includes coordinate data and timestamp data based on different spatial references; Identify the original coordinate system used by each trajectory point in the original positioning trajectory point sequence, and uniformly transform the coordinate data of all trajectory points to the preset global standard coordinate system; In the converted trajectory point sequence, the instantaneous velocity and displacement between consecutive trajectory points are calculated. When the instantaneous velocity exceeds the preset physical velocity threshold or the displacement direction physically conflicts with the preset road network allowed direction, the corresponding trajectory point is marked as a drift point and removed from the sequence. For the trajectory point sequence after removing drift points, the timestamp of each trajectory point is compensated and calibrated according to the deviation between the built-in clock of the positioning terminal and the system standard time, so that the timestamps of all trajectory points are aligned with the standard time reference. The time-calibrated trajectory points are sorted and smoothed according to the timestamp order to form the vehicle running path with the geographic coordinates and time aligned.

10. The logistics end-to-end scheduling and management system based on positioning and multi-platform data according to claim 4, characterized in that, Based on the updated vehicle's destination coordinates, speed, and direction, and combined with preset digital map road network data, the predicted path of the vehicle within a preset future time period is calculated, including: From the updated vehicle running path, extract the latest trajectory point as the path endpoint, and obtain its coordinates, instantaneous velocity value and instantaneous direction of motion; Load the preset digital map road network data and map the coordinates of the path endpoint to the nearest road segment in the digital map road network data; Using the coordinates of the path endpoint, instantaneous velocity value, and instantaneous direction of motion as the initial motion state, the motion state is recursively deduced in the topological road network defined by the digital map road network data, according to road connectivity rules and traffic flow constraints. The motion state is recursively calculated within each recursion step, based on the current speed, road speed limit and historical average speed, to determine the position at the next moment, and the direction of motion is only allowed to be changed at road intersections according to a preset turning probability model. The motion state is continuously recursively calculated until the recursion time reaches the preset future time period. All position points generated during the recursion process are then connected in chronological order to form the predicted path of the vehicle within the preset future time period.

Images