A Rail Transit Scheduling Method and System Based on Four-Dimensional Spatiotemporal Trajectory
By using a scheduling method based on four-dimensional spatiotemporal trajectories and hierarchical control, the computational complexity and turnout reliability issues of PRT systems in large-scale urban vehicle scheduling were resolved, achieving efficient, reliable, and high-density vehicle operation.
Patent Information
- Application Number
- CN202611326454.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-29
- Publication Date
- 2026-11-17
AI Technical Summary
Existing PRT systems suffer from high computational complexity in large-scale urban vehicle dispatching, low reliability of trackside turnouts, and excessively large safety clearances that limit transport capacity, making it difficult to achieve efficient, reliable, and high-density vehicle operation.
A scheduling method based on four-dimensional spatiotemporal trajectory is adopted. The central scheduling platform plans a precise four-dimensional spatiotemporal trajectory. Combined with the hierarchical control of on-board units and track section controllers, the precise guidance and safe passage of vehicles are achieved. The traditional trackside turnout is abandoned, and autonomous track selection is achieved by using on-board track selection device.
It enables efficient, reliable, and high-density operation of large-scale urban vehicles, reduces computational complexity, decreases turnout failure points, narrows safety distances, and improves system capacity.
Smart Images

Figure CN123104736A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of urban rail transit scheduling and control technology, and in particular to a rail transit scheduling method and system based on four-dimensional spatiotemporal trajectory (three-dimensional spatial coordinates plus time coordinates), applicable to Personal Rapid Transit (PRT) systems, automated small rail transit systems, and urban micro-circulation rail transit systems. Background Technology
[0002] With the acceleration of urbanization, existing ground transportation modes are struggling to balance efficiency, safety, and personalized travel needs. High-capacity rail transit such as subways and light rail operate on fixed routes with stops at every station. While they have a large capacity, they cannot provide direct point-to-point travel and face difficulties in last-mile connections. Regular buses share municipal roads, are greatly affected by road conditions, and have low punctuality rates. Private cars offer flexibility but suffer from high energy consumption, severe congestion, and parking difficulties.
[0003] Personal Rapid Transit (PRT) systems offer a solution to these problems. PRT uses small, automated vehicles running on dedicated tracks, providing point-to-point direct service through offline stations. However, existing PRT systems have the following shortcomings: (1) Poor scalability of scheduling algorithms. Most existing PRT systems adopt centralized real-time dynamic scheduling. When the number of vehicles exceeds several hundred, the computational complexity of path search and conflict resolution increases sharply, which cannot support the simultaneous online operation of thousands or even tens of thousands of vehicles in a city. Existing PRT systems in operation are limited to small-scale single-line scenarios such as airports and parks.
[0004] (2) Low reliability of trackside turnouts. Traditional PRTs rely on trackside electric turnouts to separate vehicles onto tracks. Turnouts contain a large number of mechanical moving parts and independent drive devices, which are prone to wear and jamming, resulting in a high failure rate. Moreover, once a turnout fails, it will paralyze the entire line. A few patents have proposed the concept of on-board track switching (such as CN104691581A), but its track switching device is a passive guiding structure that relies on the guide groove on the track surface for guidance. The vehicle cannot autonomously select the track selection time and lacks precise time coordination with the global scheduling system.
[0005] (3) Excessive safety distance limits transport capacity. Due to the uncertainty of vehicle trajectory, the system forces a safety distance of tens of meters to avoid collisions, which greatly reduces the line's throughput capacity and makes it difficult to reach the level of medium-capacity (5,000 to 10,000 people / hour) rail transit.
[0006] Therefore, there is a need for a rail transit scheduling method and system that can support real-time scheduling of large-scale vehicles at the city level, has high computational efficiency, and can eliminate trackside turnout faults and reduce safety distances. Summary of the Invention
[0007] The purpose of this invention is to overcome the shortcomings of the prior art and provide a rail transit scheduling method and system based on four-dimensional spatiotemporal trajectory. Through the technical concept of "pre-planning, precise execution, and hierarchical control", the scheduling problem is transformed into a deterministic spatiotemporal trajectory matching problem, thereby achieving high-efficiency, high-reliability, and high-density vehicle operation.
[0008] To achieve the above objectives, the present invention adopts the following technical solution: A rail transit scheduling method based on four-dimensional spatiotemporal trajectory includes the following steps: S1. The central dispatch platform obtains a travel time and space resource application. The travel time and space resource application includes at least origin and destination information. The travel time and space resource application can be generated by the central dispatch platform itself, or initiated by the local controller of the track segment, user terminal or vehicle. S2. The central dispatch platform searches for available four-dimensional spatiotemporal windows in the unallocated spatiotemporal right-of-way resources using an incremental matching method, based on the allocated spatiotemporal right-of-way resource occupancy table, to plan a four-dimensional spatiotemporal trajectory containing the entire time and spatial coordinates for the requesting vehicle. When further optimization of departure time or resource utilization efficiency is required, based on the allocated spatiotemporal right-of-way resources, the spatiotemporal trajectories of some online vehicles are locally adjusted to obtain a better spatiotemporal window, and a four-dimensional spatiotemporal trajectory is planned for the requesting vehicle. The four-dimensional spatiotemporal trajectory specifies the vehicle's spatial position, speed, allowable deviation threshold, track selection direction at each track splitting node, and passage time window at the track merging node at each moment. S3. The central dispatch platform sends the four-dimensional spatiotemporal trajectory to the vehicle and the local controller of each track segment the vehicle passes through. After receiving the four-dimensional spatiotemporal trajectory, the vehicle sends a confirmation to the central dispatch platform. After the confirmation is correct, the four-dimensional spatiotemporal trajectory is strictly followed by the vehicle. S4. The vehicle travels on the track according to the four-dimensional spatiotemporal trajectory. The on-board unit senses the vehicle's position and speed in real time through vehicle-road cooperative interaction, compares it with the theoretical spatiotemporal coordinates specified by the four-dimensional spatiotemporal trajectory, and controls the vehicle's movement to keep the deviation between the vehicle's actual spatiotemporal coordinates and the four-dimensional spatiotemporal trajectory within the allowable deviation threshold range. The allowable deviation threshold is determined by the central dispatching platform when planning the four-dimensional spatiotemporal trajectory based on the vehicle's travel speed at each point in time, track environment conditions, and the response speed of the vehicle and ground equipment, and is included in the four-dimensional spatiotemporal trajectory issued to the vehicle. S5. The local controller of each track section monitors the operating status data of all vehicles in the section in real time through the sensors deployed in the section, compares the actual operating status of the vehicles with the four-dimensional spatiotemporal trajectory, and when the deviation exceeds the allowable deviation threshold, it shall be handled in accordance with the local safety handling plan and reported to the central dispatch platform. S6. When the preset triggering conditions are met, the central dispatch platform generates and distributes the updated four-dimensional spatiotemporal trajectory to replace the vehicle's current four-dimensional spatiotemporal trajectory.
[0009] Furthermore, in step S2, the incremental matching method includes: the central dispatching platform maintains a network-wide spatiotemporal resource occupancy table indexed by track segments and time slices, transforming the four-dimensional spatiotemporal trajectory planning of newly added vehicles into an incremental matching problem of searching for available spatiotemporal windows segment by segment along the origin-end point path within the spatiotemporal resource gaps already occupied by online vehicles; when further optimization of departure time or resource utilization efficiency is needed, further speed fine-tuning or time window shifting is performed on some online vehicles within the affected range to free up better spatiotemporal windows without affecting the normal operation of other online vehicles. This incremental matching only involves newly added vehicles and local vehicles affected by fine-tuning, rather than a complete replanning of all vehicles in the network. Since the physical influence range of newly added vehicles in the road network is limited by the spatiotemporal neighborhood of their travel paths, the number of vehicles affected by fine-tuning has a fixed upper bound (usually a constant k) when the road network topology and vehicle density are fixed. The computational cost of a single incremental matching depends only on the number of nodes passed through by the vehicle's travel path (path length L) and the vehicle constant k within the local influence range, and is independent of the total number of online vehicles N in the entire network. The total computational resource consumption of the entire network increases approximately linearly with the scale of online vehicles in the form of O(N), rather than exponentially, which allows small computing clusters to support the scheduling of large-scale vehicles at the city level.
[0010] Furthermore, the preset triggering conditions in step S6 include fault handling triggering, passenger destination change triggering, network-wide empty vehicle balance allocation triggering, emergency avoidance demand triggering, spatiotemporal resource utilization efficiency optimization triggering, and driving route optimization triggering. After the triggering causes the four-dimensional spatiotemporal trajectory to be updated, the central dispatching platform issues a new four-dimensional spatiotemporal trajectory. After receiving the new trajectory information and confirming it with the central dispatching platform, the vehicle executes according to the new trajectory, which is different from completely real-time dynamic adjustment.
[0011] Furthermore, the four-dimensional spatiotemporal trajectory planned in step S2 also includes merging node time window information. Before merging into the main highway, the vehicle has adjusted its speed to match the main highway cruising speed and enters the main highway within the time window specified by the four-dimensional spatiotemporal trajectory. The width of the time window is not less than the duration required for the vehicle to pass through the merging point and a safety margin is left before and after, so that the vehicle does not interfere with the existing traffic flow on the main highway when merging into the main highway.
[0012] Furthermore, in step S4, before the vehicle reaches the track splitting node, the onboard track selection device is pre-controlled to switch to the state corresponding to the target track based on the pre-defined track selection timing in the four-dimensional spatiotemporal trajectory. The onboard track selection device completes the state switch before the vehicle reaches the track splitting node, guiding the vehicle into the target track through the engagement relationship with the track structure at the track splitting node. This scheme eliminates the need for traditional trackside switches, allowing the vehicle itself to complete the track selection action, significantly reducing trackside equipment failure points.
[0013] Furthermore, when the on-board track selection device experiences drive failure or mechanical jamming, the on-board track selection device automatically switches to a safe state through an elastic reset mechanism, allowing the vehicle to leave the high-speed mainline and enter the next station or buffer section, without interfering with the mainline traffic flow.
[0014] Furthermore, the method adopts a three-level computing power sharing architecture: the central scheduling platform is responsible for global four-dimensional spatiotemporal trajectory planning and macro scheduling decisions; the local controller of the track segment is responsible for detecting the spatiotemporal position deviation of all vehicles in the segment, interaction between adjacent segments, and control of facilities in the segment; the on-board unit is responsible for real-time position, speed, self-state and environmental perception, execution of local control commands, and adjustment of vehicle motion.
[0015] Furthermore, the allowable deviation threshold in step S4 is determined by the central scheduling platform when planning the four-dimensional spatiotemporal trajectory based on the vehicle's speed at each point in time, the track environment conditions, and the response speed of the vehicle and ground equipment. Different thresholds are configured for different speed levels: when the speed is low, the vehicle moves a short distance per unit time, and the position deviation threshold is correspondingly reduced; when the speed is high, the position deviation threshold is correspondingly increased.
[0016] This invention also provides a rail transit dispatching system based on four-dimensional spatiotemporal trajectory to implement the above method, comprising a central dispatching platform, multiple track segment local controllers, and on-board units. The central dispatching platform includes a computing cluster and storage devices; the track segment local controllers include processors, memory, and communication interfaces; the on-board units include a vehicle-road cooperative communication module, a positioning module, environmental sensors, status sensors, and an execution controller. The central dispatching platform and each track segment local controller are connected via a fiber optic backbone network; each track segment local controller is connected to the on-board units within its jurisdiction via a wireless redundant communication network.
[0017] Furthermore, the vehicle is also equipped with an on-board track selection device, including a track selection actuator, a drive unit, and an elastic reset mechanism.
[0018] The beneficial effects of this invention are mainly reflected in: (1) By pre-assigning a four-dimensional spatiotemporal trajectory (roadbook) containing the time and space coordinates of the entire journey to each vehicle, the complex real-time dynamic scheduling problem is transformed into an incremental spatiotemporal window matching problem. The computational complexity increases almost linearly with the number of vehicles, making it possible to operate a large-scale (ten thousand vehicles) network of vehicles at the city level.
[0019] (2) A three-level computing power sharing architecture of "centralized planning + decentralized control" is adopted. The central platform is responsible for second-level macro planning, the segment controller is responsible for millisecond-level local processing, and the vehicle unit is responsible for real-time closed-loop control, which ensures both global optimal scheduling and local rapid response.
[0020] (3) The on-board track selection device replaces the traditional trackside turnout. The vehicle completes the track selection autonomously in advance according to the route book and completes the state switching and safety verification before reaching the node. There is no need to wait for the preceding vehicle to pass before the turnout is activated. This fundamentally eliminates the limitation of the turnout activation time on the vehicle passage interval and can ensure the safe passage of high-density vehicles. It greatly reduces the number of fault points and eliminates the risk of chain congestion caused by turnout failure. The elastic reset mechanism ensures that the vehicle automatically leaves the main line when the drive fails.
[0021] (4) Based on the predictability of four-dimensional spatiotemporal trajectory, a small nominal safety distance (meter level) can be adopted between vehicles. Combined with the speed-related allowable deviation threshold and multi-level safety redundancy, the theoretical carrying capacity of a single line can reach the level of medium-capacity rail transit.
[0022] (5) Acceleration and deceleration are completed on the auxiliary track, and the main line vehicles always cruise at high speed. The merging time window is precisely matched to eliminate the interference of vehicle acceleration, deceleration and lane change on the main line traffic flow. Attached Figure Description
[0023] Figure 1 This is a schematic diagram of the overall process of the scheduling method of the present invention; Figure 2 A schematic diagram of a three-tier computing power sharing architecture; Figure 3 A schematic diagram for matching time windows at vehicle merging nodes. Detailed Implementation
[0024] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.
[0025] Example 1: Overall Flow of Scheduling Method and Incremental Matching Algorithm Reference Figure 1This embodiment provides a rail transit scheduling method based on four-dimensional spatiotemporal trajectory, applied to a city-level grid-based light rail transit system. The system includes a central scheduling platform, multiple local controllers (hereinafter referred to as "segment controllers") arranged in segments along the rail network, and multiple lightweight autonomous vehicles operating on the tracks. The rail network is formed by splicing together standard modular operating segments end-to-end, creating a grid-like transportation network covering the city.
[0026] The specific scheduling process is as follows: S1. Passengers walk to the nearest distributed micro-station and input their destination via voice interaction or touch control. After verifying their identity, the segment controller sends a travel time and space resource request containing the origin and destination station information to the central dispatch platform. In addition, travel time and space resource requests can also be initiated remotely by the user terminal, by the vehicle in operation, or automatically by the central dispatch platform based on the network's empty vehicle allocation needs.
[0027] S2. After receiving a travel time and space resource application, the central dispatch platform executes an incremental matching algorithm. The central dispatch platform maintains a network-wide time and space resource occupancy table, using track segments and time slices as basic index units. Each record contains information such as the occupying vehicle identifier, start time, end time, and travel speed. The algorithm searches for available time and space windows segment by segment along the path from the start point to the destination: it prioritizes searching for continuous available windows that meet the overall constraints within the gaps in time and space resources already occupied by online vehicles, and if such windows exist, they are directly allocated; when further optimization of departure time or resource utilization efficiency is needed, local fine-tuning is performed on some online vehicles within the affected area (such as fine-tuning speed within a certain range or shifting time windows) to free up better time and space windows, rather than re-planning all vehicles in the network.
[0028] The computational complexity of this incremental matching increases nearly linearly with the number of online vehicles. Specifically, since the physical influence range of any newly added vehicle in the road network is limited by the spatiotemporal neighborhood of its travel path, the number of vehicles affected by fine-tuning has a fixed upper bound (usually a constant k) given a fixed road network topology and vehicle density. Therefore, the computational cost of a single incremental matching depends only on the number of nodes traversed by the vehicle's travel path (path length L) and the vehicle constant k within the local influence range, and is independent of the total number of online vehicles N. Compared to the exponentially increasing complexity of traditional real-time rescheduling algorithms with N, the incremental matching mechanism of this invention makes the total computational resource consumption of the entire network increase approximately linearly with the scale of online vehicles—that is, when the total number of vehicles doubles, only double the number of independent single matching requests need to be processed, while the computation time of a single request remains almost constant, thus supporting second-level real-time scheduling of large-scale vehicles at the city level. Taking a typical city-level network as an example, even with tens of thousands of vehicles online simultaneously, a small computing cluster can complete trajectory planning in seconds.
[0029] The generated four-dimensional spatiotemporal trajectory (hereinafter referred to as the "route book") specifies the vehicle's entire journey from the starting point to the end point, including: the spatial position (track segment and mileage) at each moment, the driving speed, the allowable deviation threshold, acceleration / deceleration sections, the track selection direction at each track splitting node, and the merging time window. Typical parameters of the route book are: cruising speed of 120 km / h, nominal steady-state headway of 5 m, and corresponding time interval of approximately 150 ms.
[0030] S3. The central dispatch platform distributes the route book to the requesting vehicle and all segment controllers along the route. Upon receiving the route book, the vehicle sends a confirmation to the central dispatch platform. Once confirmed, the route book officially takes effect and must be strictly followed by the vehicle. Each segment controller receives and stores the route book segment relevant to itself for subsequent deviation comparison. This confirmation mechanism ensures no communication errors or data loss during the trajectory distribution process.
[0031] S4. The vehicle travels on the track according to the route map. The onboard unit senses the vehicle's position and speed in real time through vehicle-to-infrastructure (V2I) interaction. The positioning methods include absolute positioning using barcode rulers laid along the track (the onboard reader scans continuously arranged position-coded barcode rulers on the track to obtain absolute mileage), inertial navigation, and multi-source fusion of track beacon positioning. This data is compared with the theoretical spatiotemporal coordinates specified in the route map, and the vehicle's movement is controlled. The vehicle speed is adjusted by controlling the drive and braking systems to keep the deviation within the allowable deviation threshold. The vehicle maintains a constant cruising speed of 120 km / h on the high-speed mainline, and all acceleration and deceleration are completed on the auxiliary track.
[0032] The permissible deviation threshold is determined by the central dispatch platform during route planning based on the vehicle's speed at various points in time, track environment conditions (such as curvature, gradient, and coefficient of friction), and the response speed of the vehicle and ground equipment. This threshold is then distributed along with the track as part of the route plan. For example, at a cruising speed of 120 km / h (approximately 33.3 m / s), the typical position deviation threshold is ±1 m, corresponding to a time deviation of approximately 30 ms, which is far less than the nominal headway of 150 ms. Therefore, even with deviation, a collision will not occur. At lower speeds (such as 30 km / h in station areas), the vehicle travels a shorter distance per unit time, and the position deviation threshold decreases accordingly (e.g., ±0.3 m); at higher speeds, the position deviation threshold increases accordingly.
[0033] S5. Simultaneously, each segment controller independently monitors the position, speed, acceleration, and distance data of all vehicles within its segment using sensors deployed within that segment, and independently compares this data with the corresponding stored route data segments. When a vehicle deviation exceeds the permissible deviation threshold, the segment controller can activate a local safety contingency plan within milliseconds (such as instructing vehicles to accelerate or decelerate, notifying other online vehicles to avoid the deviation, etc.) and report to the central dispatch platform. Adjacent segment controllers are directly connected via dedicated fiber optic cables, supporting rapid on-site handling and coordination in abnormal scenarios without waiting for instructions from the central platform.
[0034] S6. When any of the following triggering conditions occur, the central dispatch platform generates and issues an updated route book, replacing the route book currently being executed by the vehicle: ① A fault occurs in a certain section requiring a detour; ② Passengers change their destination during the trip; ③ The need for balanced allocation of empty vehicles across the entire network; ④ Emergency avoidance requirements; ⑤ Trajectory optimization to improve the efficiency of spatiotemporal resource utilization across the entire network; ⑥ Optimization of driving routes based on the overall network load. After receiving the new route book information and confirming it with the central dispatch platform, the vehicle executes the new route book. This mechanism differs from schemes that continuously control each vehicle in real time; it updates the trajectory only when needed, significantly reducing the communication and computational burden.
[0035] Example 2: Three-tier computing power sharing architecture and layered collaborative processing Reference Figure 2 This invention adopts a three-level computing power sharing architecture of central-section-vehicle, with clear division of responsibilities and response time scales at each level: The first level, the central dispatch platform, is deployed in the control center and consists of computing clusters and storage devices. It is responsible for global route planning, macro-level dispatch decisions, and maintenance of the network-wide spatiotemporal resource occupancy table, with a computation cycle in the order of seconds. The central platform does not handle millisecond-level real-time vehicle control; it is only responsible for macro-level decision-making and route generation and updates.
[0036] The second level is the local controller for each track segment: one segment controller is deployed in each standard modular operating segment, including a processor, memory, and communication interface. It is responsible for monitoring the spatiotemporal position deviation of all vehicles in the segment, interacting with adjacent segments, controlling facilities in the segment, and handling local anomalies. The segment controller is connected to the central platform through a fiber optic backbone ring network, and adjacent segment controllers are directly connected by dedicated fiber optic cables, with a response time in the millisecond range.
[0037] The third level is the on-board unit: Each vehicle is equipped with an on-board controller and sensor group (including barcode ruler reader, combined positioning module, accelerometer, environmental perception sensor, etc.). The barcode ruler reader obtains the absolute position of the vehicle by scanning the position-coded barcode ruler laid along the track. It is responsible for real-time position and speed perception, self-state monitoring, environmental perception, execution of local control commands, and closed-loop adjustment of vehicle motion. The control cycle is in the millisecond range.
[0038] The onboard unit communicates with the section controller via a vehicle-to-infrastructure (V2I) wireless communication network. In this embodiment, the V2I wireless communication network adopts a heterogeneous communication architecture with 5G-R (5G dedicated network for railways) and LTE-M (LTE dedicated network for urban rail transit) as the backbone, and Wi-Fi 6 / 7 for high-density in-vehicle access and supplementary redundancy. 5G-R provides a high-bandwidth, low-latency vehicle-to-ground data transmission channel, while LTE-M provides a highly reliable, wide-coverage control command transmission channel, with the two networks serving as hot backups for each other. Wi-Fi 6 / 7 is used for high-density data access (such as video surveillance and batch data uploads) within stations and sections, and serves as supplementary redundancy to the backbone network. During high-speed operation, the vehicle interacts with the section controller in real time with operational status information via this wireless network. When either network fails, critical control data can be seamlessly switched to the other network within milliseconds, ensuring uninterrupted communication.
[0039] Example of hierarchical collaborative handling: When a segment controller detects that the actual position of vehicle A within its segment lags behind the position specified in the roadbook and exceeds the allowable deviation threshold, the segment controller sends a deceleration or acceleration correction command directly to vehicle A via the fiber optic network within milliseconds. Simultaneously, it records the event and reports it to the central platform. Upon receiving the report, the central platform assesses whether the deviation affects global resource usage in the next scheduling cycle (seconds). If adjustments to the roadbooks for subsequent vehicles are needed, the roadbooks are recalculated and reissued. This hierarchical processing avoids the central platform becoming a bottleneck for real-time control while ensuring rapid response to anomalies.
[0040] Example 3: Merging Node Time Window Matching Reference Figure 3 When vehicles need to enter the main highway from the toll station, they will merge according to the following procedure: After exiting the station, the vehicle travels at low speed along the branch line to the merging waiting area. According to the merging time window specified in the road plan, the vehicle has already adjusted its speed to match the mainline cruising speed of 120 km / h before merging onto the expressway mainline. The road plan ensures that no other vehicles occupy the corresponding position on the mainline within this time window, the width of the time window is not less than the duration required for the vehicle to pass through the merging point, and there is a safety margin before and after. After adjusting its speed, the vehicle enters the expressway mainline at the midpoint of the designated window. At this time, the corresponding position on the mainline is exactly empty, eliminating the need to wait or slow down at the merging point and avoiding interference with existing traffic flow on the mainline.
[0041] Similarly, when a vehicle arrives at its destination area, it enters the deceleration auxiliary track at the location specified in the roadbook, slows down to a low speed on the dedicated deceleration section, and then enters the station along the station branch line. The deceleration process does not occupy mainline resources. The mainline and auxiliary track separation design ensures that mainline vehicles always maintain a high cruising speed, eliminating the disturbance to mainline traffic flow caused by acceleration, deceleration, and lane changes.
[0042] Example 4: Structure and Control of Onboard Track Selection Device The vehicle is equipped with an on-board track selection device, including a base, track selection actuator, drive unit, elastic reset mechanism and position sensor.
[0043] The base is fixedly installed under the vehicle chassis. The track selection actuator is mounted on the base via a rotating shaft, allowing it to rotate around a vertical axis and possessing at least two positioning states, corresponding to direction 1 and direction 2 respectively. The lower end of the actuator is equipped with a guide component that matches the fixed guide structure at the track node. The drive unit drives the rotating shaft to rotate to the target angle according to the track selection command issued by the on-board unit. An elastic reset mechanism is connected to the actuator, automatically resetting the actuator to a default safe state when the drive unit is powered off or fails. A position sensor is used to detect the actual state of the actuator and feed it back to the on-board unit.
[0044] In one specific embodiment, the guide component of the track selection actuator is a pair of guide wheels, fixed to the lower end of the rotating shaft, with two positioning protrusions corresponding to the straight and lateral track-changing states respectively; the drive unit includes an electromagnet and an armature, and when energized, the electromagnet attracts the armature to push the rotating shaft to rotate to the target angle; the elastic reset mechanism is a torsion spring mounted on the rotating shaft, with one end fixed to the base and the other end fixed to the actuator; the position sensor is an angle sensor.
[0045] This invention is not limited to the specific embodiments described above. The drive unit can also be driven by a motor (such as a servo motor or stepper motor), hydraulically, or pneumatically; the guide component of the track selection actuator can also be a guide slider, guide pin, or other form that matches the track guide structure; the elastic reset mechanism can also be a compression spring, leaf spring, or other elastic energy storage element; the engagement method can be an embedded guide groove, a side guide rail, or a bayonet structure. As long as the state switching can be completed before the vehicle reaches the node, the vehicle can be guided into the target track through the engagement relationship with the track fixed guide structure, and the drive can be automatically reset to a safe state when the drive fails, it falls within the protection scope of this invention.
[0046] According to the pre-defined track selection timing (format including node identifier, target direction, and execution time) in the route book, the vehicle sends a command to the drive unit before reaching the track separation node. The drive unit then drives the track selection actuator to complete the state switch, with a switch response time of no more than 0.3 seconds. After the position sensor confirms that the actuator is in position, the vehicle continues to travel according to the route book. When the vehicle passes through a node, the guide component at the lower end of the track selection actuator engages with the fixed guide structure of the track at the node. Because the shape of the guide structure matches the guide component, the vehicle is naturally guided into the target track specified in the route book. The entire track selection process is completed autonomously by the vehicle.
[0047] When the drive unit experiences a power outage, failure, or mechanical jamming, the flexible reset mechanism immediately resets the track selection actuator to the default safe state. In this state, the vehicle is guided off the main highway and onto a branch line or buffer section at the next station. After the vehicle goes offline, maintenance is arranged by the central dispatch center, without affecting the normal passage of vehicles on the main line. This fail-safe design ensures that a single-vehicle track selection failure will not trigger a chain reaction of accidents on the main line.
[0048] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.
Claims
1. A rail transit scheduling method based on four-dimensional spatiotemporal trajectory, characterized in that, Includes the following steps: S1. The central dispatch platform obtains a travel time and space resource application. The travel time and space resource application includes at least origin and destination information. The travel time and space resource application is generated by the central dispatch platform itself or comes from at least one of the following: track segment local controller, user terminal or vehicle. S2. The central dispatch platform searches for available four-dimensional spatiotemporal windows in the unallocated spatiotemporal right-of-way resources using an incremental matching method, based on the allocated spatiotemporal right-of-way resource occupancy table, to plan a four-dimensional spatiotemporal trajectory containing the entire time and spatial coordinates for the requesting vehicle. When further optimization of departure time or resource utilization efficiency is required, based on the allocated spatiotemporal right-of-way resources, the spatiotemporal trajectories of some online vehicles are locally adjusted to obtain a better spatiotemporal window, and a four-dimensional spatiotemporal trajectory is planned for the requesting vehicle. The four-dimensional spatiotemporal trajectory specifies the vehicle's spatial position, speed, allowable deviation threshold, track selection direction at each track splitting node, and passage time window at the track merging node at each moment. S3. The central dispatch platform sends the four-dimensional spatiotemporal trajectory to the vehicle and the local controller of each track segment the vehicle passes through. After receiving the four-dimensional spatiotemporal trajectory, the vehicle sends a confirmation to the central dispatch platform. After the confirmation is correct, the four-dimensional spatiotemporal trajectory is strictly followed by the vehicle. S4. The vehicle travels on the track according to the four-dimensional spatiotemporal trajectory. The on-board unit senses the vehicle's position and speed in real time through vehicle-road cooperative interaction, compares it with the theoretical spatiotemporal coordinates specified by the four-dimensional spatiotemporal trajectory, and controls the vehicle's movement to keep the deviation between the vehicle's actual spatiotemporal coordinates and the four-dimensional spatiotemporal trajectory within the allowable deviation threshold range. The allowable deviation threshold is determined by the central dispatching platform when planning the four-dimensional spatiotemporal trajectory based on the vehicle's travel speed at each point in time, track environment conditions, and the response speed of the vehicle and ground equipment, and is included in the four-dimensional spatiotemporal trajectory issued to the vehicle. S5. The local controller of each track section monitors the operating status data of all vehicles in the section in real time through the sensors deployed in the section, compares the actual operating status of the vehicles with the four-dimensional spatiotemporal trajectory, and when the deviation exceeds the allowable deviation threshold, it shall be handled in accordance with the local safety handling plan and reported to the central dispatch platform. S6. When the preset triggering conditions are met, the central dispatch platform generates and distributes the updated four-dimensional spatiotemporal trajectory to replace the vehicle's current four-dimensional spatiotemporal trajectory.
2. The rail transit scheduling method based on four-dimensional spatiotemporal trajectory according to claim 1, characterized in that, In step S2, the incremental matching method includes: the central dispatching platform maintains a network-wide spatiotemporal resource occupancy table indexed by track segments and time slices, transforming the four-dimensional spatiotemporal trajectory planning of newly added vehicles into an incremental matching problem of searching for available spatiotemporal windows segment by segment along the path from the starting point to the destination within the spatiotemporal resource gaps already occupied by online vehicles; when further optimization of departure time or resource utilization efficiency is required, the speed or time window of some online vehicles within the affected range is further fine-tuned locally to free up better spatiotemporal windows without affecting the normal operation of other online vehicles; the incremental matching only involves newly added vehicles and local vehicles affected by fine-tuning, rather than replanning all vehicles in the network, and its computational complexity increases nearly linearly with the number of online vehicles.
3. The rail transit scheduling method based on four-dimensional spatiotemporal trajectory according to claim 1, characterized in that, The preset triggering condition in step S6 includes at least one of the following: Triggered by fault handling, passenger destination change, network-wide empty vehicle balance allocation, emergency avoidance needs, optimization of time and space resource utilization efficiency, and route optimization. After the trigger causes the four-dimensional spatiotemporal trajectory to be updated, the central dispatch platform issues a new four-dimensional spatiotemporal trajectory. After receiving the new trajectory information and confirming it with the central dispatch platform, the vehicle executes according to the new trajectory.
4. The rail transit scheduling method based on four-dimensional spatiotemporal trajectory according to claim 1, characterized in that, The four-dimensional spatiotemporal trajectory planned in step S2 also includes the time window information of the merging node; Before merging into the main highway, the vehicle has adjusted its speed to match the main highway cruising speed and enters the main highway within the time window specified by the four-dimensional spatiotemporal trajectory. The width of the time window is not less than the duration required for the vehicle to pass through the merging point and there is a safety margin before and after, so that the vehicle does not interfere with the existing traffic flow on the main highway when merging into the main highway.
5. The rail transit scheduling method based on four-dimensional spatiotemporal trajectory according to claim 1, characterized in that, In step S4, before the vehicle reaches the track splitting node, the on-board track selection device is controlled to switch to the state corresponding to the target track in advance according to the track selection timing pre-defined in the four-dimensional spatiotemporal trajectory. The on-board track selection device completes the state switch before the vehicle reaches the track splitting node. The on-board track selection device engages with the track structure at the track splitting node and guides the vehicle into the target track.
6. The rail transit scheduling method based on four-dimensional spatiotemporal trajectory according to claim 5, characterized in that, When the on-board track selection device experiences a drive failure or mechanical jamming, the on-board track selection device automatically switches to a safe state through an elastic reset mechanism, allowing the vehicle to leave the high-speed mainline and enter the next station or buffer section.
7. The rail transit scheduling method based on four-dimensional spatiotemporal trajectory according to claim 1, characterized in that, The method employs a three-tier computing power sharing architecture: The central dispatch platform is responsible for global four-dimensional spatiotemporal trajectory planning and macro-dispatch decisions; The local controller for the track section is responsible for detecting the spatiotemporal position deviation of all vehicles in the section, interacting with adjacent sections, and controlling the facilities in the section. The onboard unit is responsible for real-time location, speed, self-status and environmental perception, execution of local control commands, and adjustment of vehicle motion.
8. A rail transit scheduling system based on four-dimensional spatiotemporal trajectory, used to implement the method according to any one of claims 1 to 7, characterized in that, include: The central dispatch platform, including computing clusters and storage devices, is used to store the spatiotemporal resource occupancy table of the entire network, perform incremental matching calculations, and plan, distribute, and update four-dimensional spatiotemporal trajectories. Multiple track segment local controllers are set up in each track segment. Each controller includes a processor, memory and communication interface. They are used to receive four-dimensional spatiotemporal trajectories, monitor the vehicle operation status of the segment through sensors, perform deviation detection and handle according to the plan. The on-board unit, installed on each vehicle, includes a vehicle-road cooperative communication module, a positioning module, environmental sensors, status sensors, and an execution controller. It is used to receive four-dimensional spatiotemporal trajectories, perceive the vehicle's position and speed in real time through vehicle-road cooperative interaction, and control the vehicle to travel along the four-dimensional spatiotemporal trajectory. The central dispatching platform is connected to the local controllers of each track section via a fiber optic backbone network; the local controllers of each track section are connected to the on-board units within their jurisdiction via a wireless redundant communication network.
9. The rail transit dispatching system based on four-dimensional spatiotemporal trajectory according to claim 8, characterized in that, The vehicle is also equipped with an on-board track selection device, which includes: The track selection actuator has at least two switchable engagement states, each corresponding to a different travel track direction; The drive unit is used to drive the track selection actuator to switch states before the vehicle reaches the track splitting node, based on the track selection timing command in the four-dimensional spatiotemporal trajectory received by the on-board unit. The flexible reset mechanism, connected to the track selection actuator, is configured to automatically switch the track selection actuator to a safe state that allows the vehicle to leave the main line in the event of a drive failure.
Citation Information
Patent Citations
Intelligent early warning system for closing of train and early warning method of intelligent early warning system
CN104691581A
