Port unmanned vehicle cooperative scheduling and path planning system and method

Abstract

The application discloses a kind of port unmanned vehicle coordination scheduling and path planning system and method, including cloud control platform and fleet management system FMS, it is characterized in that, cloud control platform includes algorithm module, path planning module, operation guiding module, instruction analysis module and high-precision map, the operation guiding module is connected with instruction analysis module, path planning module and algorithm module signal, the operation guiding module issues instruction / path to fleet management system FMS, the path planning module calls high-precision map and carries out path planning to vehicle, issues path and instruction information to fleet management system FMS, the fleet management system FMS again issues path and instruction information to unmanned vehicle, the unmanned vehicle executes the path.

Description

Technical Field

[0001] This invention relates to the field of unmanned vehicle cloud control platform technology, specifically a port unmanned vehicle collaborative scheduling and path planning system and method. Background Technology

[0002] In recent years, the widespread adoption of cloud computing and 5G communication technologies has enabled port autonomous driving systems to achieve multi-vehicle collaborative scheduling and global path optimization. Through cloud platforms, unmanned vehicles can share data in real time, dynamically adjust routes, and improve overall operational efficiency. However, existing cloud control platforms have the following drawbacks: there are many types of unmanned vehicles, such as driverless trucks and automated guided vehicles (AGVs). Because multiple vehicles operating simultaneously are prone to path conflicts or congestion, their collaborative capabilities are weak, and they lack global optimization capabilities. They cannot achieve optimal path planning for different types and models of unmanned vehicles, efficient intersection passage, or vehicle overtaking functions. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a collaborative scheduling and path planning system and method for unmanned vehicles in ports, which can solve the problems in the prior art.

[0004] This invention is achieved through the following technical solution: An unmanned vehicle cloud path collaboration platform of this invention includes a cloud control platform and a fleet management system (FMS). The cloud control platform includes an algorithm module, a path planning module, a task guidance module, an instruction parsing module, and a high-precision map. The task guidance module establishes a signal connection with the instruction parsing module, the path planning module, and the algorithm module. The task guidance module sends instructions / paths to the fleet management system (FMS). The path planning module calls the high-precision map and performs path planning for the vehicles, sending the path and instruction information to the fleet management system (FMS). The fleet management system (FMS) then sends the path and instruction information to the unmanned vehicle, which executes the path.

[0005] A further technical solution includes a path planning module comprising a lane-level planning module, a path generation module, and a scheduling algorithm module. The lane-level planning module is used for lane planning for the autonomous vehicle, the path generation module is used for generating a travel path for the autonomous vehicle, and the scheduling algorithm module is used for scheduling and controlling the autonomous vehicles.

[0006] A further technical solution includes an OPTIM module as the algorithm module. The OPTIM module calls the iECS algorithm and returns the algorithm result to the OPTIM module. The OPTIM module is used to send the instruction information of the cloud control platform to the job guidance module, and the job guidance module feeds back the execution status to the OPTIM module.

[0007] A control method for optimal path planning, orderly passage at intersections, and overtaking in container areas or berths using a cloud control platform is characterized by the following steps: First, the cloud control platform uses an instruction parsing module to convert the start and end point information in the instruction into logical point information for the start and end points of the unmanned vehicle's path. Combining this with the overall map information, vehicle kinematic characteristics, port traffic conditions, road closures, ship bow and stern information, and operation instruction priority information, the optimal path is planned in a high-precision map based on the road network structure diagram. A key point path is generated based on the optimal path, and the path and instruction information are sent to the fleet management system (FMS). The FMS then sends the path and instruction information to the unmanned vehicle, which executes the path.

[0008] The second step involves using the fleet management system (FMS) to provide real-time feedback on the status of autonomous vehicles. The cloud control platform acquires information such as the location, speed, and trajectory of each autonomous vehicle, calculates potential conflicts arising from vehicle interactions within the intersection, and calculates stopping points to plan the traffic flow order at the intersection.

[0009] The third step is that when the unmanned vehicle is in the container area or berth, the cloud control platform determines whether overtaking conditions are met based on the current operating conditions of the container area or berth, and dispatches unmanned vehicles that meet the conditions to overtake.

[0010] A further technical solution involves a path planning module that constructs a road network structure map based on lane data. The upper layer of the lanes is the lane segment, which is composed of lanes arranged side by side. The lanes within a lane segment are arranged sequentially from left to right. The reversible lane relationship between adjacent lanes is determined by the type of lane line. Dashed lines represent lane segments that can change lanes, while solid lines represent lane segments that cannot change lanes. The upper layer of the lane segment is the road, which is composed of lane segments arranged in sequence.

[0011] Further technical solutions include lane data such as its left and right boundaries, reference lines, unique ID identifier, and relationships with the lanes in front, behind, and adjacent lanes.

[0012] A further technical solution, the optimal path planning method is as follows: First, preprocess the sub-topology map according to port operation rules, road closure information, congestion status of each road segment, and instruction priority; then, based on vehicle performance constraints, use Dijkstra's algorithm to search for the shortest feasible path that meets the requirements in the road network structure map, i.e., the optimal path; Second, based on the optimal path, provide key points at fixed intervals during the lane-keeping phase; during non-lane-keeping phases such as intersections and lane changes, provide key points for the start and end of the behavior, and the path between key points is planned by the vehicle; Third, generate key point paths based on lane-level paths and scheduling results; Fourth, path segmentation and update, segment the optimal path to form short paths, and assemble special point types on the key paths.

[0013] Further technical solutions and intersection traffic flow scheduling methods are as follows: First, based on the future trajectory of a single vehicle uploaded by the fleet management system (FMS), the separating axis theorem is used to determine whether there is a path conflict between two vehicles. If there is no conflict, no right-of-way priority allocation is performed; if there is a conflict, priority allocation is performed. The strong priority judgment condition is determined based on the autonomous vehicle's speed and current position and distance from the intersection, etc., to determine that the vehicle cannot complete an effective stop before the intersection. Second, according to the right-of-way priority, stop points are issued to the lower-priority autonomous vehicles for parking control. After the path conflict is resolved, the stop points are released, and the autonomous vehicles can drive normally into the intersection.

[0014] A further technical solution involves the cloud control platform controlling vehicles to overtake in the container area. Upon reaching the container area entrance, the cloud control platform performs the following judgment process: First, filtering vehicles operating in the container area: vehicles already within the container area and whose target operating type is a container area road. Second, determining if overtaking conditions are met: whether there is sufficient space behind the vehicle's operating position to allow for a lane change, and whether there is sufficient space in front of the operating position to allow for a following distance. Third, determining if overtaking is successful; if successful, the vehicle performs a lane change maneuver. Fourth, after the vehicle performs the lane change maneuver, determining if overtaking is complete; if successful, the vehicle returns to the operating lane.

[0015] A further technical solution involves the cloud control platform controlling the vehicle to overtake in a parking space. When the unmanned vehicle arrives at the approach ramp, the cloud control platform performs the following judgment process: First, filtering vehicles operating in the parking space: vehicles already in the operating lane and whose target is that lane. Second, vehicle state transition, determining if overtaking conditions are met: whether there is sufficient space behind the vehicle's operating position to allow for a lane change, and whether there is sufficient space in front of the operating position to allow for a following distance. Third, determining if overtaking can be successful; if successful, the vehicle performs a lane change maneuver. Fourth, after the vehicle performs the lane change maneuver, determining if overtaking is complete; if successful, the vehicle returns to the operating lane.

[0016] The beneficial effects of this invention are as follows: First, the path planning module in the cloud control platform calls a high-precision map and plans paths for different types of vehicles, then sends the path and instruction information to the fleet management system (FMS). The FMS then sends the path and instruction information to the unmanned vehicles, which execute the path. Furthermore, the instruction parsing module converts the start and end point information in the instruction into logical point information of the start and end points of the unmanned vehicle's path. Combining the overall map information, vehicle kinematic characteristics, port traffic conditions, overall road closures and ship bow and stern information, and operation instruction priority information, the optimal path is planned in the high-precision map based on the road network structure diagram, thereby realizing the function of optimal vehicle scheduling using the cloud platform.

[0017] Second, the fleet management system (FMS) provides real-time feedback on the status of unmanned vehicles. The cloud control platform obtains information such as the location, speed, and trajectory of each unmanned vehicle, calculates potential conflict relationships formed by vehicle interactions within the intersection, and realizes the function of planning the passage order of vehicles at the intersection by calculating the stopping point. This effectively improves the passage efficiency of unmanned vehicles of different models at the intersection. Furthermore, by combining cloud-based planning and vehicle-side planning, efficient vehicle control can be achieved during both lane-keeping and non-lane-keeping phases.

[0018] Third, when the unmanned vehicle is in the container area or berth, the cloud control platform determines whether overtaking conditions are met based on the current operating vehicles in the container area or berth, and dispatches unmanned vehicles that meet the conditions to overtake, thereby improving the traffic efficiency of unmanned vehicles during normal operation and avoiding obstruction and stopping.

[0019] Fourth, by constructing a road network structure map, the path is divided into multiple lane segments, and dashed and solid lines are used to limit whether overtaking is allowed, so as to improve the order of autonomous vehicle travel and improve the efficiency of autonomous vehicle scheduling.

[0020] 5. When overtaking, pre-determine whether there is sufficient space behind the vehicle's working position to allow for lane changing distance, and whether there is sufficient space in front of the working position to allow for following distance, so as to trigger the overtaking command and avoid vehicle damage or stoppage caused by overtaking failure. Attached Figure Description

[0021] For ease of explanation, the present invention will be described in detail below with reference to specific embodiments and accompanying drawings.

[0022] Figure 1 This is a schematic diagram of a module of an unmanned vehicle cloud path collaboration platform according to the present invention;

[0023] Figure 2 This is a topology diagram of the lane segment;

[0024] Figure 3 This is a schematic diagram of the key points of the autonomous vehicle's path when it turns.

[0025] Figure 4 This is a schematic diagram of an autonomous vehicle encountering a path conflict.

[0026] Figure 5 This is a schematic diagram showing the positional relationship of the vehicle's operating positions. Detailed Implementation

[0027] like Figures 1-5 As shown, the present invention will be described in detail. For ease of description, the directions mentioned below are defined as follows: the directions of up, down, left, right, front, and back mentioned below are the same as... Figure 1The projection relationships are consistent in all directions (up, down, left, right, front, back). This invention provides a collaborative scheduling and path planning system and method for unmanned vehicles in ports. It includes a cloud control platform and a fleet management system (FMS). The cloud control platform is equipped with an algorithm module, a path planning module, a work guidance module, an instruction parsing module, and a high-precision map. The work guidance module establishes signal connections with the instruction parsing module, the path planning module, and the algorithm module, enabling the work guidance module to send instructions / paths to the fleet management system (FMS). The FMS then feeds back the execution status to the work guidance module. The path planning module calls the high-precision map and performs path planning for the vehicles, then feeds back the path planning results to the work guidance module.

[0028] Beneficially, the path planning module includes a lane-level planning module, a path generation module, and a scheduling algorithm module. The lane-level planning module is used to plan lanes for the autonomous vehicle, the path generation module is used to generate travel paths for the autonomous vehicle, and the scheduling algorithm module is used to schedule and control the autonomous vehicles.

[0029] Beneficially, the algorithm module includes the OPTIM module, which calls the iECS algorithm and returns the algorithm result to the OPTIM module. The OPTIM module is used to send the instruction information of the cloud control platform to the job guidance module, and the job guidance module feeds back the execution status to the OPTIM module.

[0030] Fleet Management System (FMS) is a comprehensive management system that integrates modern information technology, GPS positioning technology, big data analysis, and other advanced technologies. Through FMS, a series of functions such as real-time vehicle monitoring, dispatch management, maintenance, fuel consumption analysis, and driving behavior monitoring can be realized, comprehensively improving the level of intelligent vehicle management.

[0031] The OPTIM module is a parameter optimization library in PtTorch that enables the calling of the iECS algorithm.

[0032] The following is the workflow of the cloud control platform:

[0033] The first step involves the upstream system sending operation instructions to the cloud control platform. The system uses the instruction parsing module to convert the start and end point information in the instructions into logical point information for the start and end points of the unmanned vehicle's path. Combining this with information such as the overall map, vehicle kinematics characteristics, port traffic conditions, road closures, ship bow and stern information, and operation instruction priority, the system plans the optimal path in the high-precision map based on the road network structure. Based on the optimal path, a key point path is generated, and the path and instruction information are sent to the fleet management system (FMS). The FMS then sends the path and instruction information to the unmanned vehicle, which executes the path.

[0034] The second step involves real-time feedback of the unmanned vehicle status through the fleet management system (FMS). The cloud control platform can obtain information such as the location, speed, and trajectory of each unmanned truck, calculate potential conflict relationships formed by vehicle interactions within the intersection, and realize the function of planning the vehicle passage order at the intersection by calculating the stopping point, thereby improving the smoothness of traffic flow at the intersection.

[0035] The third step involves the cloud control platform determining whether overtaking is possible based on the current status of vehicles in the container area or parking space. If the conditions are met, the platform will dispatch overtaking vehicles that meet the requirements, thereby improving operational efficiency and reducing queuing time.

[0036] Beneficially, the road network structure is constructed as follows: the basic unit of the high-precision map is the lane, and the lane data includes its left boundary, right boundary, reference line, unique ID identifier, and the relationship with the preceding and following lanes and adjacent lanes; for example... Figure 2 As shown, the path planning module can use these relationships to construct a road network structure diagram (also known as a topology diagram); the upper layer of the lanes is the lane segment, which is composed of lanes side by side. The lanes in the lane segment are arranged in order from left to right. The lane change relationship between adjacent lanes is determined by the type of lane line. Dashed lines represent lane segments that can change lanes, and solid lines represent lane segments that cannot change lanes. The upper layer of the lane segment is the road, which is composed of lane segments.

[0037] Beneficially, the optimal path planning method is as follows: First, the sub-topology map is preprocessed based on port operation rules (e.g., designated access ramps, non-passable container areas, etc.), road closure information, congestion status of each road segment, and instruction priority. Then, based on vehicle performance constraints, Dijkstra's algorithm is used to search for the shortest feasible path (i.e., the optimal path) in the road network structure map. Second, based on the optimal path, key points are provided at fixed intervals during lane-keeping phases; during non-lane-keeping phases such as intersections and lane changes, key points for the start and end of the behavior are provided, and the path between key points is planned by the vehicle. Third, key point paths are generated based on lane-level paths and scheduling results. Fourth, path segmentation and updating are performed, dividing the optimal path into shorter paths of 200m in length, and assembling special point types on the key paths to cope with real-time traffic and environmental changes, such as... Figure 3 As shown, the cloud control platform plans the key point path and the passable distance, the vehicle plans the local trajectory in the turning area, and the vehicle travels according to the key point path.

[0038] Beneficially, among which, such as Figure 3 As shown, the arc-shaped trajectory is a local trajectory planned by the vehicle, and each point represents the key point path and travel distance planned by the cloud platform.

[0039] Beneficially, the geographic coordinate system of the road network structure map uses WGS84, the location (x, y) sent to the vehicle end uses global UTM coordinates, and the direction (heading) is represented by the angle between the vehicle's heading and true north.

[0040] Beneficially, the coordinates of the IGV (Intelligent Guided Vehicle) are defined at the center point of the vehicle body, while the coordinates of the container truck are defined at the center point of the host vehicle.

[0041] Beneficially, the key point path refers to providing multiple key points for the lane at fixed intervals during the lane keeping phase, and providing key points for the start and end of vehicle behavior during non-lane keeping phases such as intersections and lane changes. The path between the key points is planned automatically by the vehicle.

[0042] Beneficially, the traffic flow scheduling method at intersections is as follows: First, based on the future trajectory of each vehicle uploaded by the fleet management system (FMS), the Separating Axis Theorem (SAT) is used to determine if there is a path conflict between two vehicles. If there is no conflict, no right-of-way priority allocation is performed; if there is a conflict, priority allocation is performed. The strong priority determination condition is based on the autonomous vehicle's speed and current position relative to the intersection, determining that the vehicle cannot effectively stop before the intersection. Second, based on the right-of-way priority, stop points are issued to lower-priority autonomous vehicles for parking control. After the path conflict is resolved, the stop points are released, and the autonomous vehicles proceed normally into the intersection. Figure 4 As shown, turning vehicles have low priority, while straight-line vehicles have high priority. Turning vehicles stop at the stopping point.

[0043] Beneficially, among which, such as Figure 4 As shown, the circle in the middle represents the stopping point, the vertical square represents vehicles going straight, and the horizontal square represents vehicles turning.

[0044] Beneficially, weak priority refers to the priority of work instructions and the waiting time at intersections.

[0045] Beneficially, among which, such as Figure 5As shown, when a vehicle controls another vehicle to overtake in the container area, upon reaching the container area entrance, the cloud control platform performs the following judgment process: First, filter vehicles operating in the container area: vehicles within the container area whose target operation type is container area road. Second, determine if overtaking conditions are met: whether there is sufficient space behind the vehicle's operating position to allow for a lane change, and whether there is sufficient space in front of the operating position to allow for a following distance. Third, determine if overtaking is successful; if successful, the vehicle performs a lane change. Fourth, after the vehicle performs the lane change, determine if overtaking is complete; if completed, the vehicle returns to the operating lane. Fifth, the cloud control platform assembles a scheduling task based on the vehicle's status: Normal driving: returns to "Normal driving" vehicle status; Waiting to overtake: returns to the calculated overtaking closure area and "Waiting to overtake" vehicle status; Overtaking in progress: returns to the overtaking closure area and "Overtaking in progress" vehicle status; Being overtaken: returns to "Being overtaken" vehicle status; Overtaking completed: returns to "Overtaking completed" vehicle status.

[0046] Beneficially, among which, such as Figure 5 As shown, when the cloud control platform controls the vehicle to overtake in a parking space, upon the arrival of the unmanned vehicle at the approach bridge, the cloud control platform performs the following judgment process: First, filtering vehicles operating in the parking space: vehicles in the operating lane whose target is that lane. Second, vehicle state transition, determining if overtaking conditions are met: whether there is sufficient space behind the vehicle's operating position to allow for a lane change distance, and whether there is sufficient space in front of the operating position to allow for a following distance. Third, determining if overtaking can be successful; if successful, the vehicle performs a lane change maneuver. Fourth, after the vehicle performs the lane change maneuver, determining if overtaking is complete; if completed, the vehicle returns to the operating lane. Fifth, the cloud control platform assembles and schedules tasks based on the vehicle's status: Normal driving: returns to "Normal driving" vehicle status; Waiting to overtake: returns to the calculated overtaking closure area and "Waiting to overtake" vehicle status; Overtaking in progress: returns to the overtaking closure area and "Overtaking in progress" vehicle status; Being overtaken: returns to "Being overtaken" vehicle status; Overtaking completed: returns to "Overtaking completed" vehicle status.

[0047] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions conceived without creative effort should be included within the scope of protection of the present invention; therefore, the scope of protection of the present invention should be determined by the scope defined in the claims.

Claims

1. An unmanned vehicle cloud path collaboration platform, comprising a cloud control platform and a fleet management system (FMS), characterized in that, The cloud control platform includes an algorithm module, a path planning module, a job guidance module, an instruction parsing module, and a high-precision map. The job guidance module establishes a signal connection with the instruction parsing module, the path planning module, and the algorithm module. The job guidance module sends instructions or paths to the fleet management system (FMS). The path planning module calls the high-precision map and plans paths for the vehicles, then sends the path and instruction information to the FMS. The FMS then sends the path and instruction information to the unmanned vehicle, which executes the path. The cloud control platform's control methods for optimal path planning, orderly passage at intersections, and overtaking in container areas or berths are as follows: First, the cloud control platform uses the instruction parsing module to convert the start and end point information in the instructions into logical point information for the start and end points of the unmanned vehicle's path. Combining this with overall map information, vehicle kinematic characteristics, port traffic conditions, road closures, ship bow and stern information, and operational instruction priority information, the platform plans the optimal path based on the road network structure diagram in a high-precision map. Based on this optimal path, it generates key point paths and sends the path and instruction information to the fleet management system (FMS). The fleet management system (FMS) then sends the path and instruction information to the unmanned vehicles (UAVs), which then execute the path. The second step involves the FMS providing real-time feedback on the UAV status. The cloud control platform acquires the location, speed, and trajectory information of each UAV, calculates potential conflicts arising from vehicle-to-vehicle interactions at intersections, and calculates stopping points to plan the vehicle passage order at intersections. The third step involves the cloud control platform determining whether overtaking is possible based on the current vehicle situation in the container area or parking space, and dispatching UAVs that meet the conditions for overtaking. When a vehicle controls another vehicle to overtake in the container area, upon reaching the container area entrance, the cloud control platform performs the following judgment process: First, filter vehicles operating in the container area: vehicles within the container area whose target type is the container area road; Second, determine if overtaking conditions are met: whether there is sufficient space behind the vehicle's operating position to allow for lane changing distance, and whether there is sufficient space in front of the operating position to allow for following distance; Third, determine if overtaking can be successful. If overtaking is successful, the vehicle performs a lane change maneuver; Fourth, after the vehicle performs the lane change maneuver, determine if overtaking is completed. If overtaking is completed, the vehicle returns to the operating lane. When the cloud control platform controls the vehicle to overtake in the parking space, when the unmanned vehicle arrives at the approach bridge, the cloud control platform performs the following judgment process: First, filtering of parking space operation vehicles: vehicles that are in the operation lane and whose operation target is that lane. The second step is to transfer the vehicle's status and determine whether the overtaking conditions are met: whether there is sufficient space behind the vehicle's working position to allow for a lane change distance, and whether there is sufficient space in front of the working position to allow for a following distance; the third step is to determine whether overtaking can be successful. If overtaking can be successful, the vehicle will perform a lane change maneuver; the fourth step is to determine whether overtaking is completed after the vehicle performs a lane change maneuver. If overtaking is completed, the vehicle will return to the working lane.

2. The unmanned vehicle cloud path collaboration platform according to claim 1, characterized in that: The path planning module includes a lane-level planning module, a path generation module, and a scheduling algorithm module. The lane-level planning module is used to plan lanes for the autonomous vehicle, the path generation module is used to generate travel paths for the autonomous vehicle, and the scheduling algorithm module is used to schedule and control the autonomous vehicles.

3. The unmanned vehicle cloud path collaboration platform according to claim 2, characterized in that: The algorithm module includes the OPTIM module, which calls the iECS algorithm and returns the algorithm result to the OPTIM module. The OPTIM module is used to send the instruction information of the cloud control platform to the job guidance module, and the job guidance module feeds back the execution status to the OPTIM module.

4. The unmanned vehicle cloud path coordination platform according to claim 1, and the control method for optimal path planning, orderly passage at intersections, and overtaking in parking areas or spaces for unmanned vehicles, are characterized in that: The route planning module constructs a road network structure map based on lane data. The upper layer of the lanes is the lane segment, which is composed of lanes side by side. The lanes in the lane segment are arranged in order from left to right. The change lane relationship between adjacent lanes is determined by the type of lane line. Dashed lines represent lane segments that can change lanes, and solid lines represent lane segments that cannot change lanes. The upper layer of the lane segment is the road, which is composed of lane segments.

5. The unmanned vehicle cloud path collaboration platform according to claim 4, characterized in that: Lane data includes its left and right boundaries, reference lines, unique ID identifier, and its relationship with the lanes in front, behind, and adjacent lanes.

6. The unmanned vehicle cloud path collaboration platform according to claim 5, characterized in that: The optimal path planning method is as follows: First, the sub-topology map is preprocessed according to port operation rules, road closure information, congestion status of each road segment, and command priority. Then, based on vehicle performance constraints, Dijkstra's algorithm is used to search for the shortest feasible path that meets the requirements in the road network structure map, i.e., the optimal path. Second, based on the optimal path, key points are given at fixed intervals during the lane-keeping phase. At intersections and lane-changing non-lane-keeping phases, key points for the start and end of the behavior are given, and the path between key points is planned by the vehicle. Third, key point paths are generated based on lane-level paths and scheduling results. The fourth step is path segmentation and update, which involves segmenting the optimal path to form shorter paths and assembling special point types on the critical path.

7. The unmanned vehicle cloud path collaboration platform according to claim 1, characterized in that: The traffic flow scheduling method at the intersection is as follows: First, based on the future trajectory of a single vehicle uploaded by the fleet management system (FMS), the separating axis theorem is used to determine whether there is a path conflict between two vehicles. If there is no conflict, no right-of-way priority allocation is performed. If there is a conflict, priority allocation is performed. The strong priority judgment condition is determined based on the autonomous vehicle's speed and the distance between its current position and the intersection, which indicates that the vehicle cannot complete an effective stop before the intersection. Second, based on the right-of-way priority, stop points are issued to the lower-priority autonomous vehicles for parking control. After the path conflict is resolved, the stop points are released, and the autonomous vehicles can drive normally into the intersection.

Images