Port automatic driving truck cooperative path planning method based on space-time feature sampling

Abstract

The application relates to a port automatic driving truck cooperative path planning method based on space-time feature sampling, which comprises the following steps: uploading real-time vehicle motion state information to a port scheduling system by an automatic driving truck; issuing task information to the automatic driving truck by the scheduling system; starting from a parking lot at a set interval, reaching a specified import berth, loading a container, and recording a current time by the automatic driving truck; establishing a cooperative path planning mathematical model according to the task information and the current time; solving the cooperative path planning model by using an optimization algorithm, outputting space-time sampling point information, and issuing the space-time sampling point information to a vehicle control system of the automatic driving truck; executing after the vehicle trajectory is optimized; reaching a target yard by the automatic driving truck, waiting for work equipment to complete unloading, and completing a scheduling task. Compared with the prior art, the application can cooperatively optimize the path and speed of the automatic driving truck, avoid conflicts at intersection nodes, and improve the freedom degree of the bottom layer planning control of the automatic driving truck.

Description

Technical Field

[0001] This invention relates to the field of cooperative control technology for autonomous trucks in port areas, and in particular to a cooperative path planning method for autonomous trucks in port areas based on spatiotemporal feature sampling. Background Technology

[0002] Currently, with the rapid development of artificial intelligence and information technology, autonomous driving has become a major trend in the transportation sector. Port autonomous driving is a typical application of autonomous driving technology in closed environments and low-speed operations, which is beneficial for improving port efficiency, safety, and energy conservation. Port logistics transportation is one of the important application scenarios. Because autonomous trucks integrate high-precision positioning inertial navigation, LiDAR, and other technologies, they can achieve autonomous environmental positioning, autonomous perception, and intelligent control, providing intelligent logistics solutions for ports. Therefore, the application of autonomous trucks in port operations has attracted widespread attention. Technically, port autonomous driving needs to solve complex problems such as scene perception, path planning, and decision-making control, while ensuring high-precision positioning and parking effects, as well as high reliability and safety. It can be said that the precise collaborative control and collaborative transportation of autonomous trucks in port areas provides an opportunity to further reduce port operating costs, improve operational efficiency, and ensure green driving.

[0003] However, existing research on autonomous trucks in port areas still has the following shortcomings:

[0004] 1. Existing research overlooks the controllability advantage of autonomous trucks, namely, the ability to optimize routes and speeds online in real time and collaboratively. Traditional autonomous trucks only consider route optimization and assume average speed, which leads to several shortcomings. First, it fails to utilize the adjustable speed capability of autonomous trucks, reducing the optimization space for truck control and making it difficult to match the actual arrival time of some trucks at the work point with the planned time, thus making scheduling strategies difficult to execute or ineffective. Second, because existing research assumes a constant speed, trucks can only avoid conflicts at intersections by adjusting routes, increasing vehicle travel distance, reducing transportation efficiency, and increasing energy consumption.

[0005] 2. Existing autonomous driving truck path planning focuses on all node information of the road network, which requires real-time updates of the spatiotemporal resource occupancy status of vehicles at all nodes of the road network, resulting in increased computational load. Furthermore, the lack of consideration for key nodes limits the degree of freedom of autonomous driving planning and control, affecting the optimization effect. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the existing technology by providing a collaborative path planning method for autonomous trucks in port areas based on spatiotemporal feature sampling. This method can collaboratively optimize the path and speed of autonomous trucks, avoid conflicts at intersections, and improve the degree of freedom of the underlying planning and control of autonomous trucks.

[0007] The objective of this invention can be achieved through the following technical solution: a cooperative path planning method for automated guided vehicles (AGVs) in port areas based on spatiotemporal feature sampling, comprising the following steps:

[0008] S1. The autonomous driving truck will upload the real-time vehicle motion status information to the port area dispatch system.

[0009] S2. The port area dispatch system sends task information to the autonomous driving trucks;

[0010] S3. Based on the received task information, the autonomous truck departs from the parking lot at set intervals, arrives at the designated import berth, loads the container, and records the current time.

[0011] S4. Based on the task information and current time of the autonomous trucks in the port area, establish a mathematical model for collaborative path planning;

[0012] S5. Solve the collaborative path planning model using optimization algorithms and output spatiotemporal sampling point information;

[0013] S6. Send the spatiotemporal sampling point information to the vehicle control system of the autonomous truck, optimize and generate feasible vehicle trajectories and execute them;

[0014] S7. The autonomous driving truck arrives at the target yard and waits for the operating equipment to finish unloading, thus completing one dispatching task.

[0015] Furthermore, in step S1, the autonomous driving truck is equipped with multiple sensors for collecting vehicle motion state information, including but not limited to the vehicle's position coordinates in the road coordinate system, the current road segment, speed, and acceleration.

[0016] Furthermore, the task information issued by the port area scheduling system in step S2 includes, but is not limited to, the container task number, the target yard location, and the planned arrival time at the target yard.

[0017] Furthermore, step S4 specifically includes the following steps:

[0018] S41. Define parameters and decision variables, where the parameters include: V is the set of autonomous driving trucks, N is the set of nodes in the port area road network, and n... start，k and n end，k These are the import berth node number and the target yard node number of the kth container truck, respectively.start，k and s end，k These are the arrival time of the k-th container truck at the import berth and its planned arrival time at the target yard, respectively, d ij t represents the distance between the road segments (i, j) formed by nodes i and j. ijk This represents the time taken for the k-th truck to pass through road segment (i, j);

[0019] Decision variables include path decision variables, average speed decision variables, and arrival time decision variables.

[0020] S42. Taking the arrival at the target yard on schedule as the objective and considering constraints such as road speed limits and avoiding conflicts between trucks, establish a collaborative path planning model for autonomous trucks in the port area. By optimizing the path and average speed of the autonomous trucks on the road, the overall collaborative path planning of the trucks is completed.

[0021] Furthermore, the decision variables in step S41 are specifically:

[0022] x ijk Indicate whether the k-th truck travels directly from node i to node j; if so, then x ijk =1, otherwise x ijk =0;

[0023] v ijk Let v represent the average speed of the k-th truck traversing segment (i, j). If the truck's route does not include segment (i, j), then v ijk =0;

[0024] s ik This represents the time when the k-th truck arrives at node i. If the truck's path does not include node i, then s ik =0.

[0025] Furthermore, the objective function of the cooperative path planning model in step S42 is specifically:

[0026]

[0027] Where |V| is the total number of trucks, the first term in the objective function represents the difference between the planned arrival time and the optimized arrival time at the target yard, the second term represents the shortest travel time, α and β represent the weights, the third term makes the arrival time of vehicles that have not passed through node i at node i 0, and the fourth term makes the speed of vehicles that have not passed through road segment (i,j) on road segment (i,j) 0, thus unifying with the definition of the decision variables.

[0028] Furthermore, the constraints of the collaborative path planning model in step S42 include: path continuity constraints and decision variable x. ijkThe constraints include 0-1 integer constraints, road segment speed limit constraints, vehicle departure time constraints at the entrance berth, vehicle arrival times at nodes i and j, and path decision variables x. ijk Constraints and decision variables between them ik The constraints include the value of the given information, the time difference of arrival of any two vehicles k1 and k2 at the same node i, and the time difference of arrival of the two vehicles at the node.

[0029] Furthermore, the constraints of the cooperative path planning model in step S42 are specifically as follows:

[0030]

[0031]

[0032]

[0033]

[0034]

[0035]

[0036] s ik =s start，k i = n start，k

[0037]

[0038] s lower_bound ≤s ik ≤s upper_bound

[0039]

[0040]

[0041] The first three constraints are used to ensure the continuity of the path, that is, a continuous path from the starting point to the ending point, where each node is visited at most once.

[0042] The fourth constraint is the decision variable x. ijk 0-1 integer condition;

[0043] The fifth constraint is used to ensure that the speed meets the speed limit requirements of the road segment;

[0044] The sixth constraint represents the relationship between travel time and speed on a road segment;

[0045] The seventh constraint is the vehicle's departure time at the imported berth, which is known information;

[0046] The eighth constraint establishes the arrival times of vehicles at nodes i and j, as well as the path decision variable x. ijk The connection between them;

[0047] The ninth constraint is used to limit the decision variable s ik The range of values ​​for ;

[0048] The tenth constraint represents the time difference between the arrival of any two vehicles k1 and k2 at the same node i;

[0049] The eleventh constraint, to avoid conflict between the two vehicles at the node, is the arrival time difference between the two vehicles. It needs to be greater than or equal to the preset safety threshold δ T .

[0050] Furthermore, step S5 specifically includes the following steps:

[0051] S51. After linearizing the optimization problem, solve the optimization problem and output the optimized path of the autonomous truck and the average speed of each road segment.

[0052] S52. Based on the solution results for each autonomous driving truck, determine the key nodes in the path that need to avoid conflicts and the corresponding vehicle arrival times, and output the corresponding spatiotemporal sampling point information.

[0053] Furthermore, step S51 specifically involves using a mixed-integer nonlinear programming solver to solve the optimization problem.

[0054] Compared with the prior art, the present invention has the following advantages:

[0055] I. This invention acquires vehicle motion state information and combines it with the task information of autonomous trucks to establish a collaborative path planning mathematical model. By solving this model, the path and speed of the autonomous truck can be collaboratively optimized, improving the optimality of the vehicle's driving process. This overcomes the limitations of traditional autonomous trucks that assume average speed and only optimize the vehicle path. Furthermore, it can identify key nodes in the path that need to avoid conflicts and their corresponding vehicle arrival times. By outputting corresponding spatiotemporal feature sampling information, it effectively improves the underlying planning and control freedom of the autonomous truck, fully leveraging the advantages of precise execution in autonomous driving, improving the optimality of the driving process, and reliably solving the problem of excessive restrictions on vehicle planning and control.

[0056] Second, in this invention, when establishing a collaborative path planning mathematical model for autonomous trucks, relevant parameters and decision variables are first defined. Then, the goal is to reach the target yard on schedule. Constraints such as road speed limits and avoiding conflicts between trucks are fully considered. That is, the time difference between vehicles arriving at intersection nodes is considered to avoid vehicle conflicts due to excessively small time differences, thereby ensuring the safety of vehicle operation and improving overall transportation efficiency. Attached Figure Description

[0057] Figure 1 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0058] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0059] Example

[0060] like Figure 1 As shown, a collaborative path planning method for automated driving trucks in port areas based on spatiotemporal feature sampling includes the following steps:

[0061] S1. The autonomous driving truck will upload the real-time vehicle motion status information to the port area dispatch system.

[0062] S2. The port area dispatch system sends task information to the autonomous driving trucks;

[0063] S3. Based on the received task information, the autonomous truck departs from the parking lot at set intervals, arrives at the designated import berth, loads the container, and records the current time.

[0064] S4. Based on the task information and current time of the autonomous trucks in the port area, establish a mathematical model for collaborative path planning;

[0065] S5. Solve the collaborative path planning model using optimization algorithms and output spatiotemporal sampling point information;

[0066] S6. Send the spatiotemporal sampling point information to the vehicle control system of the autonomous truck, optimize and generate feasible vehicle trajectories and execute them;

[0067] S7. The autonomous driving truck arrives at the target yard and waits for the operating equipment to finish unloading, thus completing one dispatching task.

[0068] In practical applications, the port dispatching system will also obtain the time when the autonomous trucks arrive at the target yard based on the trucks' execution results, as information input for the next dispatch.

[0069] This embodiment applies the above technical solution, and its main contents include:

[0070] Step 1: The autonomous truck is equipped with a variety of sensors, which can acquire real-time vehicle motion status information, including: the vehicle's position coordinates in the road coordinate system, the current road segment, speed, acceleration and other longitudinal information, and upload it to the port area dispatch system;

[0071] Step 2: The port dispatch system sends task information such as container task number, target yard location, and planned arrival time at the target yard to the autonomous driving truck.

[0072] Step 3: The autonomous truck departs from the parking lot at set intervals, arrives at the designated import berth, loads the container, and records the current time;

[0073] Step 4: Based on the task information and current time of the autonomous trucks in the port area, establish a mathematical model for cooperative path planning, including the following steps:

[0074] Step 4.1: Define parameters and decision variables.

[0075] Step 4.1.1: The main parameters include: V is the set of autonomous driving trucks, N is the set of nodes in the port area road network, and n start，k and n end，k These are the import berth node number and the target yard node number of the kth container truck, respectively. start，k and s end，k These are the arrival time of the k-th container truck at the import berth and its planned arrival time at the target yard, respectively, d ij t represents the distance between the road segments (i, j) formed by nodes i and j. ijk This represents the time taken for the k-th truck to pass through road segment (i, j);

[0076] Step 4.1.2: Decision variables include: x ijk Indicate whether the k-th truck travels directly from node i to node j; if so, then x ijk =1, otherwise x ijk =0; v ijk Let v represent the average speed of the k-th truck traversing segment (i, j). If the truck's route does not include segment (i, j), then v ijk =0; s ik This represents the time when the k-th truck arrives at node i. If the truck's path does not include node i, then s ik =0.

[0077] Step 4.2: Establish a cooperative path planning model for automated driving trucks in the port area. Taking the goal of arriving at the target yard on schedule as the objective, and considering constraints such as road speed limits and avoiding conflicts between trucks, the overall cooperative path planning for the trucks is completed by optimizing the routes and average speeds of the automated driving trucks. This includes the following steps:

[0078] Step 4.2.1: The specific form of the objective function is:

[0079]

[0080] Where |V| is the total number of trucks, the first term represents the difference between the planned arrival time and the optimized arrival time at the target yard, the second term represents the shortest travel time, α and β represent the weights, the third term makes the arrival time of vehicles that have not passed through node i 0, and the fourth term makes the speed of vehicles that have not passed through road segment (i,j) 0 on road segment (i,j), thus unifying with the definition of decision variables.

[0081] Step 4.2.2: The specific constraints are as follows:

[0082]

[0083]

[0084]

[0085]

[0086]

[0087]

[0088] s ik =s start，k i = n start，k

[0089]

[0090] s lower_bound ≤s ik ≤s upper_bound

[0091]

[0092]

[0093] The first three constraints ensure the continuity of the path, meaning it's a continuous path from the starting point to the ending point, with each node visited at most once. The fourth is the decision variable x. ijk The first constraint is a 0-1 integer condition. The second constraint ensures that the speed meets the speed limit requirements of the road segment. The third constraint represents the relationship between the travel time and speed of the road segment. The fourth constraint is the vehicle's departure time at the entrance berth, which is known information. The fifth constraint establishes the vehicle's arrival time at nodes i and j, as well as the path decision variable x. ijk The relationship between them, the ninth constraint restricts the decision variable sik The range of values ​​for is given. The tenth constraint represents the arrival time difference between any two vehicles k1 and k2 at the same node i. The eleventh constraint requires the arrival time difference to meet a certain safety threshold to avoid conflicts between the two vehicles at the node.

[0094] Step 5: Solve the cooperative path planning model for autonomous trucks using optimization algorithms, and output spatiotemporal sampling points, including the following steps:

[0095] Step 5.1: After linearizing the optimization problem, use a mixed-integer nonlinear programming solver to solve the problem and output the optimized path of the autonomous truck and the average speed of each road segment.

[0096] Step 5.2: Based on the solution results of each autonomous driving truck, determine the key nodes in the path that need to avoid conflicts and the corresponding vehicle arrival times, and output the spatiotemporal sampling point information.

[0097] Step 6: The port area dispatch system sends the spatiotemporal sampling information to the vehicle control system of the autonomous truck, optimizes and generates feasible vehicle trajectories, and executes them;

[0098] Step 7: Based on the truck's execution results, the port dispatch system obtains the time when the autonomous truck arrives at the target yard, and then waits for the operating equipment to complete unloading, thus completing one dispatch task.

[0099] In summary, this technical solution proposes a collaborative path planning method for automated container trucks in port areas based on spatiotemporal feature sampling, utilizing vehicle scheduling models, path optimization, and mixed-integer nonlinear programming. Based on the origin and destination information of the transportation task and the planned arrival time, and taking key transit points in port operations as the target, this method establishes a collaborative path planning scheme by considering the spatiotemporal resource occupancy and conflict characteristics of these transit points and focusing on the path speed control of the automated container trucks. This solution supports container task allocation in existing TOS scheduling systems and provides reliable technical support for speed guidance and control of automated container trucks in port areas.

[0100] This solution, based on autonomous driving technology, combines container task allocation information from existing terminal operating systems with the truck's own location, speed, and acceleration information. Through collaborative optimization of the autonomous truck's path and speed, it reduces travel time and improves overall transportation efficiency. Simultaneously, it considers the vehicle's arrival time at intersection nodes, effectively avoiding vehicle conflicts. Furthermore, the optimization results focus only on potential conflict nodes. By outputting corresponding spatiotemporal feature sampling information, it expands the underlying planning and control freedom of the autonomous truck, fully leveraging the advantages of precise execution in autonomous driving.

Claims

1. A method for cooperative path planning of automated guided vehicles (AGVs) in port areas based on spatiotemporal feature sampling, characterized in that, Includes the following steps: S1. The autonomous driving truck will upload the real-time vehicle motion status information to the port area dispatch system. S2. The port area dispatch system sends task information to the autonomous driving trucks; S3. Based on the received task information, the autonomous truck departs from the parking lot at set intervals, arrives at the designated import berth, loads the container, and records the current time. S4. Based on the task information and current time of the autonomous trucks in the port area, establish a mathematical model for collaborative path planning; Step S4 specifically includes the following steps: S41. Define parameters and decision variables, where the parameters include: It is a collection of autonomous driving trucks. It is a collection of nodes in the port area's road network. and They are the first Import berth node number and target yard node number for container trucks. and They are the first The arrival time of the container trucks at the import berth and the planned arrival time at the target storage yard. Represents a node and nodes The formed road section The distance between them Indicates the first The section of road passed by the container truck Time; Decision variables include path decision variables, average speed decision variables, and arrival time decision variables. S42. Taking the arrival at the target yard on schedule as the objective, and considering the constraints of road speed limits and avoiding conflicts between trucks, establish a cooperative path planning model for autonomous trucks in the port area. By optimizing the path and average speed of the autonomous trucks on the road, the overall cooperative path planning of the trucks is completed. The decision variables in step S41 are as follows: Indicates the first Are trucks directly from the node? Reaching the node If so, then ,otherwise ; Indicates the first The section of road passed by the container truck The average speed, if the truck's route does not include road segments. ,but ; Indicates the first The trucks arrived at the node If the truck's path does not contain the node... , ; The objective function of the collaborative path planning model in step S42 is as follows: in, This represents the total number of container trucks. The first term in the objective function represents the difference between the planned arrival time and the optimized arrival time at the target yard, while the second term represents minimizing the travel time. and The first two terms represent the weights, and the third term makes the node not visited... The vehicle arrived at the node The time is 0, and the fourth term makes the road segment not passed through. Vehicles on the road section The velocity on the surface is 0, thus aligning with the definition of the decision variable; The constraints of the collaborative path planning model in step S42 include: path continuity constraints and decision variables. 0-1 integer constraints, road segment speed limit constraints, vehicle departure time constraints at the entrance parking space, and vehicle arrival node constraints. and nodes Time and path decision variables Constraints and decision variables between Value constraints, any two cars At the same node Arrival time difference constraint at the node; arrival time difference constraint between the two vehicles at the node. S5. Solve the collaborative path planning model using optimization algorithms and output spatiotemporal sampling point information; S6. Send the spatiotemporal sampling point information to the vehicle control system of the autonomous truck, optimize and generate feasible vehicle trajectories and execute them; S7. The autonomous driving truck arrives at the target yard and waits for the operating equipment to finish unloading, thus completing one dispatching task.

2. The method for cooperative path planning of automated guided vehicles in port areas based on spatiotemporal feature sampling as described in claim 1, characterized in that, In step S1, the autonomous driving truck is equipped with multiple sensors for collecting vehicle motion state information, including but not limited to the vehicle's position coordinates in the road coordinate system, the current road segment, speed, and acceleration.

3. The method for cooperative path planning of automated guided vehicles in port areas based on spatiotemporal feature sampling according to claim 1, characterized in that, In step S2, the task information issued by the port area dispatch system includes, but is not limited to, the container task number, the target yard location, and the planned arrival time at the target yard.

4. The method for cooperative path planning of automated guided vehicles in port areas based on spatiotemporal feature sampling according to claim 1, characterized in that, The specific constraints of the cooperative path planning model in step S42 are as follows: , The first three constraints are used to ensure the continuity of the path, that is, a continuous path from the starting point to the ending point, where each node is visited at most once. The fourth constraint is the decision variable. 0-1 integer condition; The fifth constraint is used to ensure that the speed meets the speed limit requirements of the road segment; The sixth constraint represents the relationship between travel time and speed on a road segment; The seventh constraint is the vehicle's departure time at the imported berth, which is known information; The eighth constraint establishes the vehicle arrival node. and nodes Time and path decision variables The connection between them; The ninth constraint is used to limit decision variables. The range of values ​​for ; The tenth constraint represents any two vehicles At the same node Time difference of arrival; The eleventh constraint, to avoid conflict between the two vehicles at the node, is the arrival time difference between the two vehicles. It needs to be greater than or equal to the preset safety threshold. .

5. The method for cooperative path planning of automated guided vehicles in port areas based on spatiotemporal feature sampling according to claim 1, characterized in that, Step S5 specifically includes the following steps: S51. After linearizing the optimization problem, solve the optimization problem and output the optimized path of the autonomous truck and the average speed of each road segment. S52. Based on the solution results for each autonomous driving truck, determine the key nodes in the path that need to avoid conflicts and the corresponding vehicle arrival times, and output the corresponding spatiotemporal sampling point information.

6. The method for cooperative path planning of automated guided vehicles in port areas based on spatiotemporal feature sampling according to claim 5, characterized in that, Specifically, step S51 involves using a mixed-integer nonlinear programming solver to solve the optimization problem.

Images