WO2024016457A1 - 基于自主绕障的异构型多智能体网联协同调度规划方法 - Google Patents
基于自主绕障的异构型多智能体网联协同调度规划方法 Download PDFInfo
- Publication number
- WO2024016457A1 WO2024016457A1 PCT/CN2022/119207 CN2022119207W WO2024016457A1 WO 2024016457 A1 WO2024016457 A1 WO 2024016457A1 CN 2022119207 W CN2022119207 W CN 2022119207W WO 2024016457 A1 WO2024016457 A1 WO 2024016457A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- path
- agent
- time window
- time
- task
- Prior art date
Links
- 238000013439 planning Methods 0.000 title claims abstract description 67
- 238000000034 method Methods 0.000 title claims abstract description 43
- 230000006855 networking Effects 0.000 title abstract 3
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 36
- 238000007726 management method Methods 0.000 claims abstract description 13
- 230000008569 process Effects 0.000 claims abstract description 7
- 239000003795 chemical substances by application Substances 0.000 claims description 238
- 238000003780 insertion Methods 0.000 claims description 25
- 230000037431 insertion Effects 0.000 claims description 25
- 239000013598 vector Substances 0.000 claims description 15
- 238000011156 evaluation Methods 0.000 claims description 11
- 238000005070 sampling Methods 0.000 claims description 8
- 238000007689 inspection Methods 0.000 claims description 6
- 238000012384 transportation and delivery Methods 0.000 claims description 2
- 238000005516 engineering process Methods 0.000 abstract description 10
- 238000004364 calculation method Methods 0.000 abstract description 2
- 238000004088 simulation Methods 0.000 description 5
- 238000004891 communication Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
- G05D1/0214—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory in accordance with safety or protection criteria, e.g. avoiding hazardous areas
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
- G05D1/0217—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory in accordance with energy consumption, time reduction or distance reduction criteria
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Definitions
- the invention belongs to the technical field of multi-agent scheduling and path planning, and specifically relates to a heterogeneous multi-agent networked collaborative scheduling and planning method based on autonomous obstacle avoidance technology. It is a central and The distributed fusion heterogeneous multi-agent network collaborative scheduling planning method realizes efficient, flexible, robust and scalable heterogeneous multi-agent scheduling planning by exploring the autonomous intelligence capabilities of the agents. It can be used for Scheduling and path planning of multi-agent in complex scenarios such as smart logistics and smart warehousing.
- the existing technology mainly uses two different architectures of scheduling and planning systems, namely, a centralized architecture system and a distributed architecture system.
- distributed architecture systems Compared with centralized architecture systems, distributed architecture systems have the advantages of greater flexibility and scalability.
- each agent "does its own thing", which lays hidden dangers for the security of the system. And it is not conducive to the unified management and control of the intelligent agents in the scene by users. Therefore, the current distributed architecture system is rarely used in real environments such as smart logistics and smart warehousing.
- Centralized architecture systems are currently widely used. They have the advantages of easy control and easy acquisition of optimal solutions.
- the intelligent capabilities of agents in existing centralized architecture systems are not fully utilized, and the control center needs to complete all calculations in the system.
- the present invention provides a heterogeneous multi-agent network collaborative scheduling planning method based on autonomous obstacle avoidance technology that integrates central and distributed systems. This method fully utilizes the autonomous intelligence capabilities of the agents and combines the central architecture with The advantages of distributed architecture realize a heterogeneous scheduling planning method that is more flexible, efficient, robust and can accommodate a larger number of agents.
- the invention includes a control center, an intelligent body and a data sharing terminal.
- the intelligent body refers to an intelligent robot that performs transportation tasks in an application scenario.
- the algorithm part of the control center is deployed on a computer, and the data sharing terminal is deployed on a server.
- the control center performs task allocation and collaborative path search.
- Task allocation adopts basic nearest allocation logic, and collaborative path search is implemented based on the Dijkstra algorithm.
- This invention also considers the repeated road sections and opposite driving sections in the path that bring problems to the intelligent driving. Safety hazards. If the paths of two agents both contain the same road segment, then the road segment is a duplicate road segment of the two agents. If the paths of the two agents both contain the same road segment and their paths pass through the road segment in opposite directions.
- the road section is a repeated road section and an opposite driving section of the two agents; the agent uses the existing time window algorithm to perform motion planning and trace driving based on the control center path search results.
- the agent uses the existing local path planning algorithm carried by itself to circumvent the obstacle.
- the time window inspection and adjustment strategy designed by the present invention is executed to ensure the collaborative relationship between the agents in the scene during subsequent operations;
- the data sharing terminal is used to store the time window data structure required for collision avoidance between agents. Each agent can independently access or modify the data stored in the data sharing terminal.
- the method of the present invention mainly includes five parts: task management part, collaborative path search part, motion planning part, autonomous obstacle avoidance part and time window adjustment part; wherein the task management part and collaborative path search part are implemented in the control center, and the motion planning part , the autonomous obstacle avoidance part and the time window adjustment part are implemented separately on each agent side.
- the task management part realizes dynamic addition of tasks and task allocation;
- the collaborative path search part is implemented based on the dijkstra algorithm, and at the same time, the safety hazards caused by repeated road sections and opposite driving sections in the path are considered for the intelligent agent, making the method of the present invention an intelligent agent planning
- the path produced is safer and more reliable; in the motion planning part, each agent uses the time window algorithm for each path based on the results of the collaborative path search by the control center to avoid conflicts with other agents in the scene and obtain specific speeds. level execution instructions; the present invention innovatively introduces the autonomous planning of intelligent agents into the multi-agent scheduling and planning method and the multi-agent collaborative application system.
- the intelligent agent uses The local path planning algorithm of the autonomous obstacle avoidance part for autonomous obstacle avoidance; autonomous obstacle avoidance will break the driving behavior planned by the motion planning part, so after the autonomous obstacle avoidance is completed, the time window adjustment part needs to perform time window inspection and adjustment.
- the heterogeneous multi-agent network collaborative scheduling planning method based on autonomous obstacle avoidance technology includes the following specific steps:
- Step 1 Carry out system configuration and transportation task management for multi-agent network collaborative scheduling planning
- the system includes a control center, an intelligent agent and a data sharing terminal; the task management part is used to add transportation tasks to the system and assign tasks.
- the agent is an intelligent robot that performs transportation tasks in the scene; the control center is used for task allocation and collaborative path search; the data sharing terminal is used to store the time window data structure required for collision avoidance between agents; each agent can be accessed independently Or modify the data stored in the data sharing terminal.
- the configuration of the multi-agent networked collaborative scheduling planning system specifically includes the startup of the control center and each agent, the setting of the scene map in the control center and each agent, and the initialization of the time window data structure of the data sharing end.
- the scene map includes node information and road section information.
- the nodes need to be positioned at fixed intervals (such as 10% of the scene width) according to the length and width of the scene. Some key points in the scene (such as pickup points, delivery points, etc.) It should also be set as a node. If there is an existing driving route in the scene, the location of the node should be set at a fixed distance on the existing driving route; a road segment is a line segment formed by two adjacent nodes. Multiple Road segments are connected to form a path.
- the transportation task can take any node in the map as the task end point and the position of the corresponding agent performing the task as the task starting point; during specific implementation, task allocation can be carried out according to the order in which tasks are added.
- Step 2 Collaborative path search part - perform collaborative path search for each agent that is about to perform the task, and obtain the corresponding path list.
- this invention considers the safety hazards caused by repeated road sections and opposite driving sections for the intelligent agent. If the current intelligent agent If the paths of the current agent and other agents both contain the same road segment, then the road segment is a duplicate of the current agent. If the paths of the current agent and other agents both contain the same road segment and they pass through the road segment in opposite directions, then This road segment is the repeated road segment and the opposite driving road segment of the current agent; during specific implementation, check each path in the path list PATH, and set a certain path in the path list to path 0.
- path 0 is different from any other agent
- the ratio of the number of repeated segments of the path being executed to the total number of segments on path 0 is greater than the set threshold (such as 0.4) or the ratio of the number of opposite segments of the path being executed by any other agent to the total number of segments on path 0 If the ratio is greater than the set threshold (such as 0.3), path 0 is an unavailable path, and path 0 is removed from the path list PATH.
- Step 3 Motion planning part - perform motion planning to obtain the execution path
- the agent obtains the results of the collaborative path search by the control center and uses the existing time window algorithm for motion planning. It performs a time window insertion operation on each path and selects the path that can complete the task earliest as the execution path.
- Step 4 Motion planning part - calculate the driving speed based on the execution path
- the driving speed on the road segment is calculated based on the length of each road segment and the time window insertion result, and is sent to the chassis of the agent, and the agent drives at this speed.
- Step 5 Autonomous obstacle avoidance part -
- the local path planning algorithm uses the existing dynamic time window algorithm (Dynamic Window Approach, DWA).
- Step 6 Time window adjustment part - Since the time for the intelligent agent to autonomously avoid obstacles cannot be determined, after the intelligent agent autonomously avoids obstacles, it is necessary to adopt the time window inspection and adjustment strategy designed by the method of the present invention to ensure that the scene will be in the subsequent operation process. Collaborative relationships between agents.
- multiple intelligent agents can be implemented to perform tasks in the scene flexibly and efficiently.
- the present invention provides a heterogeneous multi-agent network-connected collaborative scheduling planning method based on autonomous obstacle avoidance technology that integrates centralized and distributed technologies.
- the intelligent agent realizes Collaborative resolution of conflicts and autonomous obstacle avoidance, this method is more efficient, flexible, and robust than traditional centralized architecture systems, and the system can accommodate a larger number of agents.
- the intelligent agent has the ability to autonomously avoid obstacles.
- the control center does not need to monitor the operating status of the agent at all times, which significantly reduces the burden of network communication.
- the intelligent agent has the ability to independently respond to emergencies and abnormal conditions, and is more flexible than the traditional centralized architecture system in which the control center responds to emergencies through re-planning strategies.
- Figure 1 is the system architecture adopted for the specific implementation of the method of the present invention, including a control center, an intelligent agent and a data sharing terminal.
- Figure 2 is a flow chart of the task allocation method of the task management part of the present invention.
- FIG. 3 is a schematic diagram of a scene used to illustrate specific embodiments of the present invention.
- Figure 4 is an example of the time window data structure of the data sharing end of the present invention.
- Fig. 5 is a flow chart of the algorithm for performing path checking by the collaborative path search part of the present invention.
- Figure 6 is an example of the time window insertion operation performed by the motion planning part of the present invention.
- Figure 7 is a schematic diagram of the speed sampling space of the local path planning part of the present invention.
- Figure 8 is a schematic diagram of the obstacle avoidance path independently planned by some intelligent agents of the present invention.
- the present invention provides a heterogeneous multi-agent network-connected collaborative scheduling planning method based on autonomous obstacle avoidance technology that integrates centralized and distributed systems to achieve efficient, flexible and seamless operation of multiple agents in smart logistics, smart warehousing and other scenarios.
- the system architecture adopted for the specific implementation of the method of the present invention is shown in Figure 1.
- the method of the present invention mainly includes five parts: task management part, collaborative path search part, motion planning part, autonomous obstacle avoidance part and time window adjustment part; among them, the collaborative path search part fully considers repeated road sections and opposite driving sections as intelligent objects
- the potential safety hazards caused by driving make the path planned by the system for the agent safer and more reliable
- the method of the present invention is a heterogeneous method that integrates a centralized architecture and a distributed architecture, making full use of the autonomous intelligence of the agent Capability, combining the technical advantages of centralized architecture and distributed architecture, compared with traditional centralized architecture systems, flexibility, robustness, and scalability have been greatly improved.
- the task management part implements dynamic addition of tasks and task assignment.
- the strategic process of task allocation is shown in Figure 2.
- the collaborative path search part is implemented based on the Dijkstra algorithm.
- the motion planning part avoids conflicts between agents.
- Each agent accesses the data sharing terminal and performs time window insertion operations to obtain execution instructions specific to the speed level.
- the autonomous obstacle avoidance part realizes the autonomous avoidance of the agent when encountering obstacles, that is, using the local path planning algorithm to plan the obstacle avoidance path.
- the time window adjustment part implements the time window inspection and adjustment after the autonomous obstacle avoidance is completed to ensure the collaborative relationship between the agents in the scene during subsequent operations.
- Step 1 Configure the multi-agent network collaborative scheduling planning system.
- the system includes a control center, agents and data sharing terminals; the task management part - add tasks to the system and assign tasks.
- the agent is an intelligent robot that performs transportation tasks in the scene; the control center is used for task allocation and collaborative path search; the data sharing terminal is used to store the time window data structure required for collision avoidance between agents; each agent can be accessed independently Or modify the data stored in the data sharing terminal.
- the configuration of the multi-agent networked collaborative scheduling planning system specifically includes the startup of the control center and each agent, the setting of the scene map in the control center and each agent, and the initialization of the time window data structure of the data sharing end.
- the startup of the control center and each agent is to start the computer and each agent that deploys the algorithm part of the control center;
- the setting of the scene map in the control center and each agent means inputting the same map file to the control center and each agent to ensure The control center and each agent use the same scene map;
- the time window data structure maintains a time window vector for each road segment in the map.
- Each element in the time window vector corresponds to a period of time that an agent occupies the road segment.
- the time window data structure is empty.
- Figure 4 shows an example of the time window data structure during system operation.
- the task will be assigned to the agent whose current location is closest to the end of the task according to the proximity principle.
- Step 2 Collaborative path search part - perform collaborative path search for each agent that is about to perform the task, and obtain the corresponding path list.
- S is used to record all the nodes in the map that have found the shortest path
- U is used to record the nodes that have not yet found the shortest path.
- D stores the shortest distance from the starting point s to other nodes in the map in node number order.
- P stores the shortest route from the task starting point s to some other node z in the map in node number order. The previous node of middle node z.
- each element in D is the distance from the task starting point s to other nodes in the map (if a node v in U is not adjacent to the task starting point s, then the corresponding element value in D is ⁇ );
- the algorithm flow is shown in Figure 5.
- the length of the path PATH z is l (that is, there are l nodes on the path PATH z ).
- the path list PATH E planned for AGENT 3 includes the following 3 paths:
- path 2 After checking the paths of these three paths, there is only one feasible path in the path list PATH E , namely path 1.
- path 2 The reason why path 2 was removed is that the opposite traveling section between it and the path of AGENT 4 is different from the path total.
- the ratio of the number of road segments is greater than 0.3.
- path 3 The reason why path 3 is removed is that the ratio of the repeated segments between it and the path of AGENT 1 to the total number of path segments is greater than 0.4.
- Step 3 Motion planning part - the agent obtains the results of the collaborative path search by the control center (the path list obtained in step 2) and uses the existing time window algorithm to perform motion planning, and performs a time window insertion operation on each path. Select the path that can complete the task earliest as the execution path.
- the agent After the control center sends the results of the collaborative path search to the agent, the agent first needs to access the data sharing terminal to obtain the current time window structure.
- the time window data structure maintains a time window vector for each road segment in the map. Each element in the time window vector One element corresponds to the period of time that an agent occupies the road segment, and then for each path received by the agent, the following operations are performed:
- the The location can be used as an insertion location.
- the time of entering this section should be after the time when the agent releases the previous section on its path.
- time window insertion on a path contains multiple road segments a m . After finding insertion positions that meet the conditions in the time window vector for all the road segments on the path, the time window insertion operation on this path is successful.
- the time when the agent releases the last section of the path is the estimated time to complete the task when the agent uses this path to perform the task.
- a time window insertion operation is performed on each path received by the agent, and then the path that is expected to complete the task at the earliest is selected as the final path for the agent to perform the task.
- the agent By performing a time window insertion operation on each path received by the agent, after selecting the final execution path, the agent needs to send a message to the data sharing terminal to synchronously update the time window insertion result of the selected execution path to the data sharing terminal. .
- Step 4 Motion planning part - For the execution path obtained in step 3, the driving speed on the road segment is calculated based on the length of each road segment and the time window insertion result, and is sent to the chassis of the agent. The body travels at this speed.
- the agent After the agent completes the time window insertion operation according to the current time window structure, it can obtain the length of each section of the execution path and the driving time on the section, and further calculate the length of each section of the execution path that the agent should drive on.
- the speed is sent to the chassis actuator of the smart body to realize the tracking driving of the smart body.
- Step 5 Autonomous obstacle avoidance part - when the agent encounters an obstacle, the agent uses the local path planning algorithm to autonomously avoid obstacles.
- the agent in the method of the present invention will use a local path planning algorithm to self-circulate.
- the target point of autonomous obstacle avoidance is the next node on the running path of the current node of the agent.
- the target point of autonomous obstacle avoidance is the task end.
- the method of the present invention specifically adopts the existing dynamic time window algorithm (Dynamic Window Approach, DWA), which is a dynamic decision-making algorithm based on a limited observation range.
- DWA Dynamic Window Approach
- the speed sampling space of the agent is mainly limited by the following factors: the maximum speed and minimum speed limit of the agent itself; the limit of the maximum acceleration and maximum deceleration of the agent, as shown in Figure 7, which results in the following
- the speed that can be achieved during a forward simulation period is limited to a dynamic time window.
- goal(v, ⁇ ) is the goal evaluation function, which is used to evaluate the position of the robot when it reaches the end of the simulation trajectory and the position of the target point, and the orientation angle and target orientation when it reaches the end of the simulation trajectory at the currently set sampling speed.
- the gap between the corners the greater the gap, the lower the score
- dist(v, ⁇ ) is the distance evaluation function, used to evaluate the distance between the trajectory and the nearest obstacle, the smaller the distance between the trajectory and the nearest obstacle, The larger this value is
- path (v, ⁇ ) is the path evaluation function, which is used to evaluate the shortest distance between the points on the sampling trajectory and the global path. This evaluation function is mainly used to guide the agent to search according to the global path as much as possible. Travel to reach your destination earlier.
- Step 6 Time window adjustment part - after the agent circumvents the obstacle, the time window inspection and adjustment strategy designed by the method of the present invention is used to ensure the collaborative relationship between the agents in the scene during subsequent operation. .
- the intelligent agent's autonomous obstacle avoidance is a relatively time-consuming operation and the time required cannot be estimated in advance, if the intelligent agent has not reached the end of the task after the autonomous obstacle avoidance is completed, it is necessary to check the energy consumed by the obstacle avoidance after the obstacle is completed. time, if the time consumed exceeds half of the time window protection time, the time window needs to be adjusted. Assume that the time spent on autonomous obstacle avoidance is ⁇ , and the time window protection time is w (the time window protection time refers to the minimum time difference between their time windows if two agents need to pass through a road section). Specifically The strategy is as follows:
- the time consumed by autonomous obstacle avoidance is less than or equal to half of the time window protection time.
- the agent can make up for the time consumption by adjusting the driving speed, so there is no need to adjust the time window.
- the time consumed by autonomous obstacle avoidance exceeds half of the time window protection time. If the agent only relies on the agent to adjust the speed during subsequent driving to make up for the time error, there will be a greater risk of conflict, so this It is necessary to adjust the time window and re-plan the motion.
- step 3 According to the current position of the agent, go to step 3 to determine the subsequent driving path of the agent and re-plan the movement.
Landscapes
- Engineering & Computer Science (AREA)
- Aviation & Aerospace Engineering (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
Abstract
一种基于自主绕障的异构型多智能体网联协同调度规划方法,包括:进行多智能体网联协同调度规划的系统配置和运输任务管理;系统中每一台即将执行运输任务的智能体进行协同路径搜索,得到相应的路径列表;智能体进行运动规划得到执行路径;根据执行路径计算得到行驶速度;智能体在运行的过程中,进行自主绕障;时间窗调整。通过采用时间窗算法和引入智能体的自主绕障技术,实现智能体之间冲突的协同解决和自主绕障,更加高效、灵活、鲁棒且系统可容纳的智能体数目更多。
Description
本发明属于多智能体调度及路径规划技术领域,具体涉及一种基于自主绕障技术的异构型多智能体网联协同调度规划方法,是一种引入智能体自主绕障技术的中心式与分布式融合的异构型多智能体网联协同调度规划方法,通过发掘智能体的自主智能能力,实现高效、灵活、鲁棒且可扩展性强的异构型多智能体调度规划,可用于智慧物流、智慧仓储等复杂场景下对多智能体的调度及路径规划。
近年来,由于柔性制造系统的出现、定制机器人需求的增加以及中小企业工业自动化需求的增加,移动智能体在制造业、医疗及智能物流等领域发展迅速,得到了广泛的应用。
以最大限度地提高生产效率为目的,使用者往往期望多台智能体同时在场景中运行,这对场景中智能体的安全性提出了挑战,保证多台智能体在场景中无冲突地、高效地执行运输任务是多智能体系统需要解决的首要问题。完成这一任务的就是调度规划系统,它需要协调场景中多台智能体的运行,在保证安全性的前提下,尽可能地提升系统的运行效率。
目前,现有技术主要采用两种不同架构的调度规划系统,即中心式架构系统和分布式架构系统。分布式架构系统相较于中心式架构系统有着灵活性强、可扩展性强等优势,但在此种架构的系统中,各智能体“各自为政”,这为系统的安全性埋下了隐患,且不利于使用者对场景中智能体的统一管理与控制,所以目前分布式架构的系统在智慧物流、智慧仓储等真实环境中鲜有应用。中心式架构系统目前应用非常广泛,它具有便于控制、易于获得最优解等优势,但现有的中心式架构系统中智能体的智能能力未被充分利用,控制中心需要完成系统中所有的计算任务,且系统中所有智能体需要不断地与控制中心进行信息交换,这造成了现有的中心式架构系统的控制中心计算压力大、网络通信负担大,进一步限制了系统可容纳智能体数目的提升,且此种系统存在着“单点故障”的缺陷,灵活性及鲁棒性差。
发明内容
本发明提供了一种基于自主绕障技术的中心式与分布式融合的异构型多智能体网联协同调度规划方法,该方法充分利用了智能体的自主智能能力,结合了中心式架构与分布式架构的优势,实现了一种更加灵活、高效、鲁棒且可容纳智能体数目更多的异构型调度规划方法。
本发明包括控制中心、智能体和数据共享端,智能体是指应用场景下执行运输任务的智能机器人,控制中心算法部分部署在一台计算机上,数据共享端部署在一台服务器上。其中,由控制中心进行任务分配及协同路径搜索,任务分配采用基础的就近分配逻辑,协同路径搜索基于dijkstra算法实现,本发明同时考虑了路径中重复路段及对向行驶路段为智能体行驶带来的安全隐患,若两智能体的路径中均包含同一段路段,则该路段是两台智能体的重复路段,若两台智能体的路径中均包含同一段路段且其经过该路段的方向相反,则该路段是两台智能体的重复路段和对向行驶路段;智能体根据控制中心路径搜索的结果采用现有的时间窗算法进行运动规划并循迹行驶,在智能体遇到障碍物时,智能体利用自身搭载的现有的局部路径规划算法进行绕障,绕障结束后,执行本发明设计的时间窗检查与调整策略以保障后续运行过程中场景中智能体之间的协同关系;数据共享端用于存放智能体间碰撞避免所需的时间窗数据结构,各智能体均可以独立访问或修改数据共享端中存放的数据。
本发明方法主要包括五个部分:任务管理部分、协同路径搜索部分、运动规划部分、自主绕障部分及时间窗调整部分;其中任务管理部分和协同路径搜索部分在控制中心实现,而运动规划部分、自主绕障部分及时间窗调整部分在各智能体端分别实现。
任务管理部分实现动态添加任务和任务分配;协同路径搜索部分基于dijkstra算法实现,同时考虑了路径中重复路段及对向行驶路段为智能体行驶带来的安全隐患,使得本发明方法为智能体规划出的路径更加安全可靠;在运动规划部分中,各智能体根据控制中心协同路径搜索的结果,对其中的每条路径采用时间窗算法来避免与场景中其他智能体的冲突,获得具体到速度层面的执行指令;本发明创新性地将智能体的自主规划引入到了多智能体调度与规划方法及多智能体协同应用系统中,当智能体在行驶过程中遇到障碍物时,智能体采用自主绕障部分的局部路径规划算法进行自主绕障;自主绕障会打破运动规划部分规划好的行驶行为,所以在自主绕障结束后需要由时间窗调整部分进行时间窗检查与调整。
本发明提供的基于自主绕障技术的异构型多智能体网联协同调度规划方法包括如下具体步骤:
步骤1进行多智能体网联协同调度规划的系统配置和运输任务管理;
系统包括控制中心、智能体和数据共享端;任务管理部分用于向系统中添加运输任务并进行任务分配。
智能体为场景中执行运输任务的智能机器人;控制中心用于进行任务分配及协同路径搜索;数据共享端用于存放智能体间碰撞避免所需的时间窗数据结构;各智能体均可以独立访问或修改数据共享端中存放的数据。
多智能体网联协同调度规划系统配置具体包括控制中心及各智能体的启动、控制中心和各智能体中场景地图的设置及数据共享端时间窗数据结构的初始化。场景地图中包括节点信息和路段信息,节点需根据场景的长度、宽度以固定间距(如场景宽度的10%)设定位置,场景中的一些关键点(如取货点、投货点等)也应被设定为节点,若场景中存在即有的行驶路线,则应在即有的行驶路线上以固定间距设定节点的位置;路段为两相邻节点相邻所形成的线段,多条路段相连形成一条路径。
运输任务可以以地图中任一节点为任务终点,以执行任务的相应智能体的位置为任务起点;具体实施时,可根据任务添加的顺序进行任务分配。
步骤2协同路径搜索部分——为每一台即将执行任务的智能体进行协同路径搜索,得到相应的路径列表。
基于dijkstra算法获得任务起点与任务终点间所有无环路的路径,存储在路径列表PATH中,同时本发明考虑了重复路段、对向行驶路段为智能体行驶带来的安全隐患,若当前智能体与其他智能体的路径中均包含同一段路段,则该路段是当前智能体的重复路段,若当前智能体与其他智能体的路径中均包含同一段路段且其经过该路段的方向相反,则该路段是当前智能体的重复路段和对向行驶路段;具体实施时,检查路径列表PATH中的每一条路径,设路径列表中的某一路径为path
0,若path
0与其他任一智能体正在执行的路径的重复路段数与path
0上总路段数的比值大于设定阈值(如0.4)或与其他任一智能体正在执行的路径的对向行驶路段数与path
0上总路段数的比值大于设定阈值(如0.3),则path
0为不可用路径,将path
0从路径列表PATH中移除。
步骤3运动规划部分——进行运动规划得到执行路径;
智能体获得控制中心进行协同路径搜索的结果并采用现有的时间窗算法进行运动规划,对其中的每一条路径进行时间窗插入操作,从中选取能够最早完成任务的一条路径作为执行路径。
步骤4运动规划部分——根据执行路径计算得到行驶速度;
对于步骤3中获得的执行路径,根据路径上每条路段的长度及时间窗插入结果计算出在该条路段上的行驶速度,并下发给智能体的底盘,智能体按照此速度进行行驶。
步骤5自主绕障部分——在智能体运行的过程中,若智能体遇到障碍物,则由智能体按照自主绕障部分的局部路径规划算法执行障碍物的绕开操作,具体实施时,局部路径规划算法采用了现有的动态时间窗口算法(Dynamic Window Approach,DWA)。
步骤6时间窗调整部分——由于智能体自主绕障的时间无法确定,所以在智能体自主绕障结束后,需要采用本发明方法设计的时间窗检查与调整策略以保障后续运行过程中场景中智能体之间的协同关系。
通过以上步骤,即可实现多台智能体灵活、高效地在场景中执行任务。
与现有技术相比,本发明的有益效果:
本发明提供一种基于自主绕障技术的中心式与分布式融合的异构型多智能体网联协同调度规划方法,通过采用时间窗算法和引入智能体的自主绕障技术,实现智能体之间冲突的协同解决和自主绕障,本方法相较于传统的中心式架构系统更加高效、灵活、鲁棒且系统可容纳的智能体数目更多。
本发明提供的异构型多智能体网联协同调度规划方法具有以下优点:
1)引入了智能体的自主局部路径规划,显著降低了控制中心的计算压力。
2)智能体具备自主绕障能力,控制中心无需时刻监控智能体的运行状态,显著降低了网络通信的负担。
3)控制中心计算压力及网络通信负担的降低进一步促进了系统可容纳智能体数目的提升。
4)系统的运行不再仅仅依赖于控制中心,即消除了传统的中心式架构系统“单点故障”的缺陷,系统鲁棒性更强。
5)智能体具备自主应对突发状况及异常状况的能力,相较于传统的中心式架构系统中由控制中心通过重规划应对突发状况的策略灵活性更强。
图1是本发明方法具体实施采用的系统架构,包括控制中心、智能体和数据共享端。
图2是本发明任务管理部分任务分配方法的流程框图。
图3是本发明说明具体实施方式时所采用示例的场景示意图。
图4是本发明数据共享端时间窗数据结构的一种示例。
图5是本发明协同路径搜索部分进行路径检查的算法流程框图。
图6本发明运动规划部分进行时间窗插入操作的一种示例。
图7是本发明局部路径规划部分速度采样空间示意图。
图8是本发明自主绕障部分智能体自主规划的绕障路径示意图。
为使本发明的上述目的、特征和优点能够更加明显易懂,下面结合附图和具体实施方式对本发明作进一步详细的说明。
本发明提供一种基于自主绕障技术的中心式与分布式融合的异构型多智能体网联协同调度规划方法,实现多台智能体在智慧物流、智慧仓储等场景下高效、灵活、无冲突地执行任务,本发明方法具体实施采用的系统架构参考图1所示。本发明方法主要包括五个部分:任务管理部分、协同路径搜索部分、运动规划部分、自主绕障部分及时间窗调整部分;其中协同路径搜索部分充分考虑了重复路段、对向行驶路段为智能体行驶带来的安全隐患,使得系统为智能体规划出的路径更加安全可靠;此外,本发明方法是一种中心式架构与分布式架构融合的异构型方法,充分利用了智能体的自主智能能力,结合了中心式架构与分布式架构的技术优势,相较于传统的中心式架构系统,灵活性、鲁棒性、可扩展性均得到了较大的提升。
任务管理部分实现动态添加任务和任务分配。任务分配的策略流程参考图2所示。协同路径搜索部分基于dijkstra算法实现,同时考虑了路径中重复路段、对向行驶路段为智能体行驶带来的安全隐患,使得系统为智能体规划出的路径更加安全可靠。运动规划部分实现智能体之间冲突的避免,由各智能体分别访问数据共享端并进行时间窗插入操作,获得具体到速度层面的执行指令。自主绕障部分实现智能体在遇到障碍物时的自主绕开,即利用局部路径规划算法规划出绕障路径。时间窗调整部分实现在自主绕障结束后的时间窗检查与调整,以保障后续运行过程中场景中智能体之间的协同关系。
为便于理解,在介绍本发明方法的具体实施方式时,将以图3所示的场景为例,由图可见,场景中共存在4台智能体,分别为AGENT
1、AGENT
2、AGENT
3及AGENT
4,其中只有AGENT
3处于空闲状态。
本发明方法的具体实施包括如下步骤:
步骤1进行多智能体网联协同调度规划系统配置,系统包括控制中心、智能体和数据共享端;任务管理部分——向系统中添加任务并进行任务分配。
智能体为场景中执行运输任务的智能机器人;控制中心用于进行任务分配及协同路径搜索;数据共享端用于存放智能体间碰撞避免所需的时间窗数据结构;各智能体均可以独立访问或修改数据共享端中存放的数据。
多智能体网联协同调度规划系统配置具体包括控制中心及各智能体的启动、控制中心和各智能体中场景地图的设置及数据共享端时间窗数据结构的初始化。控制中心及各智能体的启动即启动控部署控制中心算法部分的计算机和各台智能体;控制中心和各智能体中场景地图的设置是指向控制中心和各智能体输入相同的地图文件,保证控制中心和各智能体使用相 同的场景地图;时间窗数据结构为地图中每一条路段维护一个时间窗向量,时间窗向量中的每一个元素对应着一台智能体占用该路段的一段时间,初始时时间窗数据结构为空,图4所示为系统运行过程中的时间窗数据结构示例。
向系统中添加以地图中任一节点为终点的任务,并根据任务添加的顺序进行任务分配,参考图2所示,具体的任务分配策略为:
1)若当前系统中无空闲智能体,则该任务暂缓分配;
2)若当前系统中有空闲智能体,则按照就近原则将该任务分配给当前位置距离任务终点最近的智能体。
以图3所示场景为例,若添加了终点为节点21的任务,根据任务分配策略,该任务将被分配给AGENT
3执行。
步骤2协同路径搜索部分——为每一台即将执行任务的智能体进行协同路径搜索,得到相应的路径列表。
基础的dijkstra算法的执行步骤如下:
1)获取任务起点s和任务终点e;
2)定义两个集合:S和U。S用来记录地图中所有已求出最短路径的节点,U用来记录还未求出最短路径的节点。定义两个向量:D和P,向量D中以节点编号顺序存储起点s到地图中其它节点的最短距离,向量P中以节点编号顺序存储任务起点s到地图中其它某一节点z的最短路由中节点z的上一节点。初始时,S中只有任务起点s,U中包含除任务起点s之外的地图中所有节点;D中各元素的值是任务起点s到地图中其它各节点的距离(若U中某节点v与任务起点s不相邻,则D中对应的元素值为∞);
3)从U中选出距离任务起点s最近的节点k,并将节点k加入到S中,同时,从U中移除节点k;
4)更新U中各个节点到起点s的距离(考虑以节点k为中间点起点s到U中某个节点v的距离可能减小),同时更新向量D和向量P;
5)重复步骤3)和步骤4),直到遍历完地图中所有节点;
6)从终点e开始,反向检索向量P,获取起点s到终点e的最短路径。
设系统中共有A台智能体,除当前智能体外其他A-1台智能体中第i台智能体目前正在执行的路径长度为l
i(即路径上共有l
i个节点),则其他A-1台智能体中第i台智能体正在执行的路径可表示为
本发明基于dijkstra算法进行协同路径搜索的步骤如下:
1)采用基础的dijkstra算法求出一条任务起点和任务终点的两点之间的最短路径:s→k
1→k
2→…→k
n→e;其中k
1~k
n均为路径上的节点;
2)令k
0=s,k
n+1=e,对于路径上除任务终点e外的每一个节点k
m(m=0,1,2…,n),设k
m共有r个相邻节点,遍历k
m所相邻的每一个节点v
q(q=1,2,…,r);若v
q≠k
m+1,以v
q为任务起点,以e为任务终点,采用基础的dijkstra算法求出v
q与e之间的最短路径:v
q→n
1→n
2→…→n
p→e,n
1~n
p均为路径上的节点,则得到另一条s与e之间的路径:k
0→…→k
m→v
q→…→e,将其加入路径列表PATH。
3)步骤2)执行结束后,路径列表PATH中存储了任务起点s到任务终点e之间所有无环路路径,设路径列表PATH中共有S条路径,对于其中的每一条路径PATH
z(z=1,2,…,S),分别使用β
z和γ
z来记录路径PATH
z与其他智能体正在执行的路径的重复路段及对向行驶路段占路径PATH
z总路段数的比值的最大值。算法流程参考图5所示。
设路径PATH
z的长度为l(即路径PATH
z上共有l个节点),对于h=1,2,…,l-1,检查系统中除当前智能体外其他A-1台智能体路径N
i(i=1,2,…A-1),则对于w=1,2,…,l
i-1,l
i为路径N
i的长度(即路径上共有l
i个节点);从地图信息中分别查取由节点
和节点
连接而成的路段记为arc
h,及节点
和节点
连接而成的路段记为arc
w,若arc
h=arc
w,则路径PATH
z与路径N
i的重复路段数累加1,若arc
h=arc
w且
则路径PATH
z与路径N
i的对向行驶路段数也累加1。对于其他智能体当前执行路径的集合N,求得路径PATH
z与其中A-1条路径重复路段、对向行驶 路段的最大值,进而求得β
z与γ
z,若β
z≥0.4或γ
z≥0.3,则认为路径PATH
z与当前系统中正在执行任务的其他智能体有较大的产生冲突的风险,则将路径PATH
z从路径列表中移除。
在图3所示的场景中,假设AGENT
1、AGENT
2及AGENT
4当前的执行路径分别为:10→3→24→12→11→18→19,19→15→13→14及16→21→26→23→24→25,为AGENT
3规划出的路径列表PATH
E包含以下3条路径:
①22→24→3→2→9→16→21
②22→18→14→13→12→19→24→23→26→21
③22→24→12→11→18→23→26→21
对这3条路径进行路经检查后,路径列表PATH
E中仅存在一条路径可行,即路径①,路径②被移除的原因是其与AGENT
4的路径之间的对向行驶路段与路径总路段数的比值大于0.3,路径③被移除的原因是其与AGENT
1的路径之间的重复路段与路径总路段数的比值大于0.4。
步骤3运动规划部分——智能体获得控制中心进行协同路径搜索的结果(步骤2得到的路径列表)并采用现有的时间窗算法进行运动规划,对其中的每一条路径进行时间窗插入操作,从中选取能够最早完成任务的一条路径作为执行路径。
控制中心将协同路径搜索的结果发送给智能体后,智能体首先需要访问数据共享端获取当前时间窗结构,时间窗数据结构为地图中每一条路段维护一个时间窗向量,时间窗向量中的每一个元素对应着一台智能体占用该路段的一段时间,而后对于智能体收到的每一条路径,执行如下操作:
对于每条路径上的每一段路段a
m(m=1,2,…,n):
1)以最短长度的时间窗(最短时间窗长度为路段a
m的长度与智能体最大行驶速度的比值)进行插入。寻找满足以下两个条件的插入位置:
①在路段a
m的时间窗向量中,若某个位置的后一个时间段的起始时间与该位置的前一个时间段的终止时间之差足以容纳此最短长度的时间窗的长度,则该位置可作为插入位置。
②若m≠1,进入此段路段的时间应在智能体释放其路径上的上一段路段的时间之后。
2)若m≠1,判断智能体释放其路径上的上一段路段的时间与进入此段路段的时间是否相等。若不相等,则延长上一段路段的释放时间,使其等于此段路段的进入时间。
3)检查时间窗之间是否存在重叠。若不重叠,插入成功,继续对下一段路段进行操作。若重叠,返回路径上的上一段路段,并返回到步骤1)。
对一条路径进行时间窗插入操作具体是,一条路径包含多个路段a
m,对路径上的所有路段在时间窗向量中找到满足条件的插入位置后,这条路径的时间窗插入操作成功。
对一条路径进行时间窗插入操作成功后,智能体释放该条路径上最后一段路段的时间即为智能体采用该条路径执行任务时预计完成任务的时间。对智能体收到的每一条路径均进行时间窗插入操作,而后选取预计能够最早完成任务的一条路径作为智能体执行任务的最终路径。
通过对智能体收到的每一条路径进行时间窗插入操作,选取出最终的执行路径后,智能体需要向数据共享端发送消息,将选取的执行路径的时间窗插入结果向数据共享端同步更新。
仍以图3所示场景为例,在步骤2中的协同路径搜索部分求出AGENT
3只有一条候选路径,即22→24→3→2→9→16→21,设路径上的路段依次为a
9→a
16→a
7→a
1→a
17→a
22,设AGENT
3收到控制中心发送的候选路径后,在进行运动规划(也即进行时间窗插入操作)前,上述路段的时间窗数据结构为图6中(1)所示,则进行运动规划后,上述路段的时间窗数据结构如图6中(2)所示。
步骤4运动规划部分——对于步骤3中获得的执行路径,根据路径上每条路段的长度及时间窗插入结果计算出在该条路段上的行驶速度,并下发给智能体的底盘,智能体按照此速度进行行驶。
智能体根据当前时间窗结构完成时间窗插入操作后,即可获得执行路径上各段路段的长度及应在该路段上的行驶时间,进一步计算出智能体在执行路径中各段路段上应行驶的速度,将此速度下发给智能体底盘执行机构即可实现智能体循迹行驶。
步骤5自主绕障部分——在智能体遇到障碍物时,智能体利用局部路径规划算法进行自主绕障。
当智能体在行驶过程中前方遇到障碍物时,不同于传统的中心式架构系统中仅依赖于控制中心进行重规划的策略,本发明方法中的智能体将采用局部路径规划算法进行自主绕障,自主绕障的目标点为智能体当前所在节点在运行路径上的下一个节点,特别地,当运行路径上的下一个节点为路径的最后一个节点时,自主绕障的目标点为任务终点。在实施时,本发 明方法具体采用了现有的动态时间窗口算法(Dynamic Window Approach,DWA),这是一种基于有限观测范围的动态决策类算法,该算法共包含三个步骤:
1)速度采样
智能体的速度采样空间主要受以下几个因素限制:智能体本身存在的最大速度及最小速度限制;智能体存在的最大加速度、最大减速度的限制,参考图7所示,这导致智能体在下一个前向模拟周期内所能达到的速度被限制在一个动态时间窗口内。
2)轨迹评分
在采样的速度组中,有若干组轨迹均是可行的,因此采用评价函数为每条轨迹进行评价,采用的评价函数如下:
H(v,ω)=α·goal(v,ω)+β·dist(v,ω)+γ·path(v,ω) (4.5)
其中,goal(v,ω)是目标评价函数,用来评价机器人在当前设定的采样速度下,达到模拟轨迹末端时的位置与目标点的位置及到达模拟轨迹末端时的朝向角与目标方位角之间的差距,差距越大,评分越低;dist(v,ω)是距离评价函数,用来评价轨迹上与最近障碍物之间的距离,与最近障碍物之间的距离越小,此值越大;path(v,ω)为路径评价函数,用来评价采样轨迹上的点距离全局路径的最近距离,此评价函数主要是用来引导智能体尽可能地按照全局路径搜索的结果行进,以更早地到达目标点。
需要注意的是,由于goal(v,ω)、dist(v,ω)、path(v,ω)等3个评价函数的参量不同,在计算最终的评价值H(v,ω)前,需要将3部分的评价函数值进行归一化处理,而后分别与对应的系数α、β、γ相乘后求和,即可得到该组速度在下一段前向模拟周期内轨迹的评分,从所有采样轨迹中选取评分最高的一组速度作为下一周期内智能体的执行速度。
3)速度发布
在速度空间中采样多组速度并模拟出多组速度在下一前向模拟周期的轨迹后,对所有轨迹进行评分并选取出评分最高的轨迹,而后将评分最高的轨迹对应的速度下发给智能体,智能体在下一周期内按照此速度进行行驶,重复上述过程直至智能体到达目标点。
步骤6时间窗调整部分——在智能体绕障结束后,采用本发明方法设计的时间窗检查与调整策略以保障后续运行过程中场景中智能体之间的协同关系。。
由于智能体自主绕障是一项较为耗时的操作且所需时间无法提前预估,若智能体自主绕障结束后还未到达任务终点,则在绕障结束后需要检查绕障所消耗的时间,若所消耗的时间 超过了时间窗保护时间的一半,则需要进行时间窗的调整。设自主绕障所耗费的时间为δ,时间窗保护时间为w(时间窗保护时间是指若两台智能体均需要经过一条路段,它们的时间窗之间应最少相差出的时间),具体策略如下:
在此种情况下,自主绕障所消耗的时间小于或等于时间窗保护时间的一半,在后续的行驶中智能体可通过调整行驶速度来弥补时间的消耗,故无需进行时间窗的调整。
在此种情况下,自主绕障所消耗的时间超过了时间窗保护时间的一半,若仅依靠智能体在后续的行驶中通过调整速度来弥补时间误差将存在较大的冲突的风险,所以此时需要进行时间窗的调整并重新进行运动规划。
时间窗调整并重新进行运动规划的具体做法为:
1)访问数据共享端,清除当前智能体的时间窗数据。
2)获得清除当前智能体时间窗数据后的时间窗结构。
3)根据智能体当前位置,转到步骤3确定智能体后续的行驶路径并重新进行运动规划。
仍以图3所示场景为例,当AGENT
3行驶至节点24所在位置时,检测到前方行驶路径上存在障碍物,智能体随即采用步骤5中的局部路径规划算法进行自主绕障,图8为AGENT
3可能规划出的一种绕障路径;AGENT
3当绕障结束到达节点3所在位置时,若绕障额外消耗了5.2s,即δ=5.2,设时间窗保护时间为8s,即w=8,由于
所以接下来AGENT
3应清除当前时间窗数据结构中自身的时间窗,而后对后续行驶路径(3→2→9→16→21)重新进行运动规划。
需要注意的是,公布实施例的目的在于帮助进一步理解本发明,但是本领域的技术人员可以理解:在不脱离本发明及所附权利要求的精神和范围内,各种替换和修改都是可能的。因此,本发明不应局限于实施例所公开的内容,本发明要求保护的范围以权利要求书界定的范围为准。
Claims (5)
- 一种基于自主绕障的异构型多智能体网联协同调度规划方法,其特征是,包括如下步骤:步骤1:进行多智能体网联协同调度规划的系统配置和运输任务管理;系统包括控制中心、智能体和数据共享端;智能体为场景中执行运输任务的智能机器人;控制中心用于进行任务分配及协同路径搜索;数据共享端用于存放智能体间碰撞避免所需的时间窗数据结构;各智能体均可独立访问或修改数据共享端中存放的数据;多智能体网联协同调度规划的系统配置包括:控制中心及各智能体的启动、控制中心和各智能体中场景地图的设置及数据共享端的时间窗数据结构的初始化;场景地图中包括节点和路段;将场景的关键点设定为节点;场景的关键点包括取货点和投货点;另外还根据场景的长度和宽度以固定间距设定节点的位置;若场景中存在即有的行驶路线,则在即有的行驶路线上以固定间距设定节点的位置;场景的路段为两相邻节点所形成的线段,多条路段相连形成一条路径;运输任务管理是指向系统中添加运输任务并进行任务分配;运输任务以场景地图中任一节点为任务终点,以执行任务的相应智能体的位置为任务起点;步骤2:系统场景中每一台即将执行运输任务的智能体进行协同路径搜索,得到相应的路径列表;包括如下过程:21)求出一条任务起点和任务终点之间的最短路径,记作:s→k 1→k 2→…→k n→e;其中,s为任务起点;e为任务终点;k 1~k n均为路径上的节点;22)令k 0=s,k n+1=e,对于路径上除任务终点e外的每一个点k m,m=0,1,2…,n,设k m共有r个相邻节点,遍历k m所相邻的每一个节点v q,q=1,2,…,r;若v q≠k m+1,以v q为任务起点,以e为任务终点,求出v q与e之间的最短路径:v q→n 1→n 2→…→n p→e,n 1~n p均为路径上的节点,则得到另一条s与e之间的路径:k 0→…→k m→v q→…→e,将其加入路径列表;23)步骤22)执行结束后,路径列表中存储了任务起点s到任务终点e之间所有无环路路径;对于其中的每一条路径,计算并记录该路径与其他智能体正在执行的路径的重复路段及对向行驶路段占该路径总路段数的比值的最大值;所述重复路段指的是:若当前智能体与其他智能体的路径中均包含同一段路段,则该路段是当前智能体的重复路段;所述重复路段及对向行驶路段指的是:若当前智能体与其他智 能体的路径中均包含同一段路段且经过该路段的方向相反,则该路段是当前智能体的重复路段及对向行驶路段;包括如下过程:231)设路径列表PATH中共有S条路径,对于其中的每一条路径PATH z,z=1,2,…,S,分别使用β z和γ z来记录路径PATH z与其他智能体正在执行的路径的重复路段及对向行驶路段占路径PATH z总路段数的比值的最大值;232)设路径PATH z的长度为l,即路径PATH z上共有l个节点;对于h=1,2,…,l-1,检查系统中除当前智能体外其他A-1台智能体路径N i,i=1,2,…,A-1,则对于w=1,2,…,l i-1,l i为路径N i的长度,从地图中分别查取由节点 和节点 连接而成的路段记为arc h,和由节点 和节点 连接而成的路段,记为arc w;234)对于其他智能体当前执行路径的集合N,求得路径PATH z与其中A-1条路径重复路段的最大值和对向行驶路段的最大值,进而求得β z与γ z;对β z和γ z分别设置阈值;若β z超过相应阈值或γ z超过相应阈值,则将路径PATH z从路径列表中移除;步骤3:智能体进行运动规划得到执行路径;根据协同路径搜索得到的路径列表结果,采用时间窗算法进行运动规划,对其中的每一条路径进行时间窗插入操作,从中选取能够最早完成任务的一条路径作为执行路径;步骤4:根据执行路径计算得到行驶速度;根据步骤3中获得的执行路径上每条路段的长度及时间窗插入结果,计算出在该条路段上的行驶速度,并发给智能体的底盘,智能体按照此速度行驶;步骤5:智能体在运行的过程中,进行自主绕障:若智能体遇到障碍物,则由智能体按照局部路径规划算法绕开障碍物,到达自主绕障的目标点;自主绕障的目标点为智能体当前所在节点在运行路径上的下一个节点;步骤6时间窗调整:在智能体自主绕障结束后,若智能体还未到达任务终点,设计并采用时间窗检查与调整策略,保障在智能体后续运行过程中,场景中多智能体之间的协同调度;包括:6A)当智能体自主绕障所消耗的时间小于或等于时间窗保护时间的一半时,无需进行时间窗的调整;在后续的行驶中智能体可通过调整行驶速度来弥补时间的消耗;时间窗保护时间是指若两台智能体均需经过一条路段,它们的时间窗之间应最少相差出的时间;6B)当自主绕障所消耗的时间超过了时间窗保护时间的一半时,按照以下方法进行时间窗的调整:访问数据共享端,清除当前智能体的时间窗数据;获得清除当前智能体时间窗数据后的时间窗结构;再根据智能体当前位置,转到步骤3确定智能体后续的行驶路径并重新进行运动规划;通过以上步骤,即可实现基于自主绕障的异构型多智能体网联协同调度规划。
- 如权利要求1所述基于自主绕障的异构型多智能体网联协同调度规划方法,其特征是,步骤234)中,设置β z的阈值0.4;设置γ z的阈值为0.3。
- 如权利要求1所述基于自主绕障的异构型多智能体网联协同调度规划方法,其特征是,步骤21)具体是采用dijkstra算法求出一条两点间的最短路径。
- 如权利要求1所述基于自主绕障的异构型多智能体网联协同调度规划方法,其特征是,步骤3具体是采用时间窗算法进行运动规划,计算得到执行路径;包括:智能体首先访问数据共享端获取当前时间窗数据结构;时间窗数据结构为地图中每一条路段维护一个时间窗向量,时间窗向量中的每一个元素对应一台智能体占用该路段的一段时间;而后对于智能体收到的每一条路径,执行如下操作:对于每条路径上的每一段路段a m,m=0,1,2…,n:31)以最短长度的时间窗进行插入;时间窗的最短长度为路段a m的长度与智能体最大行驶速度的比值;寻找满足以下两个条件的插入位置:①在路段a m的时间窗向量中,若某个位置的后一个时间段的起始时间与该位置的前一个时间段的终止时间之差足以容纳最短长度的时间窗的长度,则该位置可作为插入位置;②若m≠1,进入此段路段的时间应在智能体释放其路径上的上一段路段的时间之后;32)若m≠1,判断智能体释放其路径上的上一段路段的时间与进入此段路段的时间是否相等;若不相等,则延长上一段路段的释放时间,使其等于此段路段的进入时间;33)检查时间窗之间是否存在重叠;若不重叠,插入成功,继续对下一段路段进行操作;若重叠,返回路径上的上一段路段,并返回到步骤31);通过对智能体收到的每一条路径进行时间窗插入操作,选取出最终的执行路径后,智能体向数据共享端发送消息,将选取的执行路径的时间窗插入结果向数据共享端同步更新。
- 如权利要求1所述基于自主绕障的异构型多智能体网联协同调度规划方法,其特征是,步骤5中,智能体按照局部路径规划算法,在运行的过程中进行自主绕障;具体是采用动态时间窗口算法DWA,包括如下步骤:51)速度采样;52)轨迹评分:采用评价函数为每条轨迹进行评分;评价指标包括:轨迹末端的位置与目标点位置及到达轨迹末端时的朝向角与目标方位角之间的差距、轨迹上与最近障碍物之间的距离、采样轨迹上的点距离全局路径的最近距离;53)速度发布:对所有轨迹进行评分并选取出评分最高的轨迹;再将评分最高的轨迹对应的速度发给智能体;智能体在下一周期内按照该速度行驶;重复上述过程直至智能体到达自主绕障的目标点。
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210854751.4 | 2022-07-18 | ||
CN202210854751.4A CN115016506B (zh) | 2022-07-18 | 2022-07-18 | 基于自主绕障的异构型多智能体网联协同调度规划方法 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2024016457A1 true WO2024016457A1 (zh) | 2024-01-25 |
Family
ID=83081190
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2022/119207 WO2024016457A1 (zh) | 2022-07-18 | 2022-09-16 | 基于自主绕障的异构型多智能体网联协同调度规划方法 |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN115016506B (zh) |
WO (1) | WO2024016457A1 (zh) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117928852A (zh) * | 2024-01-26 | 2024-04-26 | 烟台海港信息通信有限公司 | 基于深度优先的管网分析方法、系统、终端及存储介质 |
CN118358631A (zh) * | 2024-06-20 | 2024-07-19 | 华侨大学 | 基于离散速度连续时间的多智能体轨道交通路径规划方法 |
CN118550187A (zh) * | 2024-07-26 | 2024-08-27 | 湘江实验室 | 一种预设时间下异构多智能体系全分布式优化方法及装置 |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114815832A (zh) * | 2022-04-28 | 2022-07-29 | 西安交通大学 | 一种基于点云的多智能体超视距网联协同感知动态决策方法 |
CN115016506B (zh) * | 2022-07-18 | 2024-06-14 | 北京大学 | 基于自主绕障的异构型多智能体网联协同调度规划方法 |
CN115396060B (zh) * | 2022-08-30 | 2023-07-14 | 深圳市智鼎自动化技术有限公司 | 一种基于激光的同步控制方法及相关装置 |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2009294934A (ja) * | 2008-06-05 | 2009-12-17 | Ihi Corp | 自律移動装置及び自律移動装置の制御方法 |
CN106251016A (zh) * | 2016-08-01 | 2016-12-21 | 南通大学 | 一种基于动态时间窗的泊车系统路径规划方法 |
CN107727099A (zh) * | 2017-09-29 | 2018-02-23 | 山东大学 | 一种工厂内物料运输多agv调度及路径规划方法 |
CN110174111A (zh) * | 2019-05-31 | 2019-08-27 | 山东华锐智能技术有限公司 | 基于时间窗的任务分段式的多agv路径规划算法 |
CN111367294A (zh) * | 2019-12-27 | 2020-07-03 | 芜湖哈特机器人产业技术研究院有限公司 | 一种激光agv调度控制系统及其控制方法 |
CN113074728A (zh) * | 2021-03-05 | 2021-07-06 | 北京大学 | 基于跳点寻路与协同避障的多agv路径规划方法 |
CN114625140A (zh) * | 2022-03-14 | 2022-06-14 | 北京大学 | 基于改进时间窗的多智能体网联协同调度规划方法及系统 |
CN115016506A (zh) * | 2022-07-18 | 2022-09-06 | 北京大学 | 基于自主绕障的异构型多智能体网联协同调度规划方法 |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108629455B (zh) * | 2018-05-08 | 2021-09-24 | 电子科技大学 | 一种基于车辆自组织网的实时路径规划方法 |
CN110281242B (zh) * | 2019-06-28 | 2022-02-22 | 炬星科技(深圳)有限公司 | 机器人路径更新方法、电子设备及计算机可读存储介质 |
CN112747761B (zh) * | 2020-12-28 | 2022-12-06 | 北京航迹科技有限公司 | 一种基于自动驾驶车辆出行的导航方法和系统 |
-
2022
- 2022-07-18 CN CN202210854751.4A patent/CN115016506B/zh active Active
- 2022-09-16 WO PCT/CN2022/119207 patent/WO2024016457A1/zh unknown
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2009294934A (ja) * | 2008-06-05 | 2009-12-17 | Ihi Corp | 自律移動装置及び自律移動装置の制御方法 |
CN106251016A (zh) * | 2016-08-01 | 2016-12-21 | 南通大学 | 一种基于动态时间窗的泊车系统路径规划方法 |
CN107727099A (zh) * | 2017-09-29 | 2018-02-23 | 山东大学 | 一种工厂内物料运输多agv调度及路径规划方法 |
CN110174111A (zh) * | 2019-05-31 | 2019-08-27 | 山东华锐智能技术有限公司 | 基于时间窗的任务分段式的多agv路径规划算法 |
CN111367294A (zh) * | 2019-12-27 | 2020-07-03 | 芜湖哈特机器人产业技术研究院有限公司 | 一种激光agv调度控制系统及其控制方法 |
CN113074728A (zh) * | 2021-03-05 | 2021-07-06 | 北京大学 | 基于跳点寻路与协同避障的多agv路径规划方法 |
CN114625140A (zh) * | 2022-03-14 | 2022-06-14 | 北京大学 | 基于改进时间窗的多智能体网联协同调度规划方法及系统 |
CN115016506A (zh) * | 2022-07-18 | 2022-09-06 | 北京大学 | 基于自主绕障的异构型多智能体网联协同调度规划方法 |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117928852A (zh) * | 2024-01-26 | 2024-04-26 | 烟台海港信息通信有限公司 | 基于深度优先的管网分析方法、系统、终端及存储介质 |
CN118358631A (zh) * | 2024-06-20 | 2024-07-19 | 华侨大学 | 基于离散速度连续时间的多智能体轨道交通路径规划方法 |
CN118550187A (zh) * | 2024-07-26 | 2024-08-27 | 湘江实验室 | 一种预设时间下异构多智能体系全分布式优化方法及装置 |
Also Published As
Publication number | Publication date |
---|---|
CN115016506B (zh) | 2024-06-14 |
CN115016506A (zh) | 2022-09-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2024016457A1 (zh) | 基于自主绕障的异构型多智能体网联协同调度规划方法 | |
WO2022241808A1 (zh) | 一种多机器人轨迹规划方法 | |
López-González et al. | Multi robot distance based formation using Parallel Genetic Algorithm | |
CN110398967B (zh) | 一种采用离散化方法的多机器人协同轨迹信息处理方法 | |
CN110989352B (zh) | 一种基于蒙特卡洛树搜索算法的群体机器人协同搜索方法 | |
CN111638717A (zh) | 一种分布式自主机器人交通协调机制的设计方法 | |
Guo et al. | Multirobot data gathering under buffer constraints and intermittent communication | |
CN108287545A (zh) | 多移动机器人的冲突管理方法及系统 | |
CN111811511A (zh) | 一种基于降维解耦机制的无人机集群实时航迹生成方法 | |
CN111006667A (zh) | 高速场景下的自动驾驶轨迹生成系统 | |
CN110610512A (zh) | 基于bp神经网络融合卡尔曼滤波算法的无人机目标跟踪方法 | |
CN113900449A (zh) | 多无人机轨迹规划方法、装置、无人机及存储介质 | |
Qiao et al. | Dynamic self-organizing leader-follower control in a swarm mobile robots system under limited communication | |
Chen et al. | Coordinated optimal path planning of multiple substation inspection robots based on conflict detection | |
Wu et al. | Distributed multirobot path planning based on MRDWA-MADDPG | |
Zhou et al. | Multi-robot coverage path planning based on deep reinforcement learning | |
Zennir et al. | Comparison of PID and fuzzy controller for path tracking control of autonomous electrical vehicles | |
Digani et al. | A quadratic programming approach for coordinating multi-agv systems | |
CN115638804B (zh) | 一种无死锁的无人车辆在线路径规划方法 | |
CN115826590A (zh) | 一种自适应参数的移动机器人局部路径规划方法及系统 | |
CN114545971A (zh) | 一种通讯约束下多智能体分布式可飞路径规划方法、系统、计算机设备及介质 | |
Fan et al. | Hierarchical path planner combining probabilistic roadmap and deep deterministic policy gradient for unmanned ground vehicles with non-holonomic constraints | |
CN115657676A (zh) | 基于优先级的集中式多agv多径通道变道决策规划方法 | |
Chen et al. | Asynchronous Spatial Allocation Protocol for Trajectory Planning of Heterogeneous Multi-Agent Systems | |
Xu et al. | A formation keeping approach for fixed-wing uavs with exploration proximal policy optimization |