子計畫五：基於群組移動機器人之具自主避障能力動態路徑規劃演算法與協同搬運技術開發能實現人機合作模式之系統

With the development of technology and the development of the Internet, mobile agents have gradually entered the family and social environment to provideservices to meet the needs of the crowd in addition to carrying out the tasks of handling and warehousing in the factory. On the basis of the Internet of Things, smart agents that use versatility and communicate with each other are the future trends. Through the cloud integration service and high-speed wireless transmission technology, the possibility of agent communication is greatly increased, and the agent itself no longer needs a large amount of computing power, which helps to reduce costs and develop the concept of group agents. Compared to the performance of using a single agent, group agents have a great advantage in multiplex, diversity and flexibility, because the coordinated control of group agents can achieve tasks that are difficult to perform in a single agent. For example, in terms of handling, group agents with a large number of advantages can have a larger load and change the operating mode according to different needs to achieve better operational efficiency, so as to avoid the high cost of mobile agents to develop enough conditions to cope with various situations and maintenance costs. However, in order for the agent cooperation mechanism to play an extremely important role in the field of group agents applications, it is necessary to have a good mechanism design in order to exert its due advantages. Therefore, this project focuses on the actual execution ability of the group agents, including the path planning method for designing group agents, the automatic obstacle avoidance method for establishing group agents, and the cooperation mechanism for group agents to carry articles, in order to achieve the goal of human-agent collaboration in the same field. The goals and tasks of this project for first year (2019.08.01~2020.07.31) are finished as the following points: 1.Path planning for the short distance path on node-type vector map by A* algorithm 2.The path correction algorithm applying Trapezoidal/S curve velocity control to keep the continuity of velocity in high-speed turning 3.Path planning for minimum moving cost on the dynamic map 4.The grouping and path planning for the multi-agent system The goals and tasks of this project for second year(2020.08.01~2021.10.31) are finished as the following points: 1.Construct the cooperative handling mode for multi-agent system 2.The design of automatic obstacle avoidance program for multi-agent system