The Sampling Task Planning of On-board Manipulator Based on Planning Knowledge Base
Lu Yao, Qingxuan Jia · 2019 IEEE 2nd International Conference on Automation, Electronics and Electrical Engineering (AUTEEE) · 2019
Task planning is used extensively in robot fields to help it complete some complicate tasks in a certain situation. Traditional task planning use path planning algorithm to get a collision-free and optimal resource allocation path to guide the manipulator to execute the task safely and efficiently, which is unable to memorize and learn from the past planning result, and thus severely affected the planning efficiency. In this paper, a planning method based on knowledge base is put forward to improve efficiency in task planning. Firstly, the complex sampling task is divided into several atomic tasks that can be recognized and executed by manipulator directly. Then, the knowledge base is introduced into PRM to improve the planning efficiency in the process of repeating planning. Then a similarity matching method of picture based on pHash is raised to match planning graph in knowledge base. Finally, the simulation process performed by matlab proved the feasibility and correctness of this method.