Abstract: In order to deal with the complex work scene and diverse task demands of manipulator task, a manipulator task planning method based on improved graph planning is proposed. Firstly a common mathematical model of manipulator task planning is established, and then combined with the task characteristics of manipulator and the improved simulated annealing algorithm, an improved task planning algorithm based on graph planning is proposed, which extends the single planning result of traditional algorithm to the set of task action sequences. Finally, the task execution strategy is solved by fusion of different targets. A simulation of 7-degree of freedom manipulator verifies the correctness and effectiveness of the proposed method. The results show that compared with the traditional task planning algorithm, the proposed method has the ability to prioritize tasks with different targets and can shorten plan time.

Key words: manipulator, graph planning, simulated annealing algorithm, target fusion