Vision-based autonomous perceiving and planning system of a 7DOF robotic manipulator

The main contribution of this paper is the development of a system consist of a 7DOF (degree of freedom) manipulator, a three-finger robot hand and a vision system to complete autonomous environment perceiving and trajectory planning, known as APPS (Autonomous Perceiving & Planning System). The main objective of APPS is to improve intelligence of manipulator though realizing autonomous perceiving and planning for different tasks. The main novelty of this research is that a more stable environment perceiving method by improving the visual process was proposed. The performance tests show that the method could remarkably improve the registration accuracy and calibration validity. Besides, a more accurate and faster autonomous trajectory planning algorithm, combining RRT (Rapid Random Tree) with interpolation, was implemented. The performance of the algorithm was tested through simulations and experiments in real environment, which could decrease the planning time and length significantly and keep an excellent success rate all the time. The results show that APPS could effectively improve the accuracy and efficiency of 7DOF manipulator which would realize higher intelligence and collaboration.