Research On Configuration Optimization And Automatic Decomposition Of Dynamics For Constrained Reconfigurable Manipulator

With the development of science and technology, the traditional robot has been difficult to meet the needs of human beings. The robot with single function can not complete a variety of types of work, and the design of robots with different configuration for different tasks will be a waste of resources. Therefore, more intelligent, more intelligent reconfigurable robot emerge as the times require. The reconfigurable robot is formed by a variety of different functional module, it is a robot which can change the configuration according to the change of the task or the environment. According to the concept of modular design, reconfigurable robot has a good flexibility and dexterity. It shows good ability in the complexity of the environment, or high accuracy requirements of the task. Reconfigurable robots have been widely used in medical, scientific exploration, space exploration, dangerous environment and other fields. There are still many challenging issues about reconfigurable robot to be further in-depth study. Such as the configuration optimization based on task reachability, the ride control in terminal and environmental contact, the design of modular joint torque observer, the active fault tolerant of multiple faults and so on. Above on there are far-reaching theoretical and practical significance of more in depth study of reconfigurable modular robot.In this paper, the automatic decomposition of dynamics and configuration optimization of the reconfigurable manipulators have been research in theory and experiment. First, forward kinematics of the reconfigurable manipulators was analyzed based on the exponential product formula, and then a modeling method for the forward and inverse kinematics of reconfigurable manipulator is presented based on the geometric method is given, D-H equation method, the optimization method of D-H equation. Thus establishing the transformational relation model of the reconfigurable manipulators between the joint space and work space. The dynamic model of the reconfigurable manipulators in free space was set up combining with the screw theory and Newton- Euler equation. The transformational relation for contact force between the end-effector and each joint was analyzed using the Jacobi matrix. In the basis of that, dynamic model of the reconfigurable manipulators in constraint space was set up. Aiming at the problem of the configuration optimization of reconfigurable manipulator under the constraint of joint motion. The problem is transformed into the shortest path problem of the directed connected graph, and a new optimization method based on Dijkstra algorithm and genetic algorithm is designed to solve the problem. Multi objective configuration optimization under energy and time optimization is realized. And the optimization of the 3D simulation is verified. Finally, the optimization algorithm is verified by the EV-MRobot series reconfigurable manipulator, The control system of the reconfigurable manipulator is debugged and analyzed.At the end of the paper, the results obtained and the lessons are summarized, and the future research is prospected.