Abstract
The autonomous distributed system (ADS) is one of the attractive approaches for more versatile and autonomous robot systems. The present paper proposes a distributed trajectory generation method for redundant manipulators through cooperative and competitive interactions between subsystems of ADS, which is based on a concept of virtual arm as each subsystem. The virtual arm has the same kinematic structure as the manipulator except that its end-point is located on the joint or link of the manipulator. Therefore the redundant manipulator can be represented by a set of the virtual arms.
In the present paper, each subsystem corresponding to the virtual arm is implemented by a back propagation typed neural network. It is shown that the nonlinear inverse kinematics of the virtual arms can be solved through exchanging position error signals about each virtual end-point computed in each neural network. The method can generate the trajectories of redundant manipulators using only the information in the task space, and each desired end-point position of virtual arms can be planned independently. Moreover, it is robust to some failures of subsystems or actuators, since each subsystem can work fully autonomously. Finally, the effectiveness of the method is verified by computer simulations using a planar manipulator with redundant joint degrees of freedom.
