Neural network architectures for redundancy resolution of robotic manipulators

This paper presents a neural-network based computational scheme for redundancy resolution of manipulators. The Tank-Hopfield (TH) network is adopted for pseudoinverse and inverse kinematics calculations and it can provide joint velocity and joint acceleration solutions within a time frame of the order of hundred nanoseconds. The proposed TH network consists of a left-hand network and a right-hand network. The connection weights of the network can be directly obtained from the known matrix J (Jacobian matrix) and vector X˙ (end-effector velocity) or vector X¨ (end-effector acceleration). Incorporating the TH network into the redundancy resolution scheme allows planning algorithms to be implemented in real time. Simulation results for a 3-link planar manipulator are presented to demonstrate that the proposed approach is efficient and practical