Hardware implementation of a neural network controller on FPGA for a humanoid robot arm

Abstract

This paper presents the design and the implementation of the neural network controller for a humanoid robot arm, the ROBOKER whose purpose is to work on behalf of of human workers. The radial basis function network is implemented on a field programmable gate array(FPGA). The design of the floating point processor allows us to design the back-propagation learning algorithm for the neural controller. To test the functionality of the designed neural network control hardware, experiments of controlling a humanoid robot arm are conducted. Performances of the neural controller are compared with those of the PD controller.