ABSTRACT
We report on the development of a neural controller for a three degree of freedom leg of a hexapod. Biological systems are considered as motivation to develop the neural control system. Backpropagation training of multilayer perceptrons and a combination of hetrogeneous neurons are used to implement several pattern generators. Previous work [1] shows that a two degree of freedom leg controller cannot compensate for surface irregularities. The proposed controller for the leg is extended to compensate for such surface irregularities. All the components of the proposed controller are biologically feasible. Simulation results that demonstrate the performance of the leg are presented.
