Locomotion Study of a Hyper-redundant Modular Robot Using Artificial Neural Networks

Abstract

This paper deals with the study of locomotion of a hyper-redundant wheel-less modular robot. The robot has the capability of operating in a horizontal plane through the implementation of proposed locomotion, which purely depends on body undulations. Since, the body undulation takes place in two orthogonal planes simultaneously through the application of sinusoidal Joint Orientation Functions (JOFs), the contact points between the body and surface change rapidly, which makes it more complex to analyze the kinematic behaviour. At this juncture, the applicability of the ANN (Artificial Neural Network) model has been evaluated to predict the kinematic behavior (primarily the net displacement) of the robot for a given set of input JOF parameters while performing a certain locomotion. The model is trained and validated with a significant number of experimental data set and finally the simulated outputs are verified by comparing with some pilot studies.