Neural networks based control for an autonomous underwater vehicle equipped with the collective and cyclic pitch propeller

This paper presents a neural network controller for an autonomous underwater vehicle (AUV) equipped with an innovative collective and cyclic pitch propeller (CCPP). The AUV equipped with a CCPP consists of a new type of propulsion system based on the principle of a helicopter rotor. The dynamics of the AUV with CCPP is briefly described for control design. The main objective of the proposed neural networks based control algorithm is to move the AUV in all directions using only one CCPP with a shaft speed, collective pitch and cyclic angles by carrying out various underwater mission manoeuvres. The proposed control algorithm is applied for numerical simulation study using the recently developed mathematical models of an observation class underwater vehicle platform, namely Gavia, equipped with a CCPP. The work in this paper is a continuity on verification of capability of the AUV with CCPP to move in all directions. The simulation results demonstrate the good performance in the course keeping, changing and trajectory tracking controls using neural network based control algorithm.