A new bionic lateral line system applied to pitch motion parameters perception for autonomous underwater vehicles

Abstract

Attitude perception for autonomous underwater vehicles (AUVs) is a challenging task. Most of existing sensor systems are unable to detect the navigation attitude of AUVs while at the same time to identify the attitude geometric parameters. In order to bridge this research gap, a new artificial lateral line sensor (ALLS) system based on a pressure sensor array is proposed to perform pitch motion perception for AUVs. First of all, a boxfish-like robot was constructed based on the geometrical morphology of a boxfish. The proposed ALLS system was fabricated in the fish robot. Then sensing experiments in the conditions of different pitch motions of the robot were conducted and the experimental measurements were compared with numerical simulation results. The comparison showed consistent mechanisms between the numerical bionic sensor model and the ALLS in perceiving the pitch attitude of the fish robot. Subsequently, pressure measurements of the fish robot were processed by popular machine learning algorithms such as random forest and artificial neural network to establish a vehicle pitch identification model. The analysis results demonstrate that the developed ALLS system is effective in pitch motion parameters recognition, which may provide a new way for self-attitude perception and adjustment of AUVs.