A Biologically Inspired Neural Network for Autonomous Underwater Vehicles

Abstract

Autonomous underwater vehicles (AUVs) have great advantages for activities in deep oceans, and are expected as the attractive tool for near future underwater development or investigation. However, AUVs have various problems which should be solved for motion control, acquisition of sensors' information, behavioral decision, navigation without collision, self-localization and so on. This paper proposes an adaptive biologically inspired neural controller for trajectory tracking of AUVs in nonstationary environment. The kinematic adaptive neuro-controller is an unsupervised neural network, which is termed Self-Organization Direction Mapping Network (SODMN). The network uses an associative learning system to generate transformations between spatial coordinates and coordinates of propellers' velocity. The neurobiological inspired control architecture requires no knowledge of the geometry of the robot or of the quality, number, or configuration of the robot's sensors. The SODMN proposed in this paper represents a simplified way to understand in part the mechanisms that allow the brain to collect sensory input to control adaptive behaviours of autonomous navigation of the animals. The efficiency of the proposed neurobiological inspired controller for autonomous intelligent navigation was implemented on an underwater vehicle capable of operating during large periods of time for observation and monitoring tasks.