Fish School System Identification and Control Based on Artificial Neural Network

Abstract

The purpose of this paper is to introduce Artificial Neural Network (ANN) based system identification for a nonlinear dynamics structure of fish school for trajectory tracking problem. A discrete time dynamic model has been derived, and the neural network (NN) training data has been collected experimentally using a Simulink model within MATLAB program designed for this purpose. This work also seeks to design a user friendly program to automatically monitor and track the trajectory of the live fish in an aquarium using two webcams to produce a three dimensions (3D) view of the acquired trajectory. MATLAB program was used to develop a graphical user interface (GUI) which allows the monitoring user to input experiment specifications such as desired tracking time, tracking region and video source (real time directly from webcams vs. cached from the pre-saved files) in addition displaying the acquired videos from two webcams, position of the tracked fish (trajectory), and fish velocity. A predictive control approach based on NN is proposed to provide the control signal(s) that could be used to control the robotic fish to achieve trajectory tracking task in an attempt to design and build autonomous robotic fishes in future that are able to reactive to the environment and navigate toward the desired location underwater.