Development of vision-based object tracking fish robot

Abstract

This paper focuses on the development and vision-based object tracking control of a fish robot. Its efficiency towards target tracking is demonstrated in a water tank. Inspired by the maneuvering and stability of Tuna fish, the fish robot is designed by a single actuator located at the rare part of the fish body. The precise maneuverability characteristics of the fish robot is achieved by the motion of a caudal fin. A servomotor is involved to control the oscillation of the caudal fin. A vision-sensor is integrated into fish robot to gain information related to Cartesian coordinates of the targeted object by implementing a color-based filtering algorithm. An object tracking algorithm is designed which performs the decision-making task for the fish robot while identifying and following the targeted object. The locomotion of the fish robot in various directions is tested experimentally by achieving effective performance of the proposed targeted object following task.