Design and control of an embedded vision guided robotic fish with multiple control surfaces.

Junzhi Yu,Jianwei Zhang,Min Tan,Kai Wang

doi:10.1155/2014/631296

Abstract

This paper focuses on the development and control issues of a self-propelled robotic fish with multiple artificial control surfaces and an embedded vision system. By virtue of the hybrid propulsion capability in the body plus the caudal fin and the complementary maneuverability in accessory fins, a synthesized propulsion scheme including a caudal fin, a pair of pectoral fins, and a pelvic fin is proposed. To achieve flexible yet stable motions in aquatic environments, a central pattern generator- (CPG-) based control method is employed. Meanwhile, a monocular underwater vision serves as sensory feedback that modifies the control parameters. The integration of the CPG-based motion control and the visual processing in an embedded microcontroller allows the robotic fish to navigate online. Aquatic tests demonstrate the efficacy of the proposed mechatronic design and swimming control methods. Particularly, a pelvic fin actuated sideward swimming gait was first implemented. It is also found that the speeds and maneuverability of the robotic fish with coordinated control surfaces were largely superior to that of the swimming robot propelled by a single control surface.

Highlights

There has been a rapid increase in the research and development of the bioinspired mechatronic system in academia and industry over the last 20 years, due to the increasing vitality in biomimetics [1]
The propulsion modes of swimming fish, from a functionality standpoint, are classically divided into body and/or caudal fin propulsion (BCF) and median and/or paired fin (MPF) propulsion [8]
Fish classes that use varying degrees of body undulation and/or caudal fin oscillations for thrust generation are examples of the BCF mode, while fish relying on paired fins and/or median fins for thrust generation are categorized under the MPF mode

Summary

Introduction

There has been a rapid increase in the research and development of the bioinspired mechatronic system in academia and industry over the last 20 years, due to the increasing vitality in biomimetics [1]. By understanding and adapting the underlying principles of swimming motions of aquatic animals (e.g., fish and cetaceans) to artificial swimming machines (hereafter termed robotic fish), an enhanced comprehensive performance has been obtained [3,4,5,6,7]. The propulsion modes of swimming fish, from a functionality standpoint, are classically divided into body and/or caudal fin propulsion (BCF) and median and/or paired fin (MPF) propulsion [8]. Fish classes that use varying degrees of body undulation and/or caudal fin oscillations for thrust generation are examples of the BCF mode, while fish relying on paired fins (pectoral fins and pelvic fins) and/or median fins (dorsal fin and anal fin) for thrust generation are categorized under the MPF mode. It is generally thought that no absolutely superior model exists among these modes since body shape and motor function level closely depend on the fish habitats

Methods

Results

Discussion

Conclusion