Neural controller for swimming modes and gait transition on an ostraciiform fish robot

Abstract

This paper proposes a novel control strategy with the central pattern generator (CPG) model for motion control of an ostraciiform fish robot. Compared with previous methods, this method can generate diverse swimming modes and manage gait transition with only two inputs. The CPG enables the robot to achieve agile swimming in the three-dimensional space and to automatically switch between the pectoral and caudal gaits with different speeds. Immediate application of the CPG model is illustrated on a boxfish robot propelled by two pectoral fins and one caudal fin, each of which is individually driven by a motor. With a vision sensor integrated in the CPG-based feedback control, the robot performs a target tracking behavior to further validate stability and availability of the model in real time. Both numerical simulation and laboratory experimental results are provided to validate the effectiveness of the proposed CPG model.