Neural-based control of modular robotic fish with multiple propulsors

This paper presents a neural-based control method for a modular robotic fish with a pair of two-degree-of-freedom pectoral fins and a tail fin. A central pattern generator (CPG) model that consists of a network of neural oscillators (NOs) is utilized to generate the target joint angle. The properties of NO network and basic strategies of joint angle control is numerically investigated. Through the coordinated control of the propulsors, a diversity of swimming patterns are implemented. Experiments are conducted to validate the effectiveness of the proposed method.