Start-up and gait transitions in bio-inspired fish robots via modified CPG with coupled Hopf oscillators

Abstract Bio-inspired fish robots equipped with fin membranes, also referred to as amphibious swimming robots, are capable of both aquatic and terrestrial locomotion through undulating fin propulsion inspired by biological systems. This study introduces a modified central pattern generator (CPG) network based on Hopf-type oscillators to preserve natural swimming patterns during start-up, parameter transitions, and gait changes. Simulation and experimental results confirm the effectiveness of the proposed approach, demonstrating that the fin membrane maintains consistent undulatory motion when starting from arbitrary initial swimming gaits, as validated by an experimental start-up sequence consisting of a 15 s start-up, 5 s stop, and re-start. In contrast, conventional Hopf-type oscillator-based CPG networks require all fin rays to initiate motion from the zero position. Furthermore, the proposed CPG network generates smooth outputs with continuous position, velocity, and acceleration during oscillatory parameter adjustments, enabling seamless transitions between different swimming gaits despite the geometric constraints of the fin membrane. Experimental validation also confirms a successful transition from a linear swimming gait with n = 3.5 (the maximum number of wavelengths achievable with the eight fin-ray robot model) and oscillatory amplitude A = 10.3°, to another gait with n = 1 and A = 36° within approximately 270 s.