THEORETICAL ANALYSIS ON NEURAL OSCILLATOR TOWARD BIOMIMIC ROBOT CONTROL

Abstract

Neural oscillator is derived from the central pattern generator (CPG) in the biological nervous system. It can generate motor patterns for the rhythmic movements. Neural oscillator is widely adopted in biomimic robot and humanoid robot for different types of rhythmic movement controls such as swimming and walking. Theoretical analysis about neural oscillator toward biomimic robot control is presented in this paper. The methods adopted here include stability theory, describing function, and piecewise linear analysis. Some important properties of the neural oscillator, such as the determination of frequency, oscillation, and stability, are exploited. Network property of multiple neural oscillators is also studied. The insightful results will strengthen the foundation of the neural oscillator and enhance its efficient application for robotic control purpose.