Adaptive Climbing Gait Design of Snake Robot Based on Extended Kalman Filter

Abstract

We propose a method of climbing pipe with different diameter, which enables snake robot to divide itself into 3 parts to complete transition on the pipe according to the joints’ state for now. To avoid controlling the robot from building dynamic model, a gait-based open-loop control framework is adopted. An extended kalman filter (EKF) is applied here to estimate the adaptive parameter of gait-based function for different 3 parts and corresponding joints. The adaptive parameter will help snake robot perceive the environment better and climb the pipe more steadily.