Head-orientation for a sidewinding snake robot using modal decomposition

Abstract

Biological snakes exhibit a natural ability to decouple locomotion from the perceptual task of aligning their heads in a particular direction. This same multi-tasking problem is nontrivial for snake robots. A snake robot can mimic the locomotion of their biological counterparts through use of analytic gait expressions, but to orient its head the robot must solve an inverse kinematics problem at every time step. In this work we use modal decomposition to modify a snake robot’s sidewinding gait to orient the head while locomoting. This avoids the problem of determining online inverse kinematic solutions which can be computationally costly. We use knowledge of the robot’s ground contact points to vary the number of joints used for head reorientation to minimally impact locomotion stability. We further show that the resulting expression can be used as a controller to improve real-time target tracking.