Virtual Motoneuron Activation for Goal-directed Locomotion of a Hexapod Robot

Abstract

Walking animals have shown energy efficiency and versatile locomotion ability when adapting to their environments. Different methods, including machine learning, model control and biologically inspired control, have been applied in artificial legged locomotion systems to imitate these natural characteristics. In this paper, a virtual motoneuron activation module (VMAM) is built to convert goal-directed behavioral instructions into nerve signal strength, which leads to animal-like signal processing mechanisms and behavioral responses. The system is composed of a reticulospinal neuron module, a central pattern generator (CPG) and a virtual motoneuron network (VMN). The VMN receives the rhythm signals from the CPG and generates different motions that are controlled by global and local factors of muscle activation and goal-directed points; then, through an inverse kinematics module, the signals for driving the joint motors of the robot are obtained. Finally, four groups of experiments with different global and local factors of muscle activation are tested on a bionic hexapod robot to validate the effectiveness. The results prove that the method can generate active adaption and emergent behavior according to signal feedback and improve the locomotion agility, coordination, and diversity of the robot.