Walking pattern generation algorithm with waist motion for improving energy efficiency of quadruped robot

This paper deals with a walking pattern generation algorithm with waist motion for improving the energy efficiency of a quadruped robot. Most quadruped robots without a waist joint have problems that walking is unnatural compared with quadruped animals, and their energy efficiency is low owing to the limited degrees-of-freedom of the body. To solve these problems, we studied the walking pattern generation of a quadruped robot with a waist joint. First, we selected walking parameters and generated foot trajectories by referring to the actual walking data on a dog. Second, we generated the body trajectory using the preview control to maintain walking stability. Third, we realized a quadruped walking pattern generation algorithm by integrating a waist motion generation strategy capable of improving energy efficiency and animal mimicry. Finally, the walking stability and energy efficiency improvements of the quadruped robot were verified through the Webots dynamics simulator.