Bipedal robot locomotion using multivariable control

Abstract

In this research, multivariable PID control was implemented on YICAL Leg II for self-supporting bipedal locomotion. The legged robot is properly scaled to a human being. It has six DOF which is direct-driven by DC motors. The motion was limited on the sagittal plane. The sensor system of the robot is comprised of precision potentiometers for joint angle feedback and an IMU for attitude estimation. The biped has a body made up of fiberglass which is a light-weight and an elastic material. The predefined trajectories were input to the basic PID control and performance measures such as maximum error of 17.19%, maximum settling time of 1.38 seconds and maximum percent overshoot of 3.1% were achieved. Even though there were deviations from the proposed PID performance measures, the bipedal legs still achieved a walking motion.