This paper presents the process of translating formal theory and methods to efficient algorithms in the context of human-inspired control of bipedal robots, with the end result being experimentally realized robust and efficient robotic walking with AMBER. We begin by considering human walking data and find outputs (or virtual constraints) that, when calculated from the human data, are described by simple functions of time (termed canonical walking functions). Formally, we construct a torque controller, through model inversion, that drives the outputs of the robot to the outputs of the human as represented by the canonical walking function; while these functions fit the human data well, they do not apriori guarantee robotic walking (due to do the physical differences between humans and robots). An optimization problem is presented that determines the best fit of the canonical walking function to the human data, while guaranteeing walking for a specific bipedal robot; in addition, constraints can be added that guarantee physically realizable walking. We consider a physical bipedal robot AMBER and define a simple voltage based control law--utilizing only the human outputs and canonical walking function with parameters obtained from the optimization--for which we obtain walking in simulation. Since this controller does not require model inversion, it can be implemented efficiently in software. Moreover, applying this methodology to AMBER experimentally results in robust and efficient "human-like" robotic walking.
展开▼
