Walking-in-Place Foot Interface for Locomotion Control and Telepresence of Humanoid Robots

Abstract

We present a foot-tapping walking-in-place type locomotion interface and algorithm to generate high-level movement patterns for humanoid remotely controlled robots. Foot tapping motions acting on the platform are used as movement commands to remotely control locomotion of a humanoid robot. We describe two separate motion mapping algorithms suitable for wheeled and bipedal humanoid locomotion. Our interface enables remote locomotion control of humanoid robots with the help of a seated and hands-free interface and enables the use of both handheld or desktop-based interfaces for manipulation tasks. An experimental study with eight participants controlling walking speed of a virtual robot was conducted to explore if the participants could maintain distance (1-3m range) to a reference target (leading robot) moving at different speeds. All participants were able to use the proposed interface to track the leading robot efficiently for walking speeds of less than 1 m/s and an average tracking error was 0.47 m. We discuss the results of the study along with the NASA TLX and system usability surveys.