TAIL ASSISTED DYNAMIC SELF RIGHTING

Abstract

In this paper we explore the design space of tails intended for self-righting a robot’s body during free fall. Conservation of total angular momentum imposes a dimensionless index of rotational efficacy upon the robot’s kinematic and dynamical parameters whose selection insures that for a given tail rotation, the body rotation will be identical at any size scale. In contrast, the duration of such a body reorientation depends upon the acceleration of the tail relative to the body, and power density of the tail’s actuator must increase with size in order to achieve the same maneuver in the same relative time. Assuming a simple controller and power-limited actuator, we consider maneuverability constraints upon two different types of parameters — morphological and energetic — that can be used for design. We show how these constraints inform contrasting tail design on two robots separated by a four-fold length scale, the 177g Tailbot and the 8.1kg X-RHex Lite (XRL). We compare previously published empirical self-righting behavior of the Tailbot with new, tailed XRL experiments wherein we drop it nose first from a 2.7 body length height and also deliberately run it off an elevated cliff to land safely on its springy legs in both cases.\nThis was supported primarily by the ARL/GDRS RCTA and the NSF CiBER-IGERT under Award DGE-0903711.\nFor more information: Kod*Lab