Advanced Motions for Hexapods

Abstract

Mobile robots are able to access hazardous and constricted environment, otherwise impossible for humans, to carry out remote inspection, monitoring, and intervention missions. Legged robots, especially hexapods, provide greater mobility and stability in unstructured environment compared to wheeled robots. However, the motion of hexapods are typically confined to the ground plane, limiting the accessibility of the robot in areas that do not accommodate the footprint of the robot. This research aims to address this limitation by investigating the use of non-planar surface on hexapods. Motions utilising footholds on walls are termed advanced motions in this research. A set of kinematic motion primitives for the five advanced motions considered, namely chimney, chimney corner, wall, wall convex corner and wall concave corner, and the corresponding transitions are first developed. These primitives are then used by the motion planners, namely a grid-based and heuristic planner. The proposed hierarchical grid-based planning framework extends existing approaches to use wall surfaces with the inclusion of the wall and chimney walking primitive. The kinematic primitives analysed for 90 degree corners are used to generate the heuristic motion planner for navigating such corners using chimney and wall walking. Both these motion planners have been verified in simulation. The resulting paths shows the feasibility of using advanced motions in accessing areas previously thought to be inaccessible and for navigating corners. The kinematic motion primitives developed show that advanced motions are kinematically viable for the standard hexapod design with three Degrees of Freedom per leg. The quasi-static motion of chimney walking and wall transition are analysed to identify the joint requirement for executing such motions. The analysis has been verified through a series of experiments demonstrating that a hexapod with a standard design is capable of executing advanced motions.

Date of Award	3 Jan 2020
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Peter Green (Supervisor) & Simon Watson (Supervisor)