Design and Optimization of Hydraulically Actuated Hexapod Robot COMET-IV

Abstract

The development of COMET-III resulted in a completely self-contained drive system that closely approximated a practical robot. However, various problems emerged in the course of research and development. In general, there was significant scope for improvement in terms of adaptability to terrain and speed of movement. For example, owing to an insufficient amount of oil and poor durability, sustained tripod walking could not be achieved, and the achieved walking speed was slow; the possible range of motion of the legs, which lacked the ability to move sideways or diagonally, was small, and the robot could not move omnidirectionally. In particular, the preeminence of legged robots as locomotive robots is ascribed to their superior capability of discrete walking in specific environments (such as minefields) and outstanding ability in general to adapt to the terrain. These capabilities enable legged robots to easily move over difficult and uneven terrain—even in environments wherein crawler robots and wheeled robots are incapable of motion. Therefore, there is an urgent need to overcome the fatal flaws in COMET-III—for example, in terms of terrain adaptability and speed of movement.

To this end, Nonami’s team combined their various achievements to date and commenced the development of COMET-IV auxiliary robot, which is aimed at actual mine detection and removal. This chapter describes the group’s attempt to fundamentally review the basic specifications of a robot, such as the mechanism, gait, drive system, and control system, and approach the optimization-based design of COMET-IV—the hexapod dangerous-operations robot. First, based on the conceptual design of legged robots, the superiority of COMET-IV was clarified in terms of system construction, and the overall design of the robot was arrived at. Then, a single-legged robot was constructed and tested for speed and load endurance for analyzing robot performance both theoretically and experimentally. Based on the obtained test results, the design of the robot leg mechanism—essential to a walking robot—was theoretically optimized by means of iterative Jacobian-analysis-based complex evaluations. Further, a practical hexapod robot system that uses walking control for navigating uneven terrain was established by implementing force control or impedance control of end-effectors at any attitude.