A Multi-Gait Approach for Humanoid Navigation in Cluttered Environments
作者单位：Control/Robotics Research Laboratory (CRRL)Dept.of ECEPolytechnic Institute of NYU
会议名称：《第26届中国控制与决策会议》
会议日期：2014年
学科分类：080202[工学-机械电子工程] 08[工学] 0804[工学-仪器科学与技术] 0802[工学-机械工程]
关 键 词：Humanoid NAO Biped Obstacle Avoidance Path Planning Low-Profile Gaits Adaptive Gaiting
摘 要：A multi-gait approach is proposed in this paper for autonomous humanoid robot navigation and obstacle avoidance in unknown complex cluttered environments. The proposed approach is based on an environment-dependent adaptive switching between multiple gait strategies including, in particular, a new low-profile crawling gait that enables humanoid motion in tight vertically constrained spaces in addition to forward walking and side-stepping gaits. The path planning and obstacle avoidance system is based on the GODZILA algorithm that provides a computationally lightweight approach for navigation in unknown environments without requiring building of an environment map. The new low-profile crawling gait is laterally symmetric and utilizes a cooperative motion of both the hands and feet. The addition of this gait expands the set of environments that can be handled by the humanoid robot. The efficacy of the proposed approach is demonstrated through simulations and experimental studies on a NAO humanoid robot.