Ball-on-plate motion planning for six-parallel-legged robots walking on irregular terrains using pure haptic information. (November 2019)

• Record Type: Journal Article
• Authors: Chen, Zhijun; Gao, Feng; Sun, Qiao; Tian, Yuan; Liu, Jimu; Zhao, Yinan
• Is Part Of: Mechanism and machine theory. Volume 141(2019)
• Page Start: 136
• Page End: 150
• Publication Date: 2019-11
• Subjects: Ball-on-plate -- Motion planning -- Six-parallel-legged robot -- Irregular terrain -- Pure haptic information -- Impedance control; Machine theory -- Periodicals; Machinery -- Periodicals; Machines -- Périodiques; Génie mécanique -- Périodiques; Machine theory; Machinery; Periodicals; 621.81
• DOI: 10.1016/j.mechmachtheory.2019.07.009
• Languages: English
• ISSNs: 0094-114X

Abstract

Highlights: Ball-on-plate problem is developed on a six-parallel-legged robot. The robot walks on irregular terrain and balances a ball on a plate simultaneously. Ball state is estimated in real time by pure haptic information. Motion planner with five sub-planners is developed for all legs in all phases. Experiments on prototype validates the effectiveness and robustness of the method. Abstract: Current studies on the ball-on-plate (BoP) problem are mainly focused on platforms with a fixed base and use a vision system to detect the ball. However, relevant works on legged robots are not available, especially on six-parallel-legged robots using pure haptic information. This paper proposes a novel BoP motion planning method for six-parallel-legged robots walking on irregular terrain based on pure haptic information. First, a six-parallel-legged robot with seven six-dimensional force/torque sensors is introduced. Then, a complete model of the BoP system on the robot is built by considering the six-degree-of-freedom motion of the body. Further, a measurement model of the ball state based on pure haptic information is built, and a low-pass filter, lead compensator, and Kalman filter are applied to extract useful information from the noisy sensor data. An impedance controller is proposed to control the ball, and a novel motion planner with five sub-planners is introduced to address the problem of walking on irregular terrain. Finally, experiments are conducted on the prototype to validate the method. The results show that the method performs robustly when the robot traverses irregular terrains.