Model-based Force Control of a Tendon-Sheath Actuated Slender Gripper Without Output Feedback

Zhang Qi,Shen Donghua,Tian Mengqian,Wang Xingsong

DOI: https://doi.org/10.1007/s10846-022-01785-z

2022-01-01

Journal of Intelligent & Robotic Systems

Abstract:Slender robots have received a lot of attention in ruin rescue missions as they are flexible to detect narrow spaces. However, the ability to perform rescue operations is not satisfactory, which is greatly subject to the control performance of the end manipulators. Unfortunately, to adapt to the small ruined environment, the end manipulators of these robots usually lack enough sensors, moreover, the complex and unstable ruined environment further makes them hard to control. To dress this, we use tendon-sheath mechanisms to develop a slender robot with an elastic end gripper. In this paper, we propose a precise control method for the end gripper without output sensory feedback. Firstly, the gripper drive mechanism is simplified as a single tendon-sheath system with a spring load. After that, a tendon-sheath force transmission model is built. On this basis, we propose a force control method for the end gripper, it is composed of model-based friction compensation and adaptive sliding mode control. The transmission model and the force control method are then validated by experiments. The experimental results illustrate that the force transmission model can predict the distal force with high accuracy, the R2 (Coefficient of Determination) is over 0.993 and the RMSE (Root Mean Square Error) is below 0.60 N. The force control method can track the desired force precisely, the R2 is over 0.989 and the RMSE is below 0.74 N.