Performance Comparison of Compliance Control based on PI and FLC for safe Human-robot Cooperative Object Carrying

摘要

An important aspect of robotics research is that of human-robot interaction which presents the issue of cooperation between a human and a robot to allow tasks to be shared. This paper highlights on a design of robot control framework used for a human-robot cooperative object carrying task. External force/velocity control was successfully developed to make a robot capable of carrying an object together with a human safely and reliably. The main components of the system consist of a mobile robot which is Jarvis service robot developed by team Dong Yang from Prince of Songkla University, the main computer utilized for data processing and monitoring, and a multi-axis ATI gamma force-torque sensor used to detect the interactive force between human and robot. To ensure effective control, implicit velocity-based force control (external force/velocity control) was successfully implemented, and this control scheme is able to be appropriately associated with PI control and fuzzy logic control (FLC) algorithms. The stability of the system response is evaluated in terms of oscillations in the robot movements, in which the lower the variation in the robot’s movement velocity, the better the performance of the control system. The small oscillation of the robot movement was strategically evaluated in the frequency domain using a Fast Fourier Transform (FFT). The results can be concluded that the performance of the robot force/velocity control based on a PI and FLC algorithms is considered acceptable for the human-robot cooperative object carrying. However, careful observation of the FFT results revealed that the proposes FLC is slightly superior to the PI controller.