Compliant Control using Force Sensor for Industrial Robot

This study presents an innovative approach to enhance intuitive control in human-robot collaboration scenarios. It focuses on addressing challenges related to collisions that can lead to undesirable robot behavior, such as rebounding and trajectory deviations. To tackle these issues, the study proposes a teaching and control system for collaborative robots, enabling operators to have more flexible control while effectively mitigating instability caused by collision-induced rebound. The main approach is the human-robot collaboration collision index, which continuously collects force data through a force sensor. This collision index is crucial in distinguishing between normal operations and collisions, quantifying the severity of collisions. When a collision is detected, the system dynamically adjusts the robot's movement commands, rapidly increasing the gain during collisions to suppress rebound. This ensures that the robot's end-effector remains at the target position until it stabilizes before returning to normal control gain. Experimental validation was conducted using three different values of the minimum control gain (K_min), resulting in varying times taken by participants to move the robotic arm. The study found that a K_min value of 0.8 kN/m yielded consistent and efficient performance with lower variability, making it a promising solution for improving the efficiency and precision of robotic arm operations, particularly in assembly applications.