Collision Localization for a Robotic Arm Based on Fiber-Optic Sensing and Machine Learning

摘要

In recent years, robotic arms have become universal tools across various industries. To ensure regular operation and prevent severe damage, a reliable collision detection system is an indispensable component for industrial robotic arms. This article introduces a collision monitoring system for a robotic arm based on fiber-optic sensing. The system employs distributed fiber Bragg grating (FBG) sensors as the sensing element. These sensors are positioned at multiple locations on the robotic arm to measure the vibration response during collision events. By analyzing the characteristics of these vibration signals and using machine learning algorithms, the system can precisely determine the locations of collisions. Furthermore, different machine learning algorithms are applied to detect collision locations for comparison purposes. Various feature extraction processes and machine learning algorithms are also compared. As a result, the prediction accuracy for the artificial neural network (ANN) algorithm is about 90.56%. The system can accurately identify the location of collisions.