IoST-Enabled Robotic Arm Control and Abnormality Prediction Using Minimal Flex Sensors and Gaussian Mixture Models

摘要

This work presents a groundbreaking approach with a fusion of the Internet of Sensing Things (IoST) and Robotics. This system utilizes four flex sensors strategically placed on the most flexible fingers across both hands to control a Six-DoF robotic arm, offering a novel interface for those with limited mobility. This system can also be used for moving toxic objects. The Raspberry Pi is the central control unit that acquires data from the flex sensor and controls the servo motors. Moreover, the device incorporates machine learning to learn the daily movements of the users and predict abnormal finger movements. Multiple data analyses and visualization are initiated to predict the normal and abnormal data. GMMs or Gaussian Mixture Models showed successful results among various abnormality detection processes. This amalgamation of flexible sensing and mathematical modeling offers precision and adaptability in control mechanisms.