Contact event detection for robotic oil drilling

摘要: To ensure safe and reliable operation in a robotic oil drilling system, it is essential to detect contact events such as impacts slips between end-effectors workpieces. In this challenging application, where high forces are used manipulate heavy metal pipes noisy environments, acoustic emissions (AE) sensors offer promising sensing solution. Realtime AE signal features create multinomial event classifier. The sensitivity of variety including two types slip presented. Results indicate that the classifier able robustly dynamically classify with >90% accuracy using small set features.