Neuromorphic vision based control for the precise positioning of robotic drilling systems

• DOI: 10.1016/j.rcim.2022.102419
• Publication Date: 2022-08-09T16:39:16Z
• Authors: Abdulla Ayyad; Mohamad Halwani; Dewald Swart; Rajkumar Muthusamy; Fahad Almaskari; Yahya Zweiri
• Keywords: Neuromorphic engineering; Machine Vision; Visual Servoing

The manufacturing industry is currently witnessing a paradigm shift with the unprecedented adoption of industrial robots, and machine vision key perception technology that enables these robots to perform precise operations in unstructured environments. However, sensitivity conventional sensors lighting conditions high-speed motion sets limitation on reliability work-rate production lines. Neuromorphic recent potential address challenges its high temporal resolution, low latency, wide dynamic range. In this paper for first time, we propose novel neuromorphic based controller robotic machining applications enable faster more reliable operation, present complete system capable performing drilling tasks sub-millimeter accuracy. Our proposed localizes target workpiece 3D using two stages developed specifically asynchronous output cameras. stage performs multi-view reconstruction an initial estimate workpiece’s pose, second refines local region circular hole detection. robot then precisely positions end-effector drills holes combined position-based image-based visual servoing approach. solution validated experimentally nutplate workpieces placed arbitrarily environment uncontrolled lighting. Experimental results prove effectiveness our maximum positional errors less than 0.2 mm, demonstrate use overcomes speed limitations findings identify as promising can expedite robustify processes line requirements fourth revolution.