Strategies for Dealing with Problems in Robotised Unscrewing Operations

Abstract

Disassembly is the first step in a remanufacturing process chain with unscrewing being usually the most frequent task. In previously reported research in the authors’ laboratory, a new method has been developed for using robots to unfasten screws. Uncertainties and variability in the physical condition of screws induced by dirt, rust, or mechanical damage pose difficulties for such robotised unscrewing systems. There are three common failure modes: screwdriver missing screw head, screwdriver slipping on screw head and screw too tight to remove. This paper presents strategies to handle these failure modes, making the developed robotised method more robust and reliable. The strategies include conducting a second search and second unfastening trial as well as involving collaboration with a human operator. Tests were carried out to validate the proposed strategies. The results show that the strategies could deal with the failure modes, enabling 100% successful operation.

Robust Fastener Detection Based on Force and Vision Algorithms in Robotic (Un)Screwing Applications

“…The authors of [15] proposed a computer vision framework using a combination of neural networks for detecting screws and recommending related tools for disassembly. Reference [16] proposed unscrewing strategies that include conducting a second search and a second unfastening trial, as well as involving collaboration with a human operator.…”