Proposal of a Monitoring System for Collaborative Robots to Predict Outages and to Assess Reliability Factors Exploiting Machine Learning

Industry standards pertaining to Human-Robot Collaboration (HRC) impose strict safety requirements protect human operators from danger. When a robot is equipped with dangerous tools, moves at high speed or carries heavy loads, the current legislation requires continuous on-line monitoring of robot’s and suitable separation distance workers. The present paper proposes make virtue out necessity by extending scope predicting failures safe stops. This has been done implementing platform, based on open access tools technologies, monitor parameters during execution collaborative tasks. An automatic machine learning (ML) tool edge network can help perform predictions possible outages robots, especially as consequence human-robot interactions. By exploiting system, it increase reliability work, eliminating any unplanned downtimes tasks, maximising trust in interactions increasing lifetime. proposed framework demonstrates data management technique industrial robots considered physical cyber-system. Using an assembly case study, have collected fed ML model order identify most significant factors predict stops robot. Moreover, acquired study used manipulator’ joints; cobot autonomy provide predictive maintenance notifications alerts end-users vendors.