Anomaly Detection Solution 02

Predicting Failures of Robotic Arms with Novelty Detection and Continual Learning

  • Goal

    To increase the productivity at assembly lines by detecting breakdowns of industrial robots ahead of time and minimizing unplanned downtime

  • Our Approach

    We use our proprietary technology to improve the accuracy of novelty detection in predicting the failures of robotic arms. We also apply compatible continual learning algorithms to maintain the novelty detection performance in the rapidly changing environment of the assembly line

  • Results

    Detected robot failures 5 days ahead of time with an accuracy of more than 90%