A Data-driven Method for Monitoring Systems that Operate Repetitively -Applications to Wear Monitoring in an Industrial Robot Joint1

This paper presents a method for monitoring of systems that operate in a repetitive manner. Considering that data batches collected from a repetitive operation will be similar unless in the presence of an abnormality, a condition change is inferred by comparing the monitored data against a nominal batch. The method proposed considers the comparison of data in the distribution domain, which reveals information of the data amplitude. This is achieved with the use of kernel density estimates and the Kullback-Leibler distance. The method is simple to implement and can be used without process interruption, in a batch manner. The method was developed with interests in industrial robotics where a repetitive behavior is commonly found. The problem of wear monitoring in a robot joint is studied. Real data from accelerated wear tests are considered. Promising results are achieved, where the method output shows a clear response to the wear increases.