System For Predicting Anomalies In Joint Reducers Of A Robot And Method Thereof

Abstract: Present disclosure provides technique for predicting anomalies in the joint reducers of the robot. Particularly, data related to only pre-configured parameters is retrieved from the joint reducers and then unique cycles are segregated from the retrieved cycles. Further, the unique cycles may be provided to a learning model for determining the threshold value, in real-time, for a pre-determined number of unique cycles. Thereafter, subsequent cycles may be analysed by the learning model based on the determined threshold for reconstruction error. Such determination of reconstruction error may be provided to a registered user for resolving error/issue in timely manner. In this way, present disclosure avoids complete failure of the whole system and enhances the overall performance of the robot. [Fig.4]

Patent Information

Applicants

Inventors

Specification

Documents

Application Documents