In-Vehicle Motor Failure Monitoring Using Clustering and AI

The patent extracts the core PHM analysis functionality from external cloud servers and implements it locally within the vehicle using an onboard processor. The sensing module collects motor data, the processor performs clustering analysis and failure prediction locally, and only essential results are transmitted to external servers. This extraction eliminates network dependency for real-time diagnosis while maintaining advanced analysis capabilities.

The patent merges multiple functional modules into a single integrated system chip. The sensing module for data acquisition, the processor for clustering analysis, the memory for storing motor data and model information, and the communication interface are all integrated into one compact unit. This consolidation reduces device complexity compared to distributed architectures while enabling real-time diagnosis through onboard processing.

The patent applies clustering algorithms selectively to the most critical motor parameters rather than processing all available data. The system identifies key features from motor sensing data (such as current, voltage, temperature) and applies k-means clustering only to these essential parameters for failure prediction. This partial application of complex algorithms maintains high prediction accuracy while significantly reducing computational energy consumption compared to comprehensive AI processing of all sensor data.