Based on in-vehicle microphones and acceleration sensors, OHm performs edge computing on sound and vibration signals, analyzes and diagnoses vehicle health status in real time, and reports the relevant data through the connected vehicles. The noise and vibration during the vehicle operation are random in nature, so signal filtering and noise reduction processing are required. OHm extracts abnormal sound features, establishes and trains failure models to achieve robust vehicle health monitoring and management even under various road conditions.