EDGE AI-DRIVEN CONDITION MONITORING AND FAULT DIAGNOSIS OF ELECTRIC VEHICLE MOTORS: A REAL-TIME, DATA-DRIVEN APPROACH

Abstract

The rapid growth of electric vehicles (EVs) has created a strong demand for reliable and efficient motor condition monitoring systems to ensure safety, performance, and longevity. Conventional fault diagnosis methods, including rule-based and cloud-dependent approaches, often face challenges such as high latency, limited accuracy, and dependency on continuous network connectivity. To overcome these limitations, this study proposes an Edge AI-driven, real-time, data-driven framework for condition monitoring and fault diagnosis of EV motors. The system integrates multi-modal sensor data, such as vibration, temperature, and current signals, collected through IoT-enabled embedded platforms. Advanced preprocessing techniques, including noise filtering, normalization, and feature extraction, are employed to improve data quality. Lightweight deep learning models, such as CNN and LSTM, are optimized for deployment in resource-constrained edge environments to enable real-time inference and rapid fault detection. The framework also incorporates secure communication mechanisms for periodic cloud synchronization, enabling model updates and fleet-level analytics. The proposed approach is expected to deliver high diagnostic accuracy, reduced latency, and robust performance under environmental noise and intermittent connectivity conditions. It facilitates early fault detection, predictive maintenance scheduling, and autonomous fault mitigation, thereby reducing downtime and maintenance costs. Overall, the study highlights the potential of Edge AI in developing scalable, efficient, and intelligent EV motor monitoring systems, contributing to enhanced operational reliability and extended motor lifespan.