ROHM introduces AI-equipped MCUs for on-device fault prediction
ROHM has developed AI-equipped microcontrollers (MCUs), the ML63Q253x-NNNxx and ML63Q255x-NNNxx, designed for fault prediction and degradation forecasting in devices such as industrial motors, residential facilities, and home appliances. These MCUs, announced on June 4, 2025, are the first to perform both learning and inference on-device without network connectivity, as per a ROHM study.
The MCUs address the need for real-time equipment monitoring while avoiding network latency and security risks associated with traditional AI models that rely on cloud connectivity and high-performance CPUs. They use a 3-layer neural network algorithm, Solist-AI, enabling standalone AI processing for anomaly detection and maintenance co...
You've read this far — sign in to keep reading