The Role of On-Device AI in Next-Generation Industrial and EV Systems
The growing demand for intelligent, energy-efficient systems has sparked a shift toward on-device artificial intelligence, particularly in predictive maintenance and real-time monitoring. Industries such as manufacturing, home automation, and electric vehicles (EVs) are increasingly adopting edge AI to overcome the latency, bandwidth, and security concerns associated with cloud-dependent systems. By processing data locally, on-device AI enables faster decision-making, reduced operating costs, and improved safety. In electric vehicles, this technology is poised to enhance battery management, motor control, and in-vehicle diagnostics by predicting component failures before they occur. ROHM’s recent innovation aligns with this trend, delivering a breakthrough in standalone AI microcontroller units (MCU) capable of performing both training and inference directly on the hardware.
Breakthrough Standalone AI MCUs for Industrial and EV Applications
ROHM has unveiled the world’s first AI-equipped MCUs capable of independent on-device learning and inference without relying on cloud connectivity. The ML63Q253x and ML63Q255x MCU series are engineered to predict equipment faults and degradation by analyzing sensor data in real-time. These MCUs offer a critical advantage for industrial and EV applications, where uninterrupted operation and fast anomaly detection are essential. Thanks to ROHM’s proprietary “Solist-AI™” system, the devices run a simple three-layer neural network directly on the chip. This ensures rapid, localized learning and inference, significantly reducing maintenance costs, downtime, and the need for network infrastructure.
High-Speed AI Acceleration with Ultra-Low Power Consumption
Central to the MCU’s performance is the AxlCORE-ODL accelerator, which boosts AI processing speeds up to 1,000 times faster than conventional software-based systems. Operating on a 32-bit Arm® Cortex®-M0+ core, the MCU delivers impressive functionality including CAN FD communication, dual A/D converters, and three-phase motor control. It achieves this while maintaining a low power draw of just 40mW—making it an excellent fit for integration in electric vehicles, smart appliances, and industrial automation. These capabilities support retrofit installations in existing systems, offering a practical path to modernizing legacy equipment with cutting-edge AI intelligence.
Scalable Ecosystem and Seamless Integration Support
To ensure smooth deployment, ROHM offers a full ecosystem including its Solist-AI™ Sim tool for simulation and validation. This tool allows engineers to pre-train models and assess inference accuracy before real-world deployment. Additionally, ROHM collaborates with partner companies to provide development support, integration assistance, and optimized training workflows. This holistic approach encourages faster adoption of the AI MCU in industries that value flexibility, reliability, and real-time analytics—such as electric vehicles, industrial robotics, and smart infrastructure. As technology evolves, ROHM’s solution is a significant step toward sustainable, decentralized AI systems.
