Embedded NPU in automotive MCU enables real-time AI, virtual sensors, predictive maintenance, and precise motor control while reducing system complexity and BOM.
As electric vehicles evolve toward integrated x-in-1 architectures, combining inverter, OBC, DC-DC, and more, microcontrollers must handle higher intelligence and consolidate multiple functions while meeting strict safety standards. Traditional ECUs struggle with real-time AI inference, predictive maintenance, and virtual sensing at the actuation level.
STMicroelectronics addresses this with the Stellar P3E, the first automotive MCU with an embedded Neural-ART neural processing unit (NPU). By integrating AI acceleration directly on-chip, the MCU enables instantaneous edge AI, powering virtual sensors, predictive maintenance, and soft-switch applications without adding hardware. This reduces system complexity, BOM, and weight while improving reliability, energy efficiency, and vehicle range.
Stellar P3E combines high-performance Arm® Cortex®-R52+ cores with hardware isolation for safe compute, extensible xMemory based on phase-change memory, rich I/Os, and high-speed deterministic communication. It can synchronize over 90 ADC and PWM signals for dual-motor drivetrains, perform ultra-fine motor control with 102 ps PWM resolution, and optimize real-time AI workloads for EV powertrains.
Key features of the product include:
- Embedded NPU for real-time AI and virtual sensing
- Predictive maintenance for battery and motor systems
- High-performance Arm® Cortex®-R52+ cores with hardware isolation
- Extensible xMemory with PCM technology
- Supports X-in-1 powertrain ECU integration
- Ultra-fine PWM resolution for high-resolution motor control
- Reduced BOM, cable weight, and system complexity
Luca R, GP and Auto MCU GM at STMicroelectronics, said, “Stellar P3E sets a new benchmark for automotive electrification by combining high-performance real-time control and edge AI in a single, ASIL-D compliant system-on-chip. This empowers our customers to deliver safer, smarter, and more efficient X-in-1 solutions for the next generation of electric vehicles.”
