Stellar P3E

As vehicle mobility moves toward highly integrated powertrain architectures – consolidating engine control, inverters, on-board chargers (OBCs), and DC-DC converters into unified systems for hybrid and electric vehicles – microcontrollers must consolidate more functions. These systems also demand greater intelligence within the same physical footprint.

Stellar P3E shifts control from reactive to adaptive by integrating the ST Neural-ART™ Accelerator, a dedicated neural processing unit (NPU) embedded directly in silicon, enabling instantaneous AI inference that creates “smart” virtual sensors for more reliable powertrain control without additional hardware. By offloading AI workloads from the main processing cores, the NPU delivers >20× acceleration, executing sub-millisecond inference without compromising real-time behavior or functional safety. 
 

Stellar P3E combines a multi-core cluster of high-performance Arm^® Cortex^®-R52+ processors with hardware-enforced virtualization, enabling freedom from interference between software functions and more efficient consolidation of workloads. This architecture enhances responsiveness and supports safer, more deterministic behavior in modern automotive systems.

In the software-defined vehicle (SDV) era, memory is a key driver of innovation. Stellar P3E integrates extensible xMemory based on phase-change memory (PCM) technology, offering up to twice the density of traditional embedded flash. This provides robust, automotive-qualified performance and ample capacity for evolving software features throughout the vehicle’s operational lifespan.

Modern powertrain architectures require high-performance, real-time intelligence. Stellar P3E with its embedded NPU accelerator handles complex multi-sensor data directly at the actuation level. This enables functions such as predictive maintenance and anomaly detection for battery and combustion systems – identifying potential failures before they manifest. Furthermore, its virtual sensing can emulate physical sensors in inaccessible locations (e.g., inside an engine block) to provide unprecedented insights. This software-driven approach reduces hardware and complexity, boosting reliability and improving efficiency and range without compromising real-time behavior or safety.

High-fidelity AI models running on the NPU replace physical position sensors and emulate thermal sensors. 

This boosts system reliability, reduces bill of materials (BOM) cost, and lowers vehicle weight.

The system analyzes sensing and current signatures at the source to detect anomalies. 

This includes bearing wear, anode oxidation, and thermal irregularities months before they become critical failures.

The NPU offloads complex matrix computations, leading to higher performance, lower power consumption, and increased CPU headroom. 

This allows real-time optimization of switching frequencies, directly extending electric vehicle (EV) range through improved efficiency.

The Stellar P3E is engineered to master the complexities of high-integration and peak efficiency across the entire automotive spectrum. Whether optimizing internal combustion engines (ICE), managing hybrid transitions, or powering full electric vehicle (EV) platforms, the P3E consolidates multiple ECUs into a single, streamlined solution. By enabling smart sensing and high-performance processing, it delivers the intelligence required for the next generation of modern vehicle architectures: 

Stellar P3E is backed by the ST Edge AI suite - a comprehensive toolchain - connecting data scientists and embedded engineers in one flow:

Software & toolchain

• Fully supported in Stellar Studio, ST’s comprehensive automotive development environment.
• Automotive software solutions including AUTOSAR MCAL drivers

Engineering samples of Stellar P3E are available now. Contact your local ST sales office to request samples and begin evaluating your design.
Full automotive qualification and production readiness are planned for H2 2026.