Revolutionizing Vehicle Technology Through Intelligent Actuators

A look at E/E complexity at the so-called endpoints, where microcontrollers play a crucial role.

December 1, 2025

Florian Baumann and Yoann Foucher

Schematic representation of a zonal E/E architecture, consisting of four zonal control units and a central computer used for ADAS and infotainment. *(STMicroelectronics)*

Autonomous driving is often seen as just the tip of the iceberg when it comes to public perception of how artificial intelligence has been implemented in vehicles. However, this perspective often overshadows a much deeper, less visible transformation: AI's integration into the vehicle’s E/E (Electrical/Electronic) architecture.

This complex and intricate system encompasses far more than the capabilities for autonomous driving. It includes sensors that collect data, microcontrollers that process data in real-time and actuators that control safety-critical automotive functions.

The profound shift is not just about what the vehicle sees, but how AI is fundamentally redefining the experience from opening a door to driving autonomously.

The foundation

The E/E architecture (Figure 1) forms the foundation of modern vehicles, enabling advanced functions that are increasingly powered by AI. Within this system, microcontrollers and microprocessors play a crucial role, as they must process information in real-time and translate commands into physical actions. Precision and responsiveness – whether for steering movements in autonomous driving systems, electronic seat adjustments, or climate control – depend on a sophisticated interplay of data processing elements.

There are only a handful of places where automotive data is processed, particularly for the AI computation and system control aspects:

Microcontrollers (MCUs) process data locally within the vehicle, enabling fast real-time responses. For example, MCUs control various actuators, monitor the state of critical components, and can take over vehicle control in fail-safe scenarios.
Microprocessors (MPUs) and Graphics Processing Units (GPUs) handle more complex tasks like image recognition or advanced visualizations, enabling autonomous driving and enhancing the user experience (UX).
Data centers, where cloud technologies complement the extensive data analysis and optimization, which is critical for tasks like remote diagnostics and fleet management.

This three-tier architecture provides a comprehensive AI framework, revolutionizing the overall vehicle experience. It highlights the AI’s growing role as a driving force in next-generation vehicles.

Complexity at endpoints

Florian Baumann is technology advisor at STMicroelectronics. *(STMicroelectronics)*

Within vehicle architectures, an endpoint refers to a physical system typically consisting of small, resource-constrained devices such as microcontrollers, sensors, and actuators. Sensors measure physical properties such as temperature, pressure, or speed; microcontrollers process this data in real-time with minimal latency and derive commands to actuators, which convert electronic signals into physical actions like steering, braking, or accelerating.

Each vehicle contains numerous endpoints that must interact and communicate in real-time and synchronously. Figure 2 illustrates the complexity of a single endpoint using the example of a vehicle door control system. Typically, each door in a vehicle contains such an endpoint and there are numerous such endpoints in the complete vehicle (e.g., roof, trunk, lighting, etc.).

A simple action like opening a vehicle door involves a complex sequence of events that demonstrates the deep interaction of microcontrollers, sensors, and actuators:

An ultra-wideband (UWB) sensor within the door detects the proximity of the vehicle key while the microcontroller processes the data.
An actuator unlocks the door, enabling keyless entry.
An actuator unfolds the mirrors.
Blinking Indicators are activated to provide a visual signal to the driver.
A welcome sequence is initiated e.g., ambient lighting and sound greeting the driver. Exterior lighting is activated.
Digital Cockpit is activated to display useful information.
Infotainment is initialized and connects to the driver’s smartphone via Bluetooth/Wi-Fi to prepare personalized content.
Climate Control is activated to pre-cool or pre-heat the interior.
Memory seats are adjusted to pre-set positions through multiple actuators.
Steering wheel is adjusted for easier entry and pre-set to a stored position.
Mirrors are adjusted and aligned to stored preferences.

This process chain is by no means exhaustive and varies significantly across different vehicle classes. It serves to highlight the complex coordination between numerous decentralized endpoints and reflects the broader challenges faced by vehicle manufacturers.

Challenges for vehicle manufacturers

Behind a simple task such as opening the vehicle door, there are numerous complex functions running in the background not visible to the passengers. Vehicle manufacturers are starting to centralize and consolidate those functions and further, they are becoming software-defined, connected and enriched to enhance the user experience. This upcoming complexity presents several challenges for vehicle manufacturers:

The rising number of features requires sophisticated integration to ensure seamless operation.
Ensuring compatibility between different systems and suppliers is critical and challenging.
New architectures are needed to efficiently manage energy and charging systems.
Managing updates from multiple suppliers while ensuring reliability is a significant hurdle.
Balancing innovation, compliance and cost efficiency is a constant challenge.
The market demands faster development cycles, putting pressure on manufacturers to deliver without compromising quality.

These challenges are driving vehicle manufacturers to rethink and reinvent their existing E/E architectures. A global trend across all manufacturers is the adoption of so-called zonal architectures.

Leveraging AI and zonal architectures

Yoann Foucher is strategy product director for the Automotive MCU at STMicroelectronics. *(STMicroelectronics)*

The automotive industry is moving away from traditional domain-based architectures towards zonal architectures. In domain-based systems, each domain (e.g., powertrain, driver assistance, infotainment) including its endpoints operate independently, leading to complex wiring and redundant hardware. Zonal architectures, on the other hand, divide the vehicle into physical zones and consolidate multiple individual domains and endpoints into larger microcontrollers that are now integrating AI capabilities.

This approach reduces wiring complexity and weight,Decreases the number of smaller microcontrollers, consolidating them into larger multi-core microcontrollers that include AI accelerators, simplifies software management, reducing lines of code (LoC), enables centralized data processing with faster decision-making and facilitates efficient data exchange and coordination between systems.

The transition from domain-based to zonal architecture marks a significant shift in vehicle technology, with the Ethernet standard playing a key role. While previous standards are being gradually phased out, Ethernet is becoming the central communication backbone of modern vehicles. The successful implementation of zonal architecture depends heavily on the integration of Ethernet technology.

To support this transition, semiconductor manufacturers like STMicroelectronics are integrating multiple Ethernet ports and embedded Ethernet switches directly into their microcontrollers. This allows vehicle manufacturers to modernize their architectures, optimize data communication and lay the foundation for forward-looking applications like AI. Ethernet-based architecture not only offers higher bandwidths but also provides the flexibility needed for the demands of modern and future in-vehicle applications.

Application examples

Illustration of an endpoint using the example of the vehicle door control system. It demonstrates the integration of a variety of sensors, a microcontroller, 13 actuators, and other electronic components. *(STMicroelectronics)*

New AI workloads are increasingly running in zonal control units (ZCUs), which leverage the proximity of sensors and actuators for low-latency processing. The amount of data that are now collected shall even accelerate AI adoption as AI models improve with larger amount of data. Examples include:

Anomaly detection increases safety in vehicles: AI models detect patterns and anomalies, enabling real-time monitoring and proactive maintenance.
Optimization of powertrain and energy management: AI-driven systems optimize battery performance and fuel efficiency, adapting to driving conditions.
Virtual sensor and sensor fusion: AI integrates data from multiple sensors or emulates sensors enabling smarter decisions in body, chassis, and thermal management systems.
Adaptive processing: AI optimizes input and output data for faster battery charging, improved engine efficiency, and higher-performing motors.

Future challenges

Artificial intelligence is revolutionizing vehicle technology by redefining the electrical/electronic architecture and placing intelligent actuators at the forefront. From door opening and intelligent climate control to autonomous driving functions, AI enables numerous new applications for microcontrollers, making driving safer, more efficient and enhancing the driving experience. The transition from domain-based to zonal architecture reduces complexity, improves data processing, and lays the foundation for innovative applications such as predictive maintenance, optimized energy management, and sensor fusion.

Despite challenges - such as increasing system complexity, interoperability, and time-to-market pressure - AI offers enormous opportunities to make vehicles safer, more efficient, and user-friendly. The automotive industry stands on the brink of a new era, where AI not only transforms the driving experience but also sustainably shapes the future of mobility.

Yoann Foucher is strategy product director for the Automotive MCU at STMicroelectronics. Florian Baumann is technology advisor at STMicroelectronics. They wrote this article for SAE Media.