Virtual Sensing, Vehicle Dynamics, and AI Chassis Systems Europe 2026

Virtual Sensing, Vehicle Dynamics, and AI

An Exclusive Interview with Ablamvi Ameyoe, Associate Expert Chassis Control, Renault Group

An Exclusive Interview with Ablamvi Ameyoe, Associate Expert Chassis Control, Renault GroupContents

Thank you for reading X2 Virtual Sensing

To get started, can you tell us about your background and role at Renault Group?

My background is rooted in engineering disciplines focused on understanding and controlling complex systems. I hold an engineering degree in Electronics and a second engineering degree in Electromechanical Engineering, before progressively expanding my expertise into control systems, signal processing, and vehicle dynamics. This academic path was further complemented by a Master of Science in Automatic Control, Mechatronics, Aeronautics, and Space Systems at the University of Bordeaux, and later a PhD at IMT Atlantique focused on driver modelling and state diagnosis.

Today, at Renault Group, I serve as Associate Expert in Chassis Control. My work focuses on transforming raw vehicle data into high-value signals and estimations that support chassis control and Vehicle Health Monitoring. In recent years, this has included a strong focus on virtual sensing, particularly for quantities that production vehicles cannot measure directly, such as tire forces or road interaction parameters. This field is especially relevant because it combines physics-based understanding, machine learning, and embedded deployment requirements within a clear industrial and cost-efficient framework. Innovation is also a core dimension of my role, reflected in a sustained level of patent activity, with on average one patent filing per month over the last four years.

Virtual Sensing, Vehicle Dynamics, and AI

A Conversation with Ablamvi Ameyoe, Associate Expert Chassis Control, Renault Group

As vehicles become increasingly software-defined, how is the role of the chassis system evolving beyond its traditional mechanical function?

Traditionally, the chassis was primarily considered as the vehicle’s mechanical backbone, with a clear role in stability, comfort, and safety. As vehicles become increasingly software-defined, that role is expanding. The chassis is now evolving from a set of hardware-driven subsystems toward a more strategic function within overall vehicle intelligence.

This evolution is driven by the convergence of electrified actuators, x-by-wire technologies, software-defined platforms, and high-performance computing. As a result, functions such as braking, steering, suspension, and propulsion must increasingly be coordinated through more centralized architectures. In this context, the chassis is becoming a key enabler of global vehicle motion control, with a growing role in estimation, coordination, and real-time system performance.

What sensing technologies are having the greatest impact on modern chassis systems, and where do you see the biggest opportunities for virtual sensing?

Modern chassis systems continue to rely on core physical signals such as IMU data, wheel speeds, and actuator information. However, the main shift is not only in measuring more, but in extracting more value from the data already available in the vehicle. As architectures become more centralized and software-defined, sensing is increasingly about improving observability and enabling more intelligent system coordination.

This is where virtual sensing offers significant opportunities. Many of the variables that matter most for chassis control and Vehicle Health Monitoring, such as tire forces, road interaction indicators, or internal dynamic states, cannot be measured directly in a cost-efficient way in production vehicles. By combining existing signals with physics-based models, system identification, and machine learning, it becomes possible to reconstruct these hidden quantities and transform raw data into actionable information. This creates value by supporting improved control, earlier detection of degradation, and enhanced system functionality without increasing hardware complexity.

How do you decide whether a problem is best solved through additional hardware sensors, physics-based models, or machine learning approaches?

When facing that kind of problem, I usually begin with the functional objective rather than the technology itself. The first question is always simple: can the variable of interest be estimated reliably from the signals that already exist in the vehicle? If the answer is yes, then physics-based modelling is often the natural starting point, especially when the system is well understood and interpretability matters, as it often does in vehicle dynamics or safety-related applications.

Machine learning becomes especially valuable when behavior is strongly nonlinear, highly variable, or simply too complex to describe explicitly, provided there is enough data and validation to support it. Additional hardware sensors, on the other hand, become necessary when indirect estimation cannot achieve the required level of accuracy, reliability, or response time, particularly in safety-critical contexts. In practice, the most effective solutions are often hybrid. They combine direct sensing where it is essential with indirect sensing where it can enrich observability without introducing unnecessary cost or complexity. The real challenge is not choosing one philosophy against another but finding the right balance for industrial reality.

Many organizations are combining physics-based engineering with AI. What advantages does this hybrid approach offer compared with purely data-driven solutions?

The main advantage of a hybrid approach is that it combines the strengths of physics-based engineering and data-driven adaptability. Physics-based models provide structure, interpretability, and consistency with known system behaviour. Machine learning complements this by capturing nonlinearities, variability, and effects that are more difficult to model explicitly.

When combined effectively, these approaches make it possible to generate signals and estimations that are both robust and deployable. This is particularly relevant for modern chassis systems, where performance must remain compatible with embedded implementation, cost constraints, and large-scale industrial deployment. In that sense, hybrid engineering is a key enabler for production-oriented intelligent vehicle functions.

Where is machine learning currently delivering the most value within chassis, vehicle dynamics, and vehicle health monitoring applications?

Machine learning currently delivers strong value in areas where the vehicle must estimate more than it can directly measure. In chassis applications, this is particularly relevant for state estimation, signal processing, and anomaly detection. These are areas where simplified models may not be sufficient on their own, while improved observability is essential for stronger control performance.

Machine learning is also increasingly relevant for Vehicle Health Monitoring. It can help detect early degradation patterns, identify weak signals before failures become visible, and support the transition from reactive diagnostics to predictive maintenance. As vehicles become more connected, these capabilities also create value for After-Sales services, fleet efficiency, and lifecycle optimization.

What are the biggest challenges in deploying AI-enabled functions into production vehicles, particularly within the constraints of automotive ECUs and safety requirements?

AI-enabled functions offer significant potential for production vehicles, but deployment remains challenging within automotive constraints. Models must operate within ECU resource limits, real-time execution requirements, and strict safety expectations. In this context, performance alone is not sufficient. Solutions must also be predictable, integrable, and robust in production conditions.

This becomes even more important as chassis control moves toward more centralized architectures, where estimated signals directly influence coordination and control functions. A key challenge is ensuring robust behaviour across the full range of operating conditions, including rare and extreme cases. This requires validation, traceability, explainability, and, in many cases, the support of physics-based knowledge to secure industrial robustness.

How is vehicle health monitoring evolving, and what role can intelligent chassis data play in predictive maintenance and after-sales services?

Vehicle Health Monitoring is evolving from a reactive diagnostic approach toward a more predictive and lifecycle-oriented model. The objective is no longer only to detect faults once they occur, but to identify early signs of degradation, understand component ageing in real operating conditions, and support earlier intervention.

Intelligent chassis data can play a central role in this evolution because it provides direct insight into load transfer, force transmission, vibration behaviour, and road interaction. By analyzing estimations linked to vehicle dynamics and structural loading, it becomes possible to identify wear earlier and with greater relevance. This supports predictive maintenance, condition-based maintenance strategies, fleet monitoring, and more data-driven After-Sales services.

How do connected vehicle data and over-the-air software capabilities change the way chassis systems can be developed, maintained, and improved throughout a vehicle's lifecycle?

Connected vehicle data changes chassis development by creating a stronger feedback loop between vehicles in operation and engineering teams. It extends the visibility provided by development tests and validation campaigns by making it possible to observe system behaviour across a wider range of real-world usage conditions. This can support more relevant model refinement and more robust estimator development.

Over-the-air capabilities extend these benefits across the full vehicle lifecycle by enabling continuous improvement of control strategies, estimation algorithms, and monitoring functions after vehicle launch. This opens new opportunities for more adaptive and service-oriented chassis functions, while also requiring strong discipline in cybersecurity, validation, data governance, and update qualification.

Looking ahead to 2030 and beyond, what capabilities do you think will define the next generation of intelligent chassis systems?

By 2030 and beyond, the next generation of intelligent chassis systems will increasingly be defined by their ability to integrate sensing, estimation, coordination, and control within more centralized and software-defined architectures.

Virtual sensing will remain a key enabler of this evolution by providing access to information that cannot be measured directly in a cost-efficient way in production vehicles. AI will also continue to expand its role, particularly when combined with physics-based models, system identification, and embedded control intelligence.

These systems will not only improve vehicle motion control, but also support predictive maintenance, After-Sales services, and broader connected vehicle value creation throughout the vehicle lifecycle.

What do you see as the biggest opportunity for intelligent chassis systems beyond 2030?

Enhancing safety and vehicle stability
Enabling autonomous driving functions
Reducing development and hardware costs
Predictive maintenance and diagnostics

You will be speaking at Automotive Chassis Systems Europe. What can attendees expect from your session?

Attendees will gain insight into a Renault Group OEM case study focused on smart wheel impact detection for series vehicles. The session will present a production-oriented approach designed to identify and qualify events such as curb strikes, potholes, and severe road irregularities using standard chassis CAN signals only, without additional hardware sensors.

The presentation will show how simplified physics-based indicators and lightweight machine learning can be combined to create an embedded, cost-efficient solution compatible with automotive ECUs. Based on proof-of-concept vehicle testing, it will also demonstrate the broader value of embedded chassis intelligence for after-sales diagnostics, predictive maintenance, and future vehicle health monitoring in software-defined vehicle architectures.

Automotive IQ invites you to join your colleagues in Munich for the 18th Annual Chassis Systems Europe 2026 conference. Whether you work in integrated chassis control, vehicle motion control, braking, steering, suspension, Automotive IQ’s Chassis Systems Europe 2026 conference gives you the clarity needed to understand where the market is heading, how competitive dynamics are shifting, and which technologies are advancing across the chassis domain.

Taking place in Munich, Germany, this year’s event will bring together leading OEMs, Tier-1s and industry experts to explore the technologies and strategies shaping the next generation of vehicle dynamics.