IAAPS Conference 2026: Broadening the lens to pioneer net zero transport

Bio: Dan Parry-Williams is a leading automotive engineer with experience spanning Jaguar, Volvo, Aston Martin, Formula 1, and McLaren. He played a key role in iconic vehicles including the Jaguar XJ220, Aston Martin Vanquish, and McLaren P1, and later served as Director of Engineering Design at McLaren Automotive. He is also the founder of Iquad Technology, specialising in lightweight, sustainable vehicle structures.

This presentation discusses how innovation comes about in an industrial context (or doesn’t) and some experiential guidance on the process.

Small high-speed electric motors used in immersed cryogenic pumps could bring high power density and efficient pump operation to the cooling and fuel transport systems of electric aircraft; however they must be highly optimized for efficiency to mitigate heating of the cryogenic fluid, be power dense, and withstand high internal forces which change with temperature. In preparation for physical experiments, a high speed electric motor is studied using mechanical and electromagnetic simulations and analytical models to predict how it will perform whilst submerged in liquid nitrogen.

Transport is the largest emitting sector for greenhouse gases in the UK (Department for Transport, 2023). Drastic changes are needed within the transport sector in order to meet climate change mitigation targets (IPCC, 2022). Although the IPCC encourage the adoption of public transport over private vehicle use, it is not a feasible option for many due to a lack of availability (European Commission, 2023). In addition, active modes of transport are not practical for individuals with mobility issues (Cooper et al., 2019). Therefore, many people will continue to rely on the car as a main mode of transportation. Whilst car dependency still persists, then, cars need to limit their impact on the environment as much as possible.

Widespread adoption of electric vehicles powered by low-emissions electricity has the greatest potential to decarbonise land-based transport (IPCC, 2023). Electric vehicles (EVs) have successfully been adopted into fleets across the globe in recent years, yet despite having no tail-end emissions – and coinciding with the current trend towards large vehicles and SUVs – large quantities of resources are required to build them. The microcar may offer many solutions to these issues. They are compact in size, meaning they require fewer resources to build, and are relatively low in cost compared to other EVs (Grausam et al., 2022). Microcars have appeared throughout automotive history, often taking the shape of bubble cars and quadricycles. Some countries have been more successful in adopting microcars into their vehicle fleets, including Italy, India, and China. They offer functionality in their smaller size, offering a compact solution to personal mobility in densely packed urban areas, whilst also offering the luxury of an enclosed space which motorcycles and e-scooters cannot provide.

The benefit of the microcar’s size can also act as a barrier to their adoption. They often can only seat one or two people, they are limited in their top speeds due to EU regulations, and, due to their small batteries, have limited ranges. Previous research indicates that cost, perceived personal safety, seating occupancy and range are key attributes which impact consumers’ vehicle preferences and purchasing intentions (Venus et al., 2024; Axsen & Long, 2022). However, there is limited research indicating how these factors relate specifically to microcar adoption. In order to understand the barriers and enablers of microcar adoption, this study utilises a series of focus groups to identify specific factors that aid, or hinder, purchasing intention. Participants are current car owners and hold a UK driving licence, as we are interested in the views of individuals who have already made a vehicle purchase. Indeed, downsizing vehicle fleets can only be achieved by consumers replacing their current cars with smaller vehicles.

The results contribute to our understanding of the attitudes that consumers hold towards such vehicles and gain insight into what the barriers and enablers of microcar adoption are. Designers may also find the results of this study useful and develop new microcar designs based on consumers’ preferences. Furthermore, researchers and policy makers can use the results of this study to inform potential interventions to promote microcar uptake.

In-wheel electric machines are a desirable form of propulsion for lightweight electric vehicles, however the torque capability and speed range is often compromised due to the fixed gear ratio and volume/mass constraints from packaging the motor within the wheel of the vehicle. Actively changing the winding configuration of an electric machine during operation can be leveraged to enhance the torque capability, speed range, and efficiency of electric machines; the characteristic performance of which can be likened to that of a mechanical gearbox.

Reducing the flux linkage of a permanent magnet synchronous machine through winding reconfiguration from star to delta can significantly improve high speed efficiency, with a 32.78% reduction in total loss without detriment to low speed torque capability. This translates to a potential 5.43% reduction in total energy consumption over a real-world drive cycle of a solar powered electric vehicle. Improving drive cycle efficiencies for lightweight electric vehicles highlights a more sustainable transportation option for climate change mitigation strategy.

Sealing in gas turbines is vital for both efficiency and lifetime. Seals in gas turbines have historically required frequent maintenance to overcome performance degradation. Adaptive shaft seals attempt to mitigate degradation by finely-actuated movement of translatable components to track high-speed rotors during transients.

This presentation will overview the working principle of one such compliant non-contacting seal, and present a simplistic modelling approach for an otherwise complex system.

Electrification of lightweight aviation demands mass-optimised, structurally integrated battery system that meet strict requirements. These systems must be multifunctional, combining propulsion power delivery with secondary thermal functions such as airframe de-icing.

This study validates a Thermal Management System (TMS) for a wing-integrated battery submodule. An experimental campaign characterises thermal behaviour of a Lithium-ion battery pack under mission profiles for an eVTOL application. The tests quantify the heat extraction capacity of the cell-base cooling system.

These results establish the operational thermal limits of the 12-pouch cell battery submodule and the available waste heat that can be redistributed to support auxiliary aircraft functions. Thermal simulations are developed to model heat generation, across the battery submodule and scale this to a full battery module. Experimental data validate the numerical model, enabling prediction of thermal performance across wider designs. The model assesses the feasibility of using the available waste heat to support wing de-icing. The findings demonstrate the benefit of the system which can support heat demands.

Inside industrial gas turbines, there are enclosed air-filled gaps between rotating discs called compressor cavities. The temperature behaviour of these cavities directly controls how much the engine components expand and contract during operation, which in turn determines the tip clearance — the gap between the spinning compressor blades and the outer casing. If that gap is too large, efficiency drops and the engine becomes less stable; if it's too small, blades can grind against the casing. Getting this balance right is critical for efficient, reliable operation.

This research investigates a previously under-studied phenomenon: radial inflow — a leakage of air entering the cavity radially inward through mechanical joints called Hirth couplings (the serrated connections used to assemble the compressor drum). The central question is: how does this leakage change the way flow structure in the cavity, and what does that mean for the thermal behaviour of the spinning discs? Understanding this is key to improving the accuracy of thermal models used in engine design — with direct implications for efficiency and emissions in both current and future gas turbines.

This study investigated how fence height affects the performance of brush seals used in gas turbines. Experiments using a large-scale test rig show that fence height strongly influences leakage, bristle flexibility and stiffness. Larger fence heights reduce leakage at low pressures but increase bristle stress, while smaller fence heights create a stiffer, more compact seal that can increase leakage under certain conditions. The results provide insight into optimising brush seal design for improved efficiency.

Platinum is widely used in fuel cell technologies to enhance reaction rates. However, conventional platinum catalysts often suffer from poor stability and loss of active surface area during prolonged operation, leading to decreased performance. To combat this 3D platinum structures have been created using electrochemical methods to deposit the platinum into intricate lipid templates.

The result is an interconnected network of platinum nanowires which exhibit a very high surface area. The use of the lipid template allows control over the size of the platinum structure and with it tuning of its catalytic properties. The catalytic performance has been investigated using further electrochemical methods to measure the oxygen reduction reaction. The results demonstrate enhanced catalytic activity compared to conventional platinum surfaces, highlighting the potential of template-directed 3D design for improved catalyst efficiency.

Pouch cells are unique amongst Li-ion cell technologies in that they have a soft casing. This causes them to swell and contract under different state-of-charge, temperature, and aging conditions, and introduces factors such as external pressure that affect cell performance.

In this work, several sensor technologies are deployed to experimentally measure a wide range of Li-ion pouch cell properties during operation, including sensors for measuring cell displacement, ultrasound transducers, spring-loaded thermocouples, and load cells. From these experiments, the Li-ion pouch cell dataset that is obtained is one of the richest available, and it is demonstrated how the data from different sensors can be combined to measure unique cell properties, such as through-plane ultrasonic velocity.

Field experiments offer ecological validity that laboratory studies cannot match, but embedding experimental designs in live institutional processes comes with costs that are rarely discussed openly.

This presentation shares the methodological journey behind a quasi-experimental study examining whether commuter involvement in organisational travel plan development increases policy acceptance and compliance. Yet, when a carefully crafted design met 'the real world', the practice of conducting a field experiment acquired new dimensions. Recruitment and attrition, inconsistency across workshop sessions, participants not engaging with preparatory materials, and timing constraints that systematically excluded certain groups all shaped what the data could ultimately say.

This presentation reflects honestly on the gap between design and execution, and what it means for the validity of findings, and for anyone considering similar approaches. It is aimed at researchers and practitioners who work at the interface of institutional decision-making and empirical research, and who want to think seriously about what "rigour" looks like when the field plays by its own rules.

Bio: Matt is one of the UK’s leading creative thinkers on cities, inclusion, invention and technology, and their impact on the environment, the economy, and people. He is chairman of the Laboratory of Thought in Amsterdam, focusing on the issue of mobility, partner of smart cities advisory firm FieldLab, and start-up mentor at Georgetown University, Washington. Matt started his career in commercial and marketing at Diageo, then founded the creative agency Brave. He combines private and public sector advisory roles and has recently led and advised several high-growth companies in smart cities, design, and mobility – including Bigbelly, Edenspiekermann, TransitScreen and Applied Wayfinding.

Abstract: Why are our cities killing us? By Design.

“WHO predicts 70% of global deaths and 56% of the global disease burden will be attributed to lifestyle behaviours by 2030”

We have a problem

Modern mobility is great - in the short term.

Longer term, it's killing us and our planet. By design.

With more roads, more parking, parcels, and congestion, from second cars to electric cars, our daily lives are designed for convenience, passivity, and gratification.

Why does it matter? An epidemic of pollution, noise, congestion, exclusion, ill health, obesity, and depression?

Is technology the answer? Or should we focus on future generations, not future technologies?

I’ll explore how cities and mobility can improve everyday public health, reduce bias, be accessible and safe for everybody, built around the needs and rhythms of humans and the planet.

Transport systems are designed to move people from A to B, but they also represent an intervention point for two of today’s most pressing challenges: the climate crisis and the public health crisis. This presentation uses Bristol as a case study to examine how transport systems interact with the wider urban environment to shape health. Despite a population of around half a million, the city is characterised by high car dependency, severe congestion, and substantial life expectancy inequalities.

We adopt a systems-based perspective, considering how transport, built environment features, and health are interconnected. Drawing on my PhD research, I show how the urban system as a whole predicts health outcomes more effectively than individual components. This highlights the limitations of isolated interventions and the importance of coordinated, system-level change. I argue that creating walkable, integrated environments can reduce car dependency, increase physical activity, and support healthier, more sustainable cities. There is a clear opportunity to “connect the dots” and reframe urban systems as a central tool for population health improvement.

The success of decarbonisation strategies depends on the willingness of choice architects to implement them. Here we summarise insights from three studies exploring the role of SMEs in mediating modal choice. In the UK, SMEs employ 61% of the private sector but are not included in the planning permission laws that are used to secure Workplace Travel Plans (WTPs). SMEs can play a pivotal role in shaping the commuting practices of those they employ, but their adoption of WTPs is entirely voluntary. Researchers and policymakers need a greater understanding of the determinants of implementation in order to encourage the implementation of these initiatives at the scale and speed required to reach climate targets.

Our first study (n=20,017) quantifies the role of SMEs in enabling sustainable commutes in the West of England. Our second study tests the first explanatory model of adoption, including, for the first time, the effects of peer adoption (n=283). Results suggest higher capability and motivation is associated with a higher likelihood of WTP adoption. The effect of opportunity varied depending on the level of adoption, enabling non-adopters but inhibiting firms from moving from an ineffective to an effective WTP. Results of a vignette RCT suggests that increasing perceptions of peer adoption can increase reported intention to adopt or expand sustainable travel measures. Based on the results of our third study, an agent-based model, we explore a roadmap of how policymakers could trigger a positive feedback loop of WTP adoption among SMEs.

This talk presents a game-theoretic model of congestion in transport networks over time. It studies how route choices in one period affect queues and travel costs in later periods, and how these dynamics can still admit useful mathematical structure. A central motivation is to understand how individual route choices can differ from what is best for the transport system as a whole, and to explore how ideas from static congestion models can be extended to time-dependent settings. By looking at how congestion builds up over time, the work relates to the wider challenge of using limited transport capacity more effectively.

This talk will discuss traffic data collection using multi object tracking(MOT) on traffic cameras, focusing on real time single camera operation for vehicle counting, path analysis, and basic speed estimation without additional sensors. It also explores flexible short term deployments using low power cameras operating at very low frame rates to support studies with Bath Council. This will support targeted analysis of road usage for planning and development in otherwise underserviced areas. This work has applications in smart city monitoring and road and development planning, emphasising work that is directly deployable for real world insight.

What is the potential for e-bikes to reduce car use, carbon emissions and car dependency in Bath and North East Somerset?

Past research at a national level estimates electrically assisted pedal bikes (e-bikes) could significantly reduce car usage and address vulnerability to high transport costs (Philips et al., 2022).

This mini-project has re-run this national model specifically for the BathNES region, which covers 115 LSOA areas (roughly 400-1000 households in each) and compared this with the West of England Combined Authority region (WECA). This high-resolution data takes into account people’s current travel behaviour, car ownership, health/fitness and the hilliness of roads in the neighbourhood.

The study finds e-bikes could potentially replace a large amount of car travel (roughly 2,500km per person per year). The contribution is a much larger in countryside and rural areas because (a) car ownership and usage is much higher in these places, so there is more to replace (b) e-bikes have the biggest relative advantage at covering longer distances than people need to cycle in cities (c) e-bikes create cycling capability in hilly areas.

This study was undertaken proactively, presented to BathNES Council and included in the recent Movement Strategy report. Further work could make this regional analysis easily available to other authorities, and brought to life for a general audience to promote e-bike adoption.

Bio: Brian is the Global Head of Sustainable Transport & Fuels at HSBC with over 20 years of experience in the energy sector. He spent the first half of his career in utilities, working in the energy efficiency space for E.ON and LNG trading for Centrica. During this time, he went on secondment to Louisville, Kentucky to lead a smart meter roll out. The second half of his career he spent with bp, modelling transport and refined product demand & supply, leading the market intelligence function within Castrol and publishing the bp Energy Outlook.

A superconductor in extreme cold has practically zero electrical resistance. Above a certain temperature (around 90K or -180degC), it will lose its superconductivity. This is quite catastrophic, because it holds massive current densities - suddenly gaining resistance causes dramatic heating and permanent damage. This is a serious safety concern when the application is aviation or marine propulsion that uses superconducting coils inside motors. It is critical that the superconducting coils are always maintained below this temperature threshold at every point in the machine to avoid thermal runaway. Firstly, a basic setup of a superconducting motor and its benefits will be outlined for context, followed by a novel cooling system design for keeping it at cryogenic temperatures and the practical challenges involved.

Hydrogen fuel cells were first invented in the 19th century, but their widespread adoption has accelerated significantly over the past two decades. Among the various types, proton exchange membrane fuel cells (PEMFCs) have emerged as a leading option for diverse applications ranging from aerospace systems to automotive propulsion. A major challenge for PEMFC technology, however, lies in the limited lifespan of the cells and the gradual performance degradation over time. To mitigate this issue, current commercial systems are often oversized and underutilized to meet minimum durability guarantees between suppliers and end users.

This conservative approach stems largely from an incomplete understanding of degradation mechanisms across the fuel cell lifetime. This study investigates the performance decline of single PEM cells operated for 400, 600, and 1000 hours under dynamic load conditions. The cells were subjected to load profiles derived from the Worldwide Harmonized Light Vehicles Test Procedure (WLTP), which provides a more realistic and up-to-date demand model compared to the New European Driving Cycle (NEDC). In addition, voltage step protocols were applied to further characterize performance loss. The influence of external operating conditions, specifically cell temperature (55–85 °C) and relative humidity (30–100%), was systematically analyzed.

The results provide deeper insight into how operating conditions accelerate or mitigate degradation in automotive applications. Based on these findings, a hybrid degradation model is proposed to assess fuel cell health status, enabling more accurate predictions of end-of-life and improving understanding of the underlying degradation mechanisms.

We conducted a prospective LCA on carbon fibre reinforced polymer (CFRP) manufactured in the US. Our study looks at the impact of replacing fossil fuelled process heat with electrification as well as the impact of increased renewables in future grid systems.

Composite materials are becoming increasingly commonplace in modern manufacturing due to their favourable characteristics. Of these, high performance CFRP is especially popular in the transport sector. Its superb strength-to-weight ratio is favourable for aerospace and automotive applications as well as for pressure vessels, especially pressurised hydrogen fuel tanks. While it brings many beneficial properties to the table in its respective applications, CFRP is currently a major source of both cost and embodied carbon for these products which is why we conducted this study.

Key takeaways are that benefits in the short-term are limited due to a high portion of coal and gas in the US energy grid however, by 2035 it will be favourable to use grid electricity. A 100% renewable energy system using green ammonia feedstock has the potential to reduce carbon emissions by up to 76 %. Increased use of potentially lower cost electricity and the prospect of dual-fuelling provide avenues for cost saving in an energy dominated value-chain.

The net-zero targets of the United Kingdom are linked to major reductions in the transport sector, which contributed 28% of UK’s total domestic emissions in 2022. Central and local governments are prioritising electrification as a key method to decrease emissions; however, little has been done regarding indirect impacts of implementing wide-scale electric infrastructure and the application of future scenarios for national climate policies. Therefore, prospective life cycle assessment is performed on a case-study route of an electrified fleet of city buses operating on a singular line in SW England and compared to a similar analysis of a diesel-powered alternative.

Battery-electric and diesel buses were modelled, with Ecoinvent and carculator-bus being used as inventory databases. Processes for road use, wear, and infrastructure were also modelled. Per-passenger impacts, along with upfront construction and period maintenance, were considered year-by-year. For prospective assessments, premise will be employed to compare scenarios, selected based on ambitiousness of aligned policies, and derived from a REMIND-based integrated assessment model to examine scenarios that align with local government targets.

Results show that electrification of a local bus route within the UK can significantly reduce CO2-eq. emissions, with a pay-off time between 3-6 years of operation. Uncertainty in results is higher when examining the route on a per passenger-kilometre basis, with a large source originating from additional infrastructural wear, although the range of uncertainty is reduced with uptake. It is hypothesised that the application of scenarios aligned with environmental targets set by the UK government will further demonstrate the importance of electrification in reducing transport emissions, yet it is unclear how much of an impact this may make when expanded to a broader context.

The work undertaken supports existing evidence in literature that transitioning towards an electrically driven public transport network is a key step within the decarbonisation of the UK. Additionally, it further emphasises the need for electrification to be directly linked with local governmental efforts towards modal shift away from private transportation. Finally, the use of a time-explicit assessment within this context and the ease and simplicity of producing results can provide an optimal method for more granular control of activities over time-explicit stages.

Transport projects often face opposition or backlash, which can lead to project delays or cancellations and increased costs. Understanding what leads to these responses is essential to successfully delivering these projects. Emotional responses to these projects may explain these responses.

In this presentation, I will outline why we think emotions may be linked to acceptability, the mechanisms through which these effects may occur, and how perceived fairness may be a key determinant of emotional responses.

Bio: David Howey is Professor of Engineering Science at the University of Oxford and an IEEE Fellow. He leads a research group focused on the modelling, control and diagnostics of battery energy storage systems; their work combines electrochemical modelling, advanced instrumentation and data-driven methods, including probabilistic machine learning, to better understand and improve battery performance, lifetime and safety. This research has applications across electric vehicles, grid-scale storage and off-grid energy systems, contributing to the transition to low-carbon energy. He has co-authored 150+ articles, holds several patents, and works closely with academic and industrial partners worldwide. David is co-founder of the Oxford spin-out company Brill Power and the award-winning Oxford Battery Modelling Symposium, and is Research Director of a UKRI Prosperity Partnership with Fortescue Zero on Energy Storage for Decarbonisation.

Abstract: Lithium-ion batteries have enabled a revolution in communications and computing and are now underpinning a move to electric vehicles, as well as allowing us to absorb renewable energy on the power grid. Batteries, however, are a bit like people - thermally and mechanically fragile and they degrade over time! Innovation in batteries requires more than just materials and chemistry breakthroughs; it also needs systems engineering to help us design thermal management systems, diagnose internal states, and improve performance and lifetime. In this talk I will discuss several aspects of battery systems engineering, including modelling, parameterisation from lab and field data, novel diagnostics techniques, and integration within wider energy management control systems.

Anode-free lithium metal batteries offer a way to high energy density systems but remain limited by uncontrolled lithium plating and interfacial instability. I will present the first electrochemical–mechanistic model of anode-free cells incorporating a carbon black buffer layer, developed in collaboration with Lawrence Berkeley National Laboratory. The model resolves lithium transport, adsorption, and deposition processes at the carbon black interlayer, enabling better insight into the plating behavior.

A key novelty is the explicit treatment of three independent overpotentials associated with lithium adsorption onto carbon surfaces, initial lithium plating and plating on pre-existing lithium deposits. Coupled electrochemical (thermodynamic) and pressure-driven (mechanical) phenomena are used to capture the interplay between interfacial energetics and stress evolution during cycling. The model further accounts for lithium diffusion within the carbon matrix and surface adsorption kinetics, which are often neglected in conventional frameworks.

Simulation results show the conditions under which lithium preferentially plates at the current collector versus within the buffer layer, highlighting the roles of pressure, interfacial energetics, and transport limitations. These insights provide design guidelines for engineering buffer layers that mitigate dendrite formation and improve reversibility in anode-free configurations.

Hydrogen fuel cell electric vehicles (FCEVs) are a promising technology for heavy-duty applications due to hydrogen’s high gravimetric energy density and rapid refuelling capability. However, their widespread adoption depends on improving system-level efficiency. This study presents a 1D optimisation of the fuel cell cathode system, where air supply and water management are strongly coupled.

A fuel cell system with electrically assisted turbocharger and passive humidifier is optimised using parametric and genetic algorithm-based methods. Key parameters (including operating temperature, stoichiometry, back-pressure valve position, and humidifier bypass) are varied to maximise hydrogen-to-electricity efficiency while maintaining membrane hydration. Under steady state conditions, an open back pressure valve, high humidifier bypass, and load-dependent stoichiometry control improve system efficiency by up to 3%. Meanwhile, adequate hydration is maintained by reducing stack temperature by less than 15% compared to the manufacturer's conditions.

The study demonstrates that integrated cathode system optimisation, rather than isolated component improvements, is necessary to realise meaningful efficiency gains in FCEVs for heavy-duty transport.

Accurate degradation diagnostics and prognostics of lithium-ion batteries are essential for the safe and reliable operation of electric vehicles. However, existing approaches often suffer from limited generalisation and insufficient integration between physical insight and data-driven learning.

In this work, a hybrid framework combining physics-based electrochemical modelling with physics-informed neural networks (PINNs) is proposed for mechanism-aware degradation diagnostics and prognostics. The proposed approach embeds physically meaningful states and degradation dynamics into the learning process, enabling a deep integration of model-based and data-driven representations. In addition, informative features extracted from partial operating data are incorporated to enhance robustness under practical operating conditions.

The resulting model achieves stable and accurate estimation of battery health states and demonstrates strong predictive capability across varying operating regimes and degradation stages. The framework provides a scalable and interpretable solution for battery ageing analysis, with potential applications in advanced battery management systems for electric vehicles.

Hydrogen is an alternative fuel which could play an important role in decarbonising transport. To replace petrol, hydrogen production must become cheaper, and importantly, hydrogen fuel must be as pure as possible. A reliable way to produce hydrogen is through steam reforming of natural gas where the product is purified using pressure swing adsorption (PSA). However, optimising the PSA process is a wicked problem involving many inter-connected factors. One of these factors is the adsorbent used. For a long time, zeolite and activated carbon have set the standard for hydrogen purification, yet emerging metal-organic frameworks (MOFs) are showing promising performance. But given that MOFs are modular and 10,000s of diverse structures are accessible, it is a challenge to optimise the operating conditions for each MOF to yield as much pure hydrogen as possible.

Computational MOF screening is currently underpinned by molecular simulations. This involves measuring the adsorption of the key components under equilibrium conditions. However, an important limitation of this approach is that it neglects kinetics. To optimise operating conditions for every screened MOF, a dynamic model is therefore required. A physics-based breakthrough model can capture the transport of components over time, enabling the optimisation of operating factors such as adsorption and desorption pressures, temperature, and adsorbent dimensions.

In this work, molecular simulations are coupled with breakthrough simulations into a machine learning screening framework. Through this novel workflow, we demonstrate the benefits of coupling multiscale simulations in MOF screening. We also compare the outcomes of the new workflow to the results of a traditional, equilibrium-based screening.

This project focuses on helping AI understand unusual traffic events in real-world road scenes. Rather than only detecting that an incident has happened, our goal is to explain what happened, which vehicles were involved, and why the situation may be unsafe by harnessing the power of Visual Language Model. To achieve this, we created Roundabout-TAU, a new dataset built from real roundabout surveillance videos, and developed TAU-R1, a two-stage AI framework for traffic anomaly understanding. In the first stage, a lightweight model quickly identifies whether a scene is normal or abnormal. In the second stage, a larger model provides a clearer natural-language explanation of the event. This design makes the system both efficient and practical for real traffic monitoring. Overall, the project shows how modern AI can move beyond simple detection toward more meaningful scene interpretation, with potential applications in traffic safety, incident monitoring, and future intelligent transportation systems.

As automotive systems grow increasingly complex—driven by the transition to autonomous driving and sophisticated ADAS—the need for human-centric validation has never been more critical. This presentation explores the pivotal role of Driver-in-the-Loop (DiL) technology as a bridge between pure virtual simulation and physical track testing.

We begin by defining the modern Driver Simulator, moving beyond simple gaming setups to high-fidelity motion platforms that replicate real-world vehicle dynamics. The session will delve into the inherent complexities of DiL, including the synchronization of low-latency visuals, haptic feedback, and the vestibular cues required to minimize simulator sickness and ensure data validity.

Aim to provide insights into diverse Applications and possible Use Cases, ranging from subjective ride-and-handling tuning to the safety-critical testing of Human-Machine Interfaces (HMI). By integrating a real human driver into the digital thread early in the development cycle, OEMs can significantly reduce prototype costs and accelerate time-to-market. Finally, we will discuss possible Future Developments in the field, such as the integration of AI-driven traffic agents and the evolution of "cloud-connected" simulators for collaborative global engineering.