Periodic Reporting for period 1 - EmotAI (EmotAI: Emotion Driver State Monitoring System based on AI)
Période du rapport: 2024-02-01 au 2026-01-31
Road transportation plays a central role in modern societies, enabling economic activity, social inclusion, and mobility. Despite major technological advances, road accidents remain a critical societal challenge, with driver inattention, fatigue, and emotional states among the leading contributing factors. In Europe, road safety is a strategic priority, reflected in long-term policy ambitions such as reducing fatalities and serious injuries towards zero and strengthening vehicle safety through intelligent, human-centred technologies.
In this context, advanced driver monitoring systems have emerged as a key technological response. However, existing solutions largely focus on isolated indicators—such as eye closure, phone use, or head pose, often leading to fragmented interpretations of driver state and a high rate of false or poorly timed alerts. These limitations reduce system effectiveness and can undermine driver trust. There is therefore a clear need for more holistic, cognitively informed approaches that go beyond technical detection and address how drivers perceive, process, and respond to safety interventions.
The EmotAI project was conceived to address this gap by developing a new generation of driver state monitoring grounded in artificial intelligence, affective computing, and cognitive psychology. The overall objective of the project was to achieve a holistic understanding of the driver’s emotional and attentional state by integrating multiple complementary pillars: attention, activation (emotional and physiological state), driving performance, and external contextual factors. By combining these dimensions, EmotAI aimed to move from fragmented detection towards coherent interpretation of driver state over time.
A central ambition of the project was not only to improve detection accuracy but also to enhance the effectiveness of safety interventions. To this end, EmotAI introduced a cognitive control perspective, focusing on how alerts can be designed to meaningfully influence driver attention and emotional state rather than merely signalling risk. This approach reflects a clear pathway to impact: improved understanding of driver cognition enables better-designed alerts, which in turn can support safer driving behaviour, reduce accident risk, and increase acceptance of intelligent safety systems.
The project is embedded in a broader political and strategic context that promotes trustworthy, human-centric artificial intelligence and safer, more sustainable mobility. By aligning technological innovation with societal needs and regulatory trends, EmotAI contributes to ongoing efforts to strengthen road safety, support the deployment of intelligent transport systems, and enhance Europe’s leadership in responsible AI.
Social sciences and humanities played a key role in the project, particularly through the integration of cognitive psychology and human factors research. Psychological theories of attention, emotion, and cognitive load informed both the interpretation of driver state indicators and the design of cognitively appropriate alerts. This interdisciplinary integration ensured that the project addressed not only what can be detected by AI systems, but also how drivers cognitively and emotionally respond in real driving contexts. As a result, EmotAI positions itself as a human-centred innovation with the potential for significant scientific, societal, and industrial impact at scale.
In this context, advanced driver monitoring systems have emerged as a key technological response. However, existing solutions largely focus on isolated indicators—such as eye closure, phone use, or head pose, often leading to fragmented interpretations of driver state and a high rate of false or poorly timed alerts. These limitations reduce system effectiveness and can undermine driver trust. There is therefore a clear need for more holistic, cognitively informed approaches that go beyond technical detection and address how drivers perceive, process, and respond to safety interventions.
The EmotAI project was conceived to address this gap by developing a new generation of driver state monitoring grounded in artificial intelligence, affective computing, and cognitive psychology. The overall objective of the project was to achieve a holistic understanding of the driver’s emotional and attentional state by integrating multiple complementary pillars: attention, activation (emotional and physiological state), driving performance, and external contextual factors. By combining these dimensions, EmotAI aimed to move from fragmented detection towards coherent interpretation of driver state over time.
A central ambition of the project was not only to improve detection accuracy but also to enhance the effectiveness of safety interventions. To this end, EmotAI introduced a cognitive control perspective, focusing on how alerts can be designed to meaningfully influence driver attention and emotional state rather than merely signalling risk. This approach reflects a clear pathway to impact: improved understanding of driver cognition enables better-designed alerts, which in turn can support safer driving behaviour, reduce accident risk, and increase acceptance of intelligent safety systems.
The project is embedded in a broader political and strategic context that promotes trustworthy, human-centric artificial intelligence and safer, more sustainable mobility. By aligning technological innovation with societal needs and regulatory trends, EmotAI contributes to ongoing efforts to strengthen road safety, support the deployment of intelligent transport systems, and enhance Europe’s leadership in responsible AI.
Social sciences and humanities played a key role in the project, particularly through the integration of cognitive psychology and human factors research. Psychological theories of attention, emotion, and cognitive load informed both the interpretation of driver state indicators and the design of cognitively appropriate alerts. This interdisciplinary integration ensured that the project addressed not only what can be detected by AI systems, but also how drivers cognitively and emotionally respond in real driving contexts. As a result, EmotAI positions itself as a human-centred innovation with the potential for significant scientific, societal, and industrial impact at scale.
The project focused on the scientific and technical development of a holistic driver state monitoring framework that integrates artificial intelligence, affective computing, and cognitive psychology. The work performed was structured around three main pillars: scientific understanding of driver state, innovation in cognitive interpretation and alerting, and technical implementation of an integrated system.
From a scientific perspective, the project conducted an in-depth analysis of driver state by modelling four complementary dimensions: attention, activation, driving performance, and external contextual factors. Advanced AI-based methods were applied to extract and interpret driver behaviours and emotional cues using non-intrusive sensing. This work resulted in a consolidated scientific framework that explains how these dimensions interact over time and influence driving safety, providing a robust basis for subsequent system design.
Building on this scientific foundation, the project introduced an innovative cognitive control approach that integrates multiple driver state indicators into coherent safety-relevant information. Rather than treating indicators independently, the project developed mechanisms to combine them temporally and contextually, enabling a more reliable interpretation of driver state. Cognitive psychology principles were incorporated to inform the design of safety alerts, with the objective of maximising their effectiveness in restoring driver attention and emotional balance.
On the technical side, the project achieved substantial progress in the implementation of the EmotAI system. A modular system architecture was defined, and AI models for driver state detection were developed and integrated. Core components of the cognitive alerting mechanism were implemented, ensuring real-time feasibility and extensibility. While large-scale deployment and extended validation are planned beyond the project duration, the project successfully delivered a robust and functional technical foundation.
Overall, the main achievements of the project include the establishment of a holistic scientific framework for driver state monitoring, the development of a novel cognitively informed approach to driver safety alerts, and the implementation of an integrated AI-based system that advances the state of the art in human-centric driver monitoring technologies.
From a scientific perspective, the project conducted an in-depth analysis of driver state by modelling four complementary dimensions: attention, activation, driving performance, and external contextual factors. Advanced AI-based methods were applied to extract and interpret driver behaviours and emotional cues using non-intrusive sensing. This work resulted in a consolidated scientific framework that explains how these dimensions interact over time and influence driving safety, providing a robust basis for subsequent system design.
Building on this scientific foundation, the project introduced an innovative cognitive control approach that integrates multiple driver state indicators into coherent safety-relevant information. Rather than treating indicators independently, the project developed mechanisms to combine them temporally and contextually, enabling a more reliable interpretation of driver state. Cognitive psychology principles were incorporated to inform the design of safety alerts, with the objective of maximising their effectiveness in restoring driver attention and emotional balance.
On the technical side, the project achieved substantial progress in the implementation of the EmotAI system. A modular system architecture was defined, and AI models for driver state detection were developed and integrated. Core components of the cognitive alerting mechanism were implemented, ensuring real-time feasibility and extensibility. While large-scale deployment and extended validation are planned beyond the project duration, the project successfully delivered a robust and functional technical foundation.
Overall, the main achievements of the project include the establishment of a holistic scientific framework for driver state monitoring, the development of a novel cognitively informed approach to driver safety alerts, and the implementation of an integrated AI-based system that advances the state of the art in human-centric driver monitoring technologies.
The EmotAI project produced results that go beyond the current state of the art in driver monitoring systems by shifting from isolated, indicator-based detection towards a holistic, cognitively informed understanding of driver state. Existing solutions typically focus on single dimensions such as drowsiness, distraction, or specific driver actions, often generating fragmented or poorly timed alerts. EmotAI advances this paradigm by integrating multiple complementary dimensions—attention, activation, driving performance, and external contextual factors—into a unified framework that captures the dynamic and temporal nature of driver behaviour and emotional state.
A key result beyond the state of the art is the introduction of a cognitive control perspective in driver safety monitoring. Rather than treating alerts as simple reactions to threshold-based detections, EmotAI incorporates principles from cognitive psychology to inform how and when alerts should be generated in order to meaningfully influence driver attention and emotional regulation. This human-centred approach addresses a critical limitation of existing systems, which often fail to consider the cognitive and emotional impact of alerts on drivers, leading to reduced effectiveness and user acceptance.
From a technical standpoint, the project demonstrated that advanced AI and affective computing methods can be combined within a modular, real-time-capable architecture to support holistic driver state interpretation. The resulting system framework is extensible and adaptable, enabling further integration of additional sensors, models, or contextual information. This positions EmotAI as a flexible foundation for future research and industrial development in intelligent transportation systems.
In terms of potential impact, the project results are relevant to both scientific and industrial communities. Scientifically, EmotAI opens new research directions at the intersection of artificial intelligence, affective computing, and human factors, encouraging further investigation into cognitively informed safety systems. Industrially, the approach aligns with the growing demand for advanced driver assistance systems that are reliable, non-intrusive, and compliant with evolving safety regulations.
To ensure further uptake and long-term success, several follow-up actions are identified. These include extended validation and demonstration in large-scale or real-world driving scenarios, continued research to refine cognitive models and alert strategies, and integration with industrial platforms and vehicle ecosystems. Additional steps such as access to funding for system maturation, evaluation of intellectual property strategies, and alignment with regulatory and standardisation frameworks will support commercialisation and deployment. Collectively, these measures will enable the EmotAI results to transition from research innovation to practical, high-impact applications.
A key result beyond the state of the art is the introduction of a cognitive control perspective in driver safety monitoring. Rather than treating alerts as simple reactions to threshold-based detections, EmotAI incorporates principles from cognitive psychology to inform how and when alerts should be generated in order to meaningfully influence driver attention and emotional regulation. This human-centred approach addresses a critical limitation of existing systems, which often fail to consider the cognitive and emotional impact of alerts on drivers, leading to reduced effectiveness and user acceptance.
From a technical standpoint, the project demonstrated that advanced AI and affective computing methods can be combined within a modular, real-time-capable architecture to support holistic driver state interpretation. The resulting system framework is extensible and adaptable, enabling further integration of additional sensors, models, or contextual information. This positions EmotAI as a flexible foundation for future research and industrial development in intelligent transportation systems.
In terms of potential impact, the project results are relevant to both scientific and industrial communities. Scientifically, EmotAI opens new research directions at the intersection of artificial intelligence, affective computing, and human factors, encouraging further investigation into cognitively informed safety systems. Industrially, the approach aligns with the growing demand for advanced driver assistance systems that are reliable, non-intrusive, and compliant with evolving safety regulations.
To ensure further uptake and long-term success, several follow-up actions are identified. These include extended validation and demonstration in large-scale or real-world driving scenarios, continued research to refine cognitive models and alert strategies, and integration with industrial platforms and vehicle ecosystems. Additional steps such as access to funding for system maturation, evaluation of intellectual property strategies, and alignment with regulatory and standardisation frameworks will support commercialisation and deployment. Collectively, these measures will enable the EmotAI results to transition from research innovation to practical, high-impact applications.