Integration of human driving behaviour in the validation of CCAM systems (CCAM Partnership)

Project results are expected to contribute to all of the following expected outcomes:

• Validated human behavioural models representing the variety of human driving behaviour in safety-relevant scenarios, shared through a common repository and to be used:
  • to define pass criteria/ assessment criteria for CCAM systems in type approval schemes, consumer testing campaigns and industrial development processes;
  • to design safe, human-like behaviour of CCAM systems that can be anticipated easily by all road users and is acceptable to both CCAM vehicle occupants and all road users.
• Application of such human behavioural models in the virtual safety validation of CCAM systems to realistically represent the behaviour of human-driven vehicles in closed loop simulations of mixed traffic, thereby reflecting the variety of human driving behaviour, including behaviour in complex real-world and emergency conditions.

The deployment of CCAM systems in mixed traffic will mean intense interaction with all road users such as the human drivers of other vehicles as well as pedestrians and riders of two-wheelers. These interactions (including implicit and explicit communication by humans and CCAM systems) will play a crucial role in the acceptance and thereby the penetration of CCAM systems in future road transport. CCAM systems will have to show safe and human-like driving behaviour, so that their decisions and actions can be anticipated easily by all road users, respecting the variety of typical driving behaviour across different countries as well as the need for CCAM systems to respect traffic rules and support road safety.

This will require validated models of explicit and implicit human driving behaviour to design and validate such system behaviour. These models will be needed in closed loop simulations of CCAM systems in mixed traffic to realistically represent the reactions of human drivers in other vehicles to the behaviour of a CCAM system. Models representing human driving behaviour are being developed by the projects i4Driving^[1] and BERTHA^[2] for selected fields of application, i.e. they will be calibrated for a limited number of scenarios. Bringing together and building upon the results of these projects – in particular a simulation library and an innovative methodology to account for uncertainty from i4Driving and a scalable, probabilistic driver behavioural model from BERTHA, research is needed to extend the fields of application that these projects are addressing with a focus on representing driver behaviour in a multitude of safety-critical scenarios, considering the variation and statistical distribution of human behavioural patterns and the factors influencing such behaviour, including the parallel execution of non-driving related tasks.

To achieve high degrees of robustness and applicability in a wide range of scenarios, detailed calibration and parameterisation is necessary, as driver behaviour depends on factors such as the road infrastructure, vehicle types, traffic conditions and rules, as well as regional influences and driver experiences / demographics, e.g., gender, age and other relevant social variables. Considering the deviation of average from ideal human driving behaviour, proposed actions must also validate the models for their extended fields of application, going well beyond the applications and degrees of validation accomplished by the above-mentioned projects under HORIZON-CL5-2022-D6-01-03. Proposed actions are thus expected to raise the technology readiness of such models to TRL 5. Data for parameterisation and validation should be captured by monitoring real human drivers in driving simulators and/or real traffic considering what is happening inside and outside the vehicle.

Proposed actions should integrate, to the extent possible, the validated models in the virtual validation and verification approaches as developed in the projects HEADSTART^[3] and SUNRISE^[4] and complemented by the project SYNERGIES^[5]. Successful integration should be demonstrated in various safety-relevant scenarios as provided by the action(s) funded under HORIZON-CL5-2023-D6-01-02^[6]. Models should be shared via the federated data exchange platform for CCAM to be developed by an action under HORIZON-CL5-2025-D6-06^[7].

Proposals are encouraged to also explore additional fields of application of validated driver behaviour models, while the integration of relevant expertise from social sciences and humanities (SSH) is expected.

To achieve the expected outcomes, international cooperation is encouraged with research stakeholders in Japan and the United States but also with other relevant strategic partners in third countries. Such cooperation should exploit synergies as far as possible in capturing data for the parametrisation and validation of behavioural models, while considering regional and cultural differences as well as specificities of respective road infrastructures.

This topic implements the co-programmed European Partnership on ‘Connected, Cooperative and Automated Mobility’ (CCAM). As such, projects resulting from this topic will be expected to report on results to the European Partnership ‘Connected, Cooperative and Automated Mobility’ (CCAM) in support of the monitoring of its KPIs.

Projects resulting from this topic are expected to apply the European Common Evaluation Methodology (EU-CEM) for CCAM^[8].

Projects funded under this topic are encouraged to explore potential complementarities with the activities of the European Commission's Joint Research Centre’s Sustainable, Smart, and Safe Mobility Unit and, where appropriate, establish formal collaboration.

[1] Integrated 4D driver modelling under uncertainty, grant agreement ID: 101076165.

[2] BEhavioural ReplicaTion of Human drivers for CCAM, grant agreement ID: 101076360.

[3] Harmonised European solutions for testing automated road transport, grant agreement id: 824309.

[4] Safety assUraNce fRamework for connected, automated mobIlity SystEms, grant agreement ID: 101069573.

[5] Real and synthetic scenarios generated for the development, training, virtual testing and validation of CCAM systems, grant agreement ID: 101146542.

[6] Generation of scenarios for development, training, virtual testing and validation of CCAM systems

[7] Federated CCAM data exchange platform (see below).

[8] See the evaluation methodology here.