BERTHA has released Professional Publication #1 | Summary Report #RTR2025, distilling the most relevant lessons from the Results from Road Transport Research (RTR) 2025 conference and highlighting how human‑centric approaches to driver behaviour modelling and simulation can accelerate safer, predictable automated mobility across Europe.

Read it on our website: https://berthaproject.eu/professional-publication-1-summary-report-rtr2025/ and access to the full report in: https://berthaproject.eu/wp-content/uploads/2025/10/RTR2025-Summary-all-sessions-FINAL.pdf.

Why this matters

RTR is the EU’s annual stage for sharing road‑transport R&I outcomes. The 2025 edition brought record participation (over 900 attendees and 90+ projects) and featured sessions on human‑centric CCAM and driver safety, making it a prime moment to align on indicators, scenarios and validation practices that other projects can adopt quickly.

For BERTHA, whose goal is a probabilistic Driver Behavioural Model (DBM) and CARLA integration to make automated vehicles more human‑aligned and acceptable, turning conference insights into a plain‑language, reusable summary supports strategic communication to audiences beyond our consortium—exactly what Horizon Europe expects.

What BERTHA delivered

• A concise, public‑facing report that captures the essence of our contributions to the Human‑centric CCAM and driver safety track and synthesises takeaways for practitioners, cities and industry.
• Context on simulation‑based validation pathways (including CARLA), showing how behavioural modelling and human‑state monitoring can lead to safer, more predictable automated behaviour in mixed traffic.
• Links and pointers that help peers find and reuse materials more easily, supporting comparability and reducing duplication across the CCAM ecosystem.  

Key takeaways from RTR 2025 (for human‑centric CCAM)

• Comparable behavioural indicators and scenario taxonomies are essential to scale validation beyond single projects; BERTHA’s summary proposes practical alignment points that others can adopt.
• Trustworthy/explainable AI, together with ethical/legal frameworks (e.g., type‑approval processes), should be embedded early in human‑centred design to build societal trust.
• Open, plain‑language reporting (like our summary) improves discoverability and speeds up knowledge transfer from events to real‑world testing and policy.