Research Question / Objective: Advanced Driver Assistance Systems (ADASs) such as Forward Collision Warning have been developed for light passenger vehicles (LPVs) to avoid and mitigate collisions with other road users and objects. These technologies may have contributed to a reduction in LPV traffic fatalities in the EU and US. However the number of powered two wheeler (PTW) fatalities has remained relatively constant in the US. To fully realize the potential safety benefits across all vehicle categories, LPV crash avoidance technologies also need to be effective in avoiding collisions with PTWs. To accomplish this, knowledge of the pre-crash LPV-PTW vehicle trajectories and conflicts is needed to guide the development and testing of effective crash countermeasures for both LPVs and PTWs.
Methods and Data Sources: Crash scenario database development tools previously developed to evaluate LPV- LPV crash countermeasure effectiveness have been extended to LPV-PTW crash scenarios. This involved using information for a large sample of LPV-PTW crashes from the EU Motorcycle Accidents In-Depth Study (MAIDS) and US Motorcycle Crash Causation Study (MCCS) databases, which are based on in-depth crash investigations and the Organisation for Economic Co-operation and Development (OECD) Common Methodology. The vehicle pre- crash trajectories were estimated based on the coded data and digitized information from the scaled pre-crash scene diagrams. The pre-crash conflict state was then analyzed based on these trajectories.
Results: The estimated pre-crash trajectories using this method indicate that LPV-PTW pre-crash trajectories and conflicts in France, Germany, Italy, and the US have many similarities, but there are some differences as well. These results indicate that conflicts in several types of pre-crash scenarios, such as the LPV turning across the PTW path in the same direction or opposite direction, begin less than 1.5 sec before impact, which may not be sufficient time for some crash countermeasures based on conflict detection and driver warnings to be effective.
Discussions and Limitations: The accuracy of the results is based on a number of assumptions, approximations, and limitations in the data and methods used. These include the accuracy and representativeness of the data based on in-depth crash investigations, as well as the domain-of-validity and accuracy of the vehicle directional control models used.
Conclusion and relevance to session submitted: Analysis of real world accident data is critical to the development and evaluation of ADAS and automated driving systems. This analysis has shown that LPV-PTW crash countermeasures need to function with shorter pre-crash conflict epochs, or in the pre-conflict phase, in order to be effective in preventing collisions. This information may help to define requirements for LPV-PTW crash countermeasures (e.g., C-ITS V2V and Blind Spot Detection), evaluate their effectiveness, and inform the development of performance confirmation tests (e.g., New Car Assessment Programs).