How do conditionally automated cars relate to current traffic safety challenges in Finland? Identifying target accidents for passenger cars with automated driving systems for motorways and urban areas

This dissertation identified the potential of passenger cars equipped with automated driving systems (SAE3) for motorways and urban areas in improving traffic safety in Finland. It includes three sub-studies which investigated the current safety situation from a i) road network, ii) single-driver, and iii) pedestrian perspective. The results were first obtained by identifying the road network on which automated driving systems (ADS) can be used based on their operational design domain (ODD). Then, the expected number of injury accidents, fatalities, and serious injuries on the respective networks was estimated with the Empirical Bayes method.

The results show that motorway ADS in passenger cars could potentially affect a maximum of 191 injury accidents, eight fatalities, and 15 serious injuries a year. Depending on severity, this corresponds to 3.1–3.3% of the national annual average. In four major Finnish cities, urban ADS in passenger cars could potentially affect 127 injury accidents, three fatalities, and 12 serious injuries annually. These correspond to 1.1–2.5% of the national and 17.1–26.8% of the selected cities' annual average. The results also show that the current safety situation is better on the operating networks and within the operating conditions for the ADS compared to other networks and conditions not fulfilling the ODD requirements. For a single driver, the relative accident risk increases as the road weather conditions worsen and the conditions are not anymore fulfilling with the ODD requirements. This increase is largest on motorways meaning that the driver must be in control of and take over the driving task in the most dangerous conditions. Finally, the results indicate that the ADS of passenger cars in this study could have a greater effect on serious injuries than on fatalities among pedestrians.

The requirement for physical separation of driving directions currently restricts the potential to extend the network length for motorway ADS in Finland. The operating environment of ADS, and consequently their safety potential, could be substantially extended if the systems were used on high-level rural roads. Further measures should be put in place to attain traffic safety goals, since the period of transition to ADS is expected to be long. Conditionally automated cars can probably contribute to this process, especially on motorways where other highly effective traffic safety measures have already been implemented.