ADAS TESTING IN CANADA: COULD PARTIAL AUTOMATION MAKE OUR ROADS SAFER?

摘要

As part of an ongoing effort to further improve the safety of its road transportation system, Transport Canada (TC) has been evaluating Advanced Driver Assistance Systems (ADAS) for a number of years now. The main objective of this paper is to determine the potential of ADAS technology in reducing fatalities and injuries on Canadian roadways while using proven international test protocols and certified test equipment. The findings will be used to provide science-based evidence in support of future regulatory, research and policy development. Results from this study clearly demonstrated that Automatic Emergency Braking (AEB) and Pedestrian AEB (P-AEB) technologies can provide significant improvements in terms of collision mitigation which can directly result in reduced road fatalities and injuries. These findings are also in line with those of studies based on real-World data predicting significant reductions in rear-end collisions due to AEB deployment. Nonetheless, this ADAS program also exposed an important number of flaws and performance variability. While the best AEB and P-AEB systems were able to fully avoid collisions with vehicles and pedestrians at speeds up to 60 kilometers per hour (km/h), others were challenged at speeds below 10 km/h. Also, a few P-AEB systems were never able to avoid a collision with a pedestrian despite manufacturers' claims of pedestrian avoidance capabilities. Scenarios replicating AEB activation in moving traffic showed that most systems unnecessarily came to a full stop rather than match the speed of vehicles they detected on their path, potentially generating higher safety concerns than those they were designed to prevent in high density traffic. Finally, due to variability in test results and overall unpredictable system behaviour, it was not possible to gather enough data to confidently assess the potential safety benefits associated with Lane Support Systems (LSS). AEB, P-AEB and LSS are essential components of automated driving systems which will need to reliably brake and steer at all time to safely avoid other road users. That level of performance is not yet evident from the extensive testing carried out within this project. Substantial progress is therefore needed to reach the level of detection, braking and steering performance that will be required to make commercial automated driving systems a reality.