Characterizing car to two-wheeler residual crashes in China

Abstract

The fast development of vehicles not only benefits peoples’ lives, but also threatens peoples’ health in the road traffic. The Automatic Emergency Braking (AEB) system is one effective active safety system in saving lives on the road. This study proposed one AEB algorithm and it was implemented to car-to-two-wheeler crashes to evaluate the performance of the AEB, with the aim to analyze the characteristics of the remaining crashes. The algorithm was based on comfort braking and steering limits of car drivers and two-wheeler drivers, and the algorithm simulation was composed of the future path prediction of both vehicles, the braking maneuver and the steering maneuver of car drivers and two-wheeler drivers. The simulation of the future path prediction was used to check whether the car and the two-wheeler were on the collision course and the brak-ing and steering maneuvers were used to assess the collision avoidance ability of both drivers. The proposed algorithm was triggered only when the collision danger was de-tected and both drivers could not avoid the collision by steering or braking on their own within their comfort limits. The reference algorithm, which utilized the vehicle braking limitation and did not involve the collision avoidance ability of the drivers, was only applied to compare the effectiveness in collision avoidance. Both AEB algorithms were applied to the pre-crash-matrix (PCM) of the China Shanghai United Road Traffic Sci-entific Research Center (SHUFO) crash data. It was found that the proposed algorithm was triggered later than the reference algorithm in about 50% of cases. In these cases, the drivers may feel it was unnecessary to activate the AEB when the reference algo-rithm was triggered, as they were still able to avoid the collision by their own action or by the action of the collision partner. To evaluate the injury mitigation, one available motorcyclist injury model from previous research was used in the study. The injury mitigation was studied based on the three levels of injury: MAIS2+F, MAIS3+F and fatal injury. The results indicated that the effectiveness in injury mitigation for fatal injury is around 60% and for MAIS2+F and MAIS3+F, around 50% after the proposed AEB implementation. The proposed algorithm is applicable to all types of car-to-two-wheeler scenarios. The crash data was classified into nine types of scenarios and the simulation results showed that the effectiveness of the proposed AEB algorithm in col-lision avoidance varied across scenarios. Straight moving car scenarios have a higher proportion of residual crashes compared to the turning car scenarios. This may be due to the differences in the cars traveling speed.