Studying the influence of trajectory and steering model choice on crash avoidance timing in simulation of rear-end conflicts that include consideration of drivers’ comfortable steering potential

Abstract

The World Health Organization states that road crashes are expected to be the seventh leading cause of death in 2030, with road traffic injuries being the leading cause of death for children and young adults aged between 15 to 29 years of age, and the most contributing factor in road accidents is human error. Passive safety features in vehicles help depreciate serious injury to protect the vehicle occupants in a crash scenario. Some examples of this include seatbelts, crumple zones, airbags, headrests, etc. Although this has been found to be a very effective method, vehicle manufacturers are aiming to assist the driver and alleviate collisions which has resulted in an increase in the development of active safety systems which use sensors to help avoid or mitigate crashes by understanding the state of the vehicle surroundings and the driver. The modern approach to active safety from automotive industries involves building autonomous vehicles to take into account people’s vulnerability to crashes and to be forgiving of common human errors. With respect to assessing active safety systems before the systems are on the market, virtual simulations are gaining momentum when compared to driving thousands of miles for evaluation of active safety functions. But this method of virtual assessment sometimes lack the knowledge of how complex a vehicle model needs to be in order to provide sufficiently accurate results. The aim of this thesis is to study the impact of trajectory and steering model choice on crash avoidance timing by taking into account the drivers' avoidance by steering in a simulation of rear-end conflicts. Three continuous steering avoidance trajectories (Virtual Desired Trajectories; VDTs), based on previous work, is used together with a single-track vehicle model with load transfer effect. A PID controller is used to track these VDTs in simulations applied to the rear-end crashes. The performance metrics in terms of collision avoidance time is evaluated for all three VDTs. Also, a sensitivity analysis is made in three parts, namely, without load transfer effect, increasing the cornering stiffness of the tires and finally by decreasing the cornering stiffness of the tires. It was observed that there were no obvious effects on the overall results in the first two simulations, but some effects were noticeable in the third simulation. The results of the three VDTs when compared, showed an observable difference.