Design of a dependable model vehicle for rear-end collision avoidance and its evaluation

It is important for constructing a safe automobile society to decrease traffic accidents such as rear-end collision, crossing collision, and right turn collisions. Especially, the number of the traffic accidents can be decreased significantly if the occurrence of rear-end collision can be avoided. The purpose of this paper is to design a dependable model vehicle for rear-end collision avoidance even if an external disturbance is injected. For the purpose of our study, we paid attention to following distance, driving speed of the proceeding vehicle, and driving speed of the succeeding vehicle. We propose a rear-end collision avoidance mechanism based on redundant sensor signal processing, which is designed by combining three kind of sensing information using sensors with different operating principles such as infrared communication for transmitting driving speed of the proceeding vehicle, an ultrasonic sensor for following distance, and a magnetic sensor for driving speed of the succeeding vehicle. In this new concept, since the operating principles of these sensors are different, we consider that the possibility of multiple fault can be greatly reduced. It is shown in the experiment that a fault occurs only in one sensor when the sensors are exposed to sunlight or impulsive sound as an external disturbance. As a result, rear-end collision can be avoided. In conclusion, the use of sensors with different operating principles is valid to avoid rear-end collision.