An Adaptive Vehicle Rear-End Collision Warning Algorithm Based on Neural Network

• Published in: Information and Management Engineering Vol. 236; pp. 305 - 314
• Main Authors: Wei, Zhou, Xiang, Song, Xuan, Dong, Xu, Li
• Format: Book Chapter
• Language: English
• Published: Germany Springer Berlin / Heidelberg 2011; Springer Berlin Heidelberg
• Series: Communications in Computer and Information Science
• Subjects: adaptive; nerual network; rear-end collision; warning algorithm
• ISBN: 3642240968; 9783642240966
• ISSN: 1865-0929; 1865-0937
• DOI: 10.1007/978-3-642-24097-3_46

Most of the existing algorithms of vehicle rear-end collision have poor adaptive, high false alarm and missed alarm rates. A two-level early warning model based on logic algorithm of safe distance is discussed. The influence of road conditions, driver status and vehicle performance on the warning distance of rear-end collision in the driving process is analyzed. And for different driving conditions, a warning algorithm of vehicle rear-end collision based on neural network with adaptive threshold which can adapt to different status of the three main elements, human-vehicle-road is proposed. Also the comparison of the warning distance whether using adaptive strategies for the rear-end collision algorithm through changing the real-time status of human-vehicle-road is presented. The result of the simulation shows that the algorithm proposed is self-adaptive to the warning distance and region, and the feasibility of the algorithm is verified.