Journal
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
Volume 17, Issue 11, Pages 3087-3097Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TITS.2016.2537878
Keywords
Collision warning system; perception-reaction time; nonparametric method; multilayer perceptron neural network
Categories
Funding
- Korean Ministry of Land, Infrastructure and Transport (MOLIT) through the U-City Master and Doctor Course Grant Program
Ask authors/readers for more resources
The existing rear-end collision warning systems (CWS) that involve the variable perception-reaction time (PRT) have some negative effects on the collision warning performance due to the poor adaptive capability for the influence of different PRTs. To deal with the related problems, several studies have been conducted based on nonparametric approaches. However, the previous nonparametric methods are of doubtful validity with different PRTs. Moreover, there is a lack of consideration for the criterion to split the real-time data into training and testing sets in terms of enhancing the algorithm performance. In this paper, we propose multilayer perceptron neural-network-based rear-end collision warning algorithm (MCWA) to develop a real-time CWS without any influence of human PRTs. Through a sensitivity analysis, the optimal criterion for splitting real-time data into training and prediction is found in terms of a tradeoff between training time and algorithm accuracy. Comparison study demonstrates that the proposed algorithm outperforms other previous algorithms for predicting the potential rear-end collision by detecting severe deceleration in advance.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available