Driver behavior analysis based on Bayesian network and multiple classifiers

• Published in: 2010 IEEE International Conference on Intelligent Computing and Intelligent Systems Vol. 3; pp. 663 - 668
• Main Authors: Guoqing Xu, Li Liu, Zhangjun Song
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.10.2010
• Subjects: Acceleration; Adaptation model; Bayesian network; driver behavior model; Driver circuits; Hidden Markov models; multiple classifiers; support vector machine
• ISBN: 9781424465828; 1424465826
• DOI: 10.1109/ICICISYS.2010.5658384

Driver behavior model is one of the key technologies for the driver assistance and safety system which can provide useful priori knowledge for detecting the deviant and dangerous behavior. This paper proposes the hybrid model based on Bayesian network and multiple classifiers of support vector machine to analyze and recognize the driver behavior and the limited and observable features of driver behavior are extracted in the model. In addition, the relationship between the features and driver behavior is analyzed. The effect of data loss on the hybrid model is also analyzed. Finally, the hybrid model is compared with support vector machine. Experiment results show that the hybrid model can achieve better accuracy and stability.