Research on highway driving risk perception based on vehicular ad hoc networks
This paper proposes a method for highway driving risk perception based on a hybrid rough set and genetic algorithm optimized support vector machine (SVM) model in vehicular ad hoc network environments. By preprocessing and feature extraction of collected basic safety message (BSM) data,and optimizing SVM parameters using a genetic algorithm,an efficient driving state discrimination model is constructed. Real-world testing results show that the model has high accuracy in detecting potential risky driving behaviors such as abrupt acceleration,abrupt braking,lane changing,and bumpiness,with a classification accuracy of 94.4444％,surpassing traditional methods,verifying its effectiveness and robustness in practical applications.
vehicular ad hoc networkshighwaydriving risk perception
万方数据
维普
