时间递推耦合神经网络的交通路径动态诱导技术

通过"人—机—环境"耦合关系,对路况与时间变化关系进行研究。综合车辆运行过程中不同时段的路况差异和人因作用、突发道路事故随机性,以神经网络有师学习作为经验累积方法,提出时间递推预测方法确定路径最短时间,从而实现对交通路径的动态诱导。递推预测以知识库累积经验与实时路况信息作比较,为驾驶者提供实时有效的路况信息支撑。结果表明,该诱导技术可辅助驾驶者及时对路况作出正确判断,减少因经验不足和突发事件造成的时间损失,适用于安装有GPS导航的车辆。实例分析表明,所构建模型与实际数据结合收到良好效果。

Traffic route dynamic guidance technology based on coupling of time recursive and artificial neuron network

This paper researched the relation of road conditions and time on the basis of people-machine-environmental coupling, and proposed a prediction method of time recursive to confirm the shortest time of route. This method was as an accumulated experience basing on the idea of supervised learning in artificial neural network, colligating with the difference of road conditions during different time section, the human factors function, and the randomness of the accident in course of driving, thus realized the guidance of the traffic route. Comparing with the real-time road conditions and accumulated experience, the method of time guidance prediction could offer real-time and effective road information for drivers. This guidance technology assists drivers to judge correctly in time and reduces the time losses because of the lack of the experience and the accidents. The guidance technology can be applied to the vehicles, which is with GPS. The example indicates that the model is effective combined with the real data.