交通拥堵汽车流量准确预测仿真研究

发布时间：2018-12-19 21:11

【摘要】：交通拥堵汽车流量的准确预测,能够有效改善现有的道路网络的通行能力。对车流量的预测,需要利用模糊推理系统将交通流分为不同的模糊集,给出模糊隶属度函数,完成交通拥堵汽车流量的预测。传统方法先选出相关性较高的道路断面,得到路网空间关系归一化约束矩阵,但忽略了给出交通隶属度函数,导致预测精度偏低。提出基于模糊神经网络的拥堵汽车流量预测方法。将神经网络和自适应卡尔曼滤波模型相结合,组建拥堵汽车流量预测模型,将任意时刻的交通流预测期望值作为模型的输入,将实际预测误差作为模型的输出,利用模糊推理系统将交通流分为不同的模糊集,给出模糊隶属度函数,将小波函数作为模糊推理系统的输入,利用神经网络实现模糊推理,借鉴遗传理论优化模糊神经网络,完成对拥堵汽车流量的预测。实验结果表明,所提方法预测精度较高,具有较高的预测准确性。
[Abstract]:Accurate prediction of traffic congestion can effectively improve the capacity of the existing road network. To predict the traffic flow, the traffic flow should be divided into different fuzzy sets by fuzzy inference system, and the fuzzy membership function should be given to forecast the traffic congestion. The traditional method firstly selects the road section with high correlation, and obtains the normalized constraint matrix of the road network spatial relationship, but neglects to give the traffic membership function, which leads to the low prediction accuracy. A fuzzy neural network based traffic prediction method for congested vehicles is proposed. Combining the neural network with the adaptive Kalman filter model, the traffic flow prediction model of congested vehicle is constructed. The expected value of traffic flow prediction at any time is taken as the input of the model, and the actual prediction error is taken as the output of the model. Traffic flow is divided into different fuzzy sets by fuzzy inference system, fuzzy membership function is given, wavelet function is used as input of fuzzy inference system, fuzzy inference is realized by neural network, and fuzzy neural network is optimized by using genetic theory. Complete the forecast of congested car flow. The experimental results show that the proposed method is accurate and accurate.
【作者单位】：江西科技师范大学数学与计算机科学学院;
【基金】：江西省科技计划指导性项目(2015ZBAB201007) 江西科技师范大学校级科研重点项目(2016XJZD006) 江西省高校人文社会科学研究项目(TQ1505)
【分类号】：U491.14

【相似文献】