基于灰色神经网络的点速度预测模型

为了克服交通流时空不稳定性导致的检测数据误差,提高预测点速度的精度,在比较传统灰色预测模型和反向(BP)神经网络预测模型优缺点的基础上,建立了灰色神经网络点速度预测模型.该模型综合了灰色预测模型所需数据少及神经网络具有的自学习和自适应能力的特点.以实测值作为输出数据,构建不同的灰色预测模型,将各灰色预测模型的预测结果作为BP神经网络训练的输入数据,得到最佳的预测模型.实例分析表明:与传统灰色理论及BP神经网络预测模型相比较,在20、40和60s采样间隔条件下,本文模型预测结果与实测值的相对误差平均减少了32％,为交通运行状态评价和行程时间预测提供了依据.

Spot Speed Prediction Model Based on Grey Neural Network

To overcome the detecting data error due to the temporal and spatial instability of traffic condition and improve the accuracy of spot speed prediction,a spot speed prediction model based on grey neural network was developed on the basis of grey prediction model and BP(back propagation) neutral network.The model combines the characters of low data demand of grey prediction model and the self-learning and self-adaptive abilities of BP neutral network.It uses field data as output to build different grey prediction models,and then the predicted results are used as inputs to train the BP neural network to obtain the optimized model.Case study shows that compared with those of the traditional grey theory and BP neural network models,the average relative deviation between predicted and field data at 20,40,60 s sampling intervals can decrease 32% on average using the proposed model.Therefore,the proposed model can be used as a basis for traffic condition estimation and travel time prediction.