基于改进BP算法神经网络的路况评价模型研究

通过利用在BP神经网络中增加反馈信号及偏差单元的网络模型,生成内部回归网络,对BP神经网络进行改进,据此得出改进的BP神经网络的具体步骤.在此基础上,结合陕西省高速公路沥青路面的实际情况建立了路况评价模型,并对该模型的具体应用做了研究.研究结果表明:网络的收敛速度很好,训练结果与实际路况结果相差很小,利用改进BP算法神经网络建立的路况评价模型不仅方便易用,而且精度很高.

Study on Road Condition Evaluation Model Based on Improved BP Calculation Neural Networks

The internal regression networks are set forth based on the application of feedback signals and deviation unit onto the neuaral networks.The specific study procedure of the BP neural networks is put forward.Furthermore,the road condition evalution model is made on the practical condition of the expressway bituminous road by the specific operation of the model.The study concludes contain that: the deviation between training result and practical result is small;the condition evaluation model based on the improved BP calculation neural networks is not only convenient to handle,but also more accurate.