Based On Road Traffic Safety Of The Bp Neural Network Prediction

The road safety prediction is the important content of the research of traffic safety. The purpose of road safety forecasting is to analyze the tendency of traffic accidents under existing traffic conditions, evaluates the feasibility and practical effectiveness of road safety measures reasonably, controls the factors affecting road accidents, and reduces the traffic accidents. The characteristics such as nonlinearity, randomness and uncertainty in traffic system make it difficult to forecast road accidents, the behavioral feature of traffic system. Based on the analysis of existing macroscopic forecasting methods of road accidents, the forecasting methods of road safety based on BP Neural networks is researched in this dissertation. The main research contents and conclusions are as follows:1. The paper introduces the current situation of traffic safety both at home and abroad; describes the background of research that the status of traffic safety is very intense.2. Analyses the causes of accident from several aspects, such as driver, vehicle, road and surroundings, as well as expatiates the relation between road accidents and its causes, all this work lay a foundation for establishing a prediction model.3. After the contrast of several relatively representative targets of the traffic safety estimate, it brings forward that traffic safety integrative mortality is more appropriate. Then this paper chooses nine factors which the relationship with road safety is very tight according to the gray relevancy theory as the influence factors to road safety. Lastly, builds up a mathematic model of road safety and its influence factors.4. This dissertation builds up a BP Neural networks model of road safety prediction, and discusses some key technique and means applying the model, including the selection and pretreatment of swatches, the selection of input-output variables, conforming the number of the nodes in hidden layers, the selection of the initial weight and value, the selection of activation function, training arithmetic as well as parameter.5. At last, through the example of the road safety prediction, it validates that this method opens out the relation of the road safety and its influence factors in some error bound ,and could be applied to the road safety predict.