Bp Artificial Neural Network-based Road Safety Evaluation Studies

The road safety evaluation is an important part of the traffic safety research. The existing traffic safety evaluation most based on traffic accident data but not the road environment; Otherwise, the traffic safety evaluation is usually in the large area but not the road; the methods used most are traffic accident rate and step analysis. For this situation, this paper advanced a new evaluation that uses the factors that influence the road safety.First, analyze and screen the factors that influence the road safety, and then ascertain the index of the traffic safety evaluation. After analyze and compare with the methods of traffic safety valuation, use the BP Artificial Neural Network (ANN) to evaluation, use the MATLAB to create the ANN. Make the ripe evaluated road as a sample to train the BP net. Meanwhile, uses the GUI from MATLAB achieve the safety evaluation. The result indicate that this method improve the efficient of the evaluation. The data from training the BP net and example evaluation is accord with the objective reality.The road traffic safety evaluation that based on ANN also can used in forecast the traffic accidents; check up the traffic accident black point and so on. It has a very important significance in traffic safety research.