Research Of Online Statistical Algorithm About Traffic Volume Based On Running Vehicles Noise

Statistics traffic is one of the important elements that determines the highway construction,is the important basis of evaluation on the national economic and the financial evaluation.It affects the development scale of Highway.Accurate statistics on the traffic volume has an important value that will increase reliability of the project investment decision.In fact.identification is the basis of traffic statistics.Now,The study for Vehicle Recognition focused on induction coil,infrared detection,pressure sensors,etc,at home and abroad Their disadvantage are inconvenient about the installation,and need to regular maintenance.The image research whose equipment is expensive contain a large of information.Therefore,this paper presents a new method on the vehicles noise for the traffic volume statistics.Distinguishing on Vehicles Noise is convenient management and small capital investment.It has the widely value on the automated statistics and traffic contr01. It is an important part of the paper that using parameter model processes and analysises the vehicle noise signal.Firstly,Seeking AR model parameters with Burg.Secondly,We research a new methOd——DelaVed Autocorrelation.The results prove the advantage of the new method.Another part of the article is how to achieve the online traflfic statistics.If the research successed,it will be applied to practice.The implementation idea is:1 detect the endpoint of the vehicles noise.2 Cutting goal flame.3 Extract feature parameters.4 classify with the BP neural network.5 calculate traffic volume.Using the matlab software does this function.The experimental results show that the method of Distinguishing on Vehicles Noise is low cost and easy to realize.