Characteristic Parameters Identification Of Traffic Flow Based On Acoustic Signal Analysis

With the rapid development of traffic construction,road traffic has become an important link of social and economic development;and traffic control is also moving toward informatization and intelligentialize accompanied by the rapid development of computer and information technology.Traffic flow characteristic parameter information is the basic information among many traffic information of the intelligent transportation system(ITS),while an important prerequisite for traffic intelligent management is to realize the intelligent,informatization and high efficiency of traffic information detection.The traffic information detection technology based on acoustic signal analysis has the advantages of low cost,small information redundancy and small external interference.This detection technique can improve the effect of traffic information detection and facilitate the intelligent detection of traffic information.According to the research of traffic flow parameter identification method based on sound and vibration signals at home and abroad,it can be found that there are two main problems in the existing methods: one is that it's difficult to eliminate the noise interference by using a single vehicle sound signal,which can't realize the identification of traffic parameters.The other is that most of the current research methods need to use multi-target signal,which means more sensors will be needed and the difficulties in synchronization,as well as easy to damage the pavement.In view of the shortcomings of the existing methods,this paper analyzes the acoustic signals of the vehicle through the vocal cords,and proposes an identification method of the traffic flow characteristic parameters based on the vocal signal analysis.The method is as follow: the real-time acoustic signals of the vehicle passing through the deceleration strip is collected by the sensor;The signal is analyzed to determine the target signal and the signal envelope is extracted to further extract the characteristic parameter-peak time;using the distance between the deceleration zone and the crest time to calculate the speed of the vehicle;the vehicle wheelbase is obtained on the basis of the speed,and the vehicle type is classified by comparing with the basic vehicle type wheelbase information;on the basis of speed and classification,the number of vehicles within a fixed time is analyzed to achieve traffic statistics.The main research contents in this paper are as follows:1.Determination of target signalBased on the analysis of the commonly used signal-processing methods like the start-stop detection method,zero-crossing rate method,short-term energy method and double-threshold method,a method of target signal detection is proposed in the paper : the fastest rising edge detection method.2.Signal Analysis: Feature Selection and ExtractionDenoising preprocessing of the collected acoustic signal and analyzing the characteristics of the signal and the features of the commonly used signal and the characteristic parameters.An algorithm for the extraction of the acoustic signal envelope based on the transformation step is proposed,which is compared with the commonly used envelope algorithm.Finally,combining the threshold and extremum to determine the time-domain characteristic parameter-the peak point.3.Vehicle parameter identificationVehicle parameters identification mainly in aspects of the speed judgment,vehicle classification and traffic statistics.The feasibility of the method for identifying the characteristic parameters of traffic flow based on vocalization signal analysis is verified by experiments.In this study,the signal acquisition only needs a kind of equipment,which overcomes the shortcomings of the synchronization in the multi-sensor detection environment.The proposed method is easy to obtain the sound waveform with the deceleration zone,and mainly in the time-domain to fulfil the vehicle speed measurement and vehicle identification,besides,it is simpler.