The Wavelet Analyzing For Traffic Flow Signal Based On Dsp

This thesis mainly studies three following problems. First is to deduce matrix expression of fast wavelet algorithm based on wavelet multi-resolution analysis. The second is using MATLAB to multi-scale decompensate the wavelet signals, then get the high-frequency and low-frequency coefficients, eliminate noise and re-constituted it. The third is to run fast wavelet algorithm on high-speed digital processor TMS320C5402 and analyze and process fact measure traffic wave signal in simulator of integrated development environment CCS . The concrete content is summarized as follows:First, according to the introduction of traffic flow this particular wave, when the road traffic flow data collection system works actually, there will have some unusual value misleading load forecast model establishment, the value will impact prediction result's precision and reliability. So it is an effective way to deal with the traffic volume of wavelet using multi-resolution analysis based on computer simulation. On the basis of wavelet multi-resolution analysis, we analyzed fast wavelet algorithm-Mallat algorithm in detail. For implementing Mallat algorithm in computer, we deduce the matrix expression of fast wavelet algorithm possessing the characteristic of convolution operation and signal cycle expandability that suits to computer advanced language programming.Secondly, using MATLAB to separate wavelet signal decomposition, withdraw the high-frequency and low-frequency coefficients of the multi-criterion wavelet transformation and carry on the high-frequency and low-frequency restructuring of the wavelet decomposition structure. We will obtain the transportation wave primary signal and three decomposition each resolution ratio chart, each resolution ratio restructuring chart and synthesis restructuring chart as well as the signal re-constituted chart after eliminated noise.Thirdly, in order to operate the fast wavelet algorithm in the traffic wave signal processing is feasible, we select the high speed digital signal processor TMS320C5402 chip as the key part of fast wavelet algorithm operation. This thesis introduces the hardware construct traits and the unique assembly instruction system and the integrated developing environment (CCS2) of the TMS320C5402 chip which suits fast wavelet algorithm, and realizes fast wavelet algorithm by the module programming method. We operate the fast wavelet algorithm in the integrated development environment CCS2 which has software simulator (Simulator), and compare the result with the operating result in MATLAB. By compare, we can know the two results are same.The upper study indicates that the wavelet multi-resolution run in TMS320C5402 chip which deal with the traffic wave signal, has the detecting effect of high efficiency and high precision, which can achieve the quite ideal effect.