The variable of road's slope could be online calculated by vehicle's NED velocity obtained from one GPS based vehicle attitude measure system. In order to wipe off measured noise and to identify the longitudinal road curve parameters more exactly, the wavelet analysis method was applied in the data processing. Using Symlets wavelet function, the row gradient data were divided into one approximate signal and ten level details signals which present the road's slope characteristics and the

more »

... noise respectively. The design parameters of longitudinal road curve were obtained from approximate signal by using least square curve fitting method. The analysis results of survey data of ring road in Xi'an show that it is efficient to eliminate the measured noise by wavelet tools and the characteristic parameters of longitudinal road curve can quickly and accurately be identified.