AD1

Monday, November 1, 2010

AR model of vehicle road spectrum Research and Implementation of Reconstruction

To speed up the pace of China's automobile self-developed, so that conditions for the development of automotive products, in line with China's road conditions, the National High Technology Research and Development Program (863 Program) set up a special - the statistical measurement of a typical vehicle road spectrum and application system development, the purpose of Establishing a typical road vehicle to explore the spectrum of databases and research applications. The research paper is part of the project, focusing on the road spectral reconstruction theory and methods to establish the basis for the virtual road. This article analyzes the characteristics of road roughness, road roughness in the summary of the reconstruction of the existing mathematical model based on the AR model is improved. Burg algorithm proposed by using the order of AR model were identified and by controlling the power spectrum fitting error to optimize the AR model order. Furthermore, using this method the principle of the spectral equivalent AR model for the standard road reconstruction, with the existing method for determining the order of (partial autocorrelation function, Akaike information criterion) than the improved reconstruction accuracy than the AR model greatly improved. Which proves that the Order to determine the feasibility and superiority. At the same time, the use of the method for determining the order of, on the road for one-dimensional AR model reconstruction. Comparative analysis of the AR model spectrum estimation with the traditional spectral estimation, the road through a standard known AR model spectrum estimation spectral estimation error of regression was significant and should be better than the traditional spectrum estimation. Established a two-dimensional surface - AR model and in-depth study of two-dimensional method for solving the AR model parameters. Two-dimensional AR model for solving the parameters involved in the process of decomposition of complex Toeplize-block Toeplitz matrix problems, Jacobi algorithm using two-dimensional autocorrelation matrix decomposition, which will reduce the problem two-dimensional AR model is one-dimensional AR model problem.At the same time using one-dimensional filter to form two-dimensional filter, thus simplifying the AR model of two-dimensional reconstruction of the road. And using the algorithm of the standard and the actual level of two-dimensional reconstruction of the road, can get high accuracy, thus proving the feasibility of this method and have good stability. 

Key words: road roughness, power spectrum, AR model, reconstruction, correlation index

 

No comments:

Post a Comment