Road roughness detection by analyzing IMU data

Abstract: Nowadays, people are resorting to many advanced measuring systems and methods to detect road roughness, among which, this paper is proposed to find out road roughness information from a popular mobile measuring/mapping system GPS/INS. Investigation of the IMU signal of the INS is focused for purpose of mining its ability of expressing road roughness. The bumps on road and road texture are used as two indicators for describing road roughness. Both time domain analysis and frequency domain analysis of IMU data are performed for detecting the bumps and road texture. Based on the idea that road bumps generate signal bumps from sensor, the location and magnitude of bumps are figured out by removing noisy signal bumps and extracting signal bumps caused by road bumps. The detection of road texture is basically based on frequency analysis, and the result is then used as the input for roughness classification. Three different types of classifiers such as fussy logic classification, distance based classification, and maximum likelihood classification are tested in this research. The results from fuzzy logic and distance based classification prove to be very good, but the maximum likelihood classification is considered as an unsuitable method in this case. Lastly, road roughness detected from IMU data is visualized on map. The results of this paper demonstrate that IMU data has a great potential for revealing the road roughness.