The Detection System Of Pavement Roughness Based On Multi-sensor Data Fusion

In recent years, the infrastructure construction has received fast development in China, especially in the highway traffic. However, the workload of road inspection and road maintenance are becoming heavy due to the increase of various kinds of roads such as expressway, national roads, provincial roads, county roads and cities and towns etc. Before the inspection and maintenance, the road conditions must be detected and evaluated objectively and accurately in order to provide the test data and evaluation results to the relevant department. The roughness of pavement accounted for about 20% of the total evaluation score in the many evaluation indexes of the road. So it is very important to obtain the roughness of the road under different environment.Firstly, this paper introduces the current development of roughness detection technology at home and abroad. Then, according to the existing problems we put forward a kind of road pavement roughness detection system, which can meet different needs including the wavelength detection at the same time, can be slow, variable speed, quickly detect the road. In order to realize the full-band detection of longitudinal section, the pavement wavelength is divi ded into large one and small one, which has realized full-band wave detection of surface profile with the information fusion processing technology. At last, this paper analyzes the wavelength response of the international roughness, and the response curve shows the requirement of wavelength.In order to detect the roughness in the different type speed including slow speed, variable speed, and quick speed, this paper adopts the non-inertial vertical section detection theory based on a small benchmark transfer principle to test the small wavelength of pavement profile. Besides, we use the GPS elevation data and gyro range and high data fusion method to achieve accurate detection of large wavelength.Before the fusion of small wavelength and large wavelength, firstly, for pavement longitudinal section of small wavelength we removed the pavement trend processing and with regard to data splicing problem we proposed an adaptive method of local signal benchmark adjustment, this method is to obtain the smooth and accurate surface profile curve of small wavelength. Secondly, we use Kalman filter algorithm to get the accurate surface profile curve of large wavelength of the GPS difference measurement system and gyro measurement system. Finally, overlay the vertical section of road surface of small wavelength curve onto large wavelength curve, we can get pavement longitudinal profile curve and analyze the pavement smoothness from the curve.Based on the above detection technology, the software and hardware experimental platform are set up. The experiment of smoothness detection is carried out repeatedly under the low-speed in outdoor road, and the relevant analysis experiment of leveling instrument is made, too. The experimental results show that the detection system is stable, reliable and has great practical value.