Research On Algorithm For Identifying Specific Road Information Based On Multi-sensor

With the improvement of people's living standar ds and the overall cost of the decline in the car,the car in people's lives play an increasingly important role.It is an important research direction to ensure the safe driving of the vehicle under the complicated road condition.Road information detecti on is the key content.However,there are few researches on road information detection,especially the research on active detection of road information by vehicles as nodes.In this context,this thesis studies the road information recognition algorithm based on multi-sensor,and the research content can promote the safe driving.The most important evaluation criteria in the road quality index are the degree of road bumps and skid resistance.The bumpiness of the road is the smoothness of the road,and anti-skid performance evaluation criteria are based on different types of road foundation.The research content of this subject is to identify the active detection road information with the vehicle as the main body.Therefore,this thesis selects the most important two kinds of information as the recognition target,one is the bump degree information of the road surface,the other is the road type information.In this thesis,we use the motion sensor to obtain the acceleration perpendicular to the pavement,and obtain the eigenvector of the type which can accurately describe the bumpy road by the feature extraction.Through a series of experimental verification and parameter tuning,it is found that motion sensor combined with feature extraction and Hidden Marko v Model can judge the bump type of the road better.For the determination of the road type,it is also necessary to specify the road type to the physical parameters which can accurately describe its information.At the present stage,the mature object recognition schemes are generally based on machine vision.Through a series of experiments,it is proved that the texture feature can express the road type better in the recognition of road type,and finally select GLCM(Gray Level Co-occurrence Matrix)as the texture feature algorithm.At the same time,in order to solve the problem of road type identification,this thesis puts forward the application method of Voting Support Vector Machine and proposes a complete road type recognition scheme.In order to verify the algorithm proposed in this thesis and the application method,the corresponding experimental platform is set up.The acceleration data in the vertical direction of the vehicle are collected in real time and the result is satisfactory.The average recognition precision can reach 94%.The application of simulation in road type recognition proves the corresponding algorithm and application.Experimental verification based on computer vision based texture feature matching improved SVM(Support Vector Machine)algorithm can better identify the road type and achieve better results in accuracy.The average recognition accuracy is 93.2%.In general,the algorithms and methods used in this thesis have a high practical value and practical value for the developmen t of vehicle-based sensor networks.