Asphalt Pavement Texture Recognition And Braking Method For Autonomous Vehicle

The autonomous vehicle makes the driving behavior appreciable.This technology improves traffic safety because of their ease of traffic pressure,reduced energy consumption,and reduced traffic accidents.This will be the mainstream form of vehicles on the road in the future.However,the brake model of the autonomous vehicle ignores the influencing factors of the asphalt pavement surface,and does not consider the impact of road surface texture on the safety of vehicle braking under different road conditions on sunny days and rainy days.The texture characteristics of the asphalt pavement surface are directly related to the anti-skid performance of the road.Good anti-skid performance of the road can provide sufficient friction for the high-speed driving vehicle to ensure safety and comfort during the driving of the vehicle.In addition,typical road surface texture measurement methods and characterization parameters are the equalization of the road surface texture level,which can not reflect the fractal characteristics of the rough asphalt pavement surface texture.The efficiency of traditional measurement methods is low and there is an uncontrollable loss of texture information.In view of this,this paper combines the technical characteristics and sensing requirements of autonomous vehicles to study the surface texture recognition method of asphalt pavement and the braking stability of autonomous vehicles in the whole life cycle of the asphalt.This paper also aims to provide theoretical guidance for braking strategy selection of autonomous vehicles.Based on the basic principle of close-range photogrammetry technology,this paper first proposed an automatic close range photogrammetry system(ACRP system),and at the same time built a close-range photogrammetry hardware platform to realize automated real-time acquisition asphalt pavement surface texture images.Then,a supporting ACRP software module was established using MATLAB and Python hybrid programming to realize the automatic preprocessing of the asphalt pavement image.3D models of the asphalt pavement reflecting the actual texture details of the pavement surface were reconstructed from 2D images based on the VisualSFM open source framework.Due to defects such as burrs and holes in the initial reconstruction of the pavement model,further pretreatment is needed to improve the accuracy of 3D reconstruction of the pavement.The refinement method of 3D model of asphalt pavement surface texture is described.Based on the obtained 3D model of asphalt pavement,the ACRP software module for visualization of 3D coordinate point data and texture parameter calculation of asphalt pavement surface texture is developed.The subroutine for calculating the mean texture depth(MTD)of the pavement surface was realized.Processing software including GeoMagic and MeshLab were used to preprocess the 3D models of the inversely reconstructed asphalt pavement surface texture to obtain accurate 3D coordinate point data.Then,the root-mean-square roughness(RMSR)and the MTD value were selected as the evaluation indexes.The validity and accuracy of the established texture parameter extraction method were verified by the on-site sand patch method and laser scanner method and provided the basis for friction coefficient calculation.Considering the wet conditions,the road surface power spectral density changes due to the barrier of the water film.Thus,the texture characteristics of the road surface under wet road conditions need to be considered.Firstly,the formation mechanism of the water film on the surface of rainy asphalt pavement was introduced.The formation process of the water film on rainy asphalt pavement and the influencing factors of water film thickness were summarized.The hardware equipment was used to realize the realtime acquisition of water film thickness of asphalt pavement.Secondly,based on the Persson friction theory model of the fractal theory,the road power spectrum density curve was calculated by Python programming.The traditional power spectrum calculation method is not applicable to the wet road surface condition.As a result,the concept of “anti-skid non-contribution area” was proposed.The dynamic friction coefficient curves of the tire-road surface under dry and wet conditions were calculated.The accuracy of the theoretical solution of the friction coefficient was verified by DFT tester.The differences in braking principle between autonomous vehicles and traditional vehicles were compared.Various braking scenarios faced by autonomous vehicles were summarized.Based on the friction coefficient curve obtained above,a braking model of the autonomous vehicle was created in Simulink based on the braking principle of the autonomous vehicle.Co-simulation of the braking of autonomous vehicles in different scenarios on dry and wet roads(emergency braking and blocked road braking)was carried out in CarSim/Simulink.Braking strategies of autonomous vehicles under different working conditions were specified.For emergency braking conditions,it is important to consider the safety of the brakes.A braking strategy of “gentle braking” was proposed for the braking situation of congested road sections,which fully considered passenger comfort.The results show that the autonomous vehicle should be equipped with professional short-range radar,long-range radar and high-definition camera to detect the environment on the vehicle's traveling route.The effective distance of the longrange radar should exceed 150m;the visual field on the rainy day would be reduced.The emergency braking distance is more than 45% compared with that in the case of sunny weather.The speed should be slowed down and the safe distance should be kept.When the vehicle distance is large,a middle-sized braking force is preferred(4 MPa~6 MPa);autonomous vehicles should keep a certain safe distance(1.1~1.2 times the average value of safe braking distance for sunny and rainy days)during the driving process on dry or wet roads.Therefore,in order to ensure the braking safety of autonomous vehicles under different road conditions,the influence of road surface texture characteristics needs to be considered in the autonomous braking model.