Road Curvature Estimation Method For High Precision Navigation Map Data

In the information age,driverless technology and driver assistance technology are increasingly attracting the attention of people in the industry.In order to better solve the driver's navigation problems during driving and give better and more comfortable driving feelings,the driver is The hands are liberated and the driverless car can safely drive on the road,all relying on the participation of navigational electronic maps with higher precision,especially the establishment of a lane-level navigation map database plays an absolutely important role.The pre-data to be obtained from the lane level feature information includes the shape of the road,the curvature of the lane,slope,course,elevation,and roll.According to the road safety concept,for the safety of the vehicle,the linear design of the road is almost always curved,and the curvature of each curve is different,especially in the expression of intelligent driving high-precision navigation electronic maps.Vehicles are reminded in real time.Therefore,the accuracy of road curvature plays an important role in the road attribute data,and it is an important issue related to whether the autonomous vehicle can run safely on the road.The high precision of high-precision navigation electronic map means on the one hand that the absolute coordinates of high-precision navigation maps are more accurate than ordinary electronic maps.That is,the coordinate accuracy on maps refers to the accuracy between the objects on the map and the real world outside;On the one hand,the attribute information of the high precision navigation map contains more,more abundant and detailed road traffic information elements.In this paper,based on the research of the development status of high-precision navigation maps and the method of road curvature estimation,and according to the characteristics of the experimental data used in this paper,a curvature estimation algorithm of“minimum fitting circle”for calculating road curvature based on GNSS data is proposed.First,preprocessing the original data curve fitting for each section.And a polyline is generated by equidistant interpolation.Then,the generated polyline is projected onto the unit ball by the stereo projection method.According to the rule of least squares,find the best fitting plane of the data point on the unit ball,and the intersecting circle of the fitting plane and the unit ball.When the intersecting circle is stereoscopically projected to the equatorial plane,corresponding arcs and curves can be obtained.The curves are categorized into five types:no curve,simple curve,compound curve,inverse curve and spiral curve.The results of the recognition show that the proposed algorithm is accurate and the road edge is extracted.This article is mainly divided into two steps in the verification of road curvature accuracy.The first step is to build the curve library.After processing the road edge data collected by the GNSS-RTK,use the identified five curves to set up the curve database to prepare for the later accuracy verification.The second step is to verify the accuracy of the curve,and to import the curve node coordinate data of the point cloud data into the system.The data detection and curve curvature value are automatically compared and verified,and the results are finally obtained,and the subsequent data analysis is directly derived.Finally,the verification result is obtained,and the subsequent data is directly exported and saved.The main examination data of this paper is the curvature estimation of the road border.The accuracy verification and data analysis of the two kinds of data show that the maximum difference between the curvature value of the same line segment is8.99×10^-3,the minimum value is5.7×10^-6,the difference mean square root is3.38×10^-3,the relative precision of the average value difference is 1/442,and the precision reaches the expected assumption.By comparing the curvature of the road edges extracted from the point cloud data,the results show that the overall completion of the method is better,and the expected goal of the result precision is ensured while improving the efficiency of the operation.