Research On Intelligent Algorithm And Security In Driverless System

In recent years,the research on unmanned driving in various countries has been extremely hot.The research on intelligent driving at home and abroad mainly focuses on the tuning and algorithm tuning between the modules of the intelligent driving system and the final real vehicle landing.The research and testing of intelligent algorithms and their safety in the unmanned driving system have not been widely studied.Concern,so it is of great significance to the study of intelligent algorithms and their security in unmanned systems.This paper first proposes the unmanned intelligent algorithm and its security architecture,and summarizes the security problems existing in each layer of the driverless system.Aiming at the problem that the local path planning algorithm based on artificial potential field method in the unmanned safety architecture planning and decision-making layer is easy to fall into the local minimum point,a deadlock-free part based on artificial potential field method and differential evolution method is proposed.Path planning algorithm;for the lane-line detection algorithm based on deep learning convolutional neural network,the lane line is missing,and the lane recognition rate is low under bad weather conditions.This paper proposes a lane line based on high-precision map and multi-sensor fusion.Detection algorithm.The main research work of this paper consists of three parts,as described below:(1)An intelligent algorithm for unmanned driving and its security architecture are proposed.According to the functions,the architecture can be divided into application layer security,network layer security,planning and decision-making layer security,and sensing layer security.This paper proposes that intelligent driverless system application layer security issues include high-precision maps that cannot match terrain under low positioning accuracy;Network layer security issues include digital signature forgery in vehicle networks and data authentication failures;planning and decision-making layer security issues include local path planning algorithms,pedestrian detection algorithms,lane detection algorithms,and obstacle detection algorithms that fail in specific scenarios.Problem;Perceptual layer security issues include security issues with GPS sensors,millimeter-wave radars,and camera sensors.(2)A local path planning algorithm based on artificial potential field method and difference method is proposed.The core idea of the algorithm is to use the differential evolution method to complete the optimal sub-object selection when the smart car falls into the local minimum point.After the selection of the optimal sub-object is completed,the local path planning algorithm can reconstruct the force field to solve the local minimum.Value problem.(3)A lane detection algorithm based on high-precision map and multi-sensor fusion is proposed.The basic principle of this algorithm is to use intelligent car's centimeter-level high-precision positioning combined with high-precision map data to complete Lane detection.In the process of high-precision map generation or high-precision map uncovered area,the use of lidar to complete the estimation of road curvature and assist Lane detection.