A Study On Autonomous Driving System For Network Connected Vehicle And Its Application In The Parking Lot Scenario

Currently with the rapid progresses of the advanced technology,especially as internet technology,communication technology,data technology,and sensing technology,the following three major changes are occurring in the automobile industry and its relative industrial fields:1)The city cars,as a kind of independent and freely running mobility,have been gradually becoming the mobile node in the traffic net,and the management of the road traffic running right and running safety responsibility gradually transfer to the online cloud by internet;2)Development of car-sharing network makes city private cars be gradually replaced by service vehicles;3)Vehicles with the devices of active safety and driving assistance are being into the market,and more and more technical parks and test-courses for the pilot operation autonomous driving are being in building in recent years.On the other hand,the current situation of urban traffic problems,like traffic jams,parking difficulties,traffic accidents,causing air pollution,and high energy consumption,sharp contradictions increasingly.In these backgrounds,the automobile manufacturing industry and travel services industry are undergoing a profound revolution and the autonomous driving technology will become a key powerful promoter.As a result,the city road traffic concept and value orientation have been transferred to be more orderly,smooth,safe,efficient,green,convenient,and economic from independent,free,and driving pleasure.The main content of this theme is studying on the network connected vehicle autonomous driving technology,that be significantly on the design and implementation of information platform,cloud vehicle control system,and related application service system.However,the most of the current researches of autonomous driving still focus on a single vehicle running independently and freely,as Google and Tesla do.And the autonomous driving vehicles are assumed full safe-responsibility.It will be impossible to solve all the problems of traffic safety in the short term and it will be difficult to use this kind of autonomous driving vehicle in practice.We believe that the autopilot of autonomous driving must be studied and applied in an intelligent traffic environment,in an orderly,smooth,efficient,safe,convenient,comfortable,green as the goal and value orientation.Our philosophy is as "subtraction in the car while addition in the cloud" by new internet and its related technologies.The main work of this theme is in the following five aspects:(1)The architecture design of autonomous driving systemBased on the background of development of the modern city road traffic information and intelligent transportation,this theme studies on several city information platform development,intelligent transportation system,and modern smart city computing.The design idea and design principle of the autonomous driving system architecture are proposed and the architecture is designed as more simplified,hierarchical,open,and extensible.(2)Traffic information service platform designThe intelligent transportation information platform system is designed to realize the core needs of application portal,to realize the extended application and to make module as alternative forms.The most important feature of the platform is be able to share the service through the platform of every application to achieve data sharing,traffic safety responsibility and data collaboration further.Based on the analysis of the basic demands of the information platform,the platform standard data format and communication protocol are defined.And the specific function and core modules are developed to be used in the information platform.(3)Vehicle cloud control systemFor the vehicle cloud control system,the functional requirements and non-functional requirements are analyzed and system functions are defined.In order to keep being modularization and replaceability,the principle of detailed design and realization of vehicle control system is made first,and then architecture is designed and the cloud system modules are developed,that include vehicle tracking algorithm and control module and vehicle front collision protection control module and its algorithm.(4)Application service system of unmanned parking lotTo meet the functional requirements of the application of parking lot,the overall parking architecture and the function of each modular are defined,and then the parking business logic follow is determined.This theme also introduces basic principle and realizing method of laser radar positioning and develop the positioning module used in the parking lot,that is equipped in the vehicle to provide real-time accurate positioning data system for vehicle.(5)Parking lot scale model and its experimentsIn order to show the research results,this paper designed and produced a complete parking lot scale sand table model.Using this model,traffic information service platform,cloud vehicle control system,application service system of unmanned parking lot,realize the coordinated operation of the system,a series of experiments are done.The experiment results show that the proposed system can provide accurate automatic tracking,and by adapting the topological map of parking lot to network connected autonomous vehicle,the parking capacity can be increased.