Estimation Road Grade And Velocity For Autonomous Driving Vehicle

Because of the development on a variety of control, communications and informationfusion technology, the autonomous driving vehicles benefit from these developments. Butthe implementation of these techniques requires the accurate vehicle states information.For now, the information of vehicle states is measured directly through the vehicle sen-sors. But with the improvement of the performance of the vehicle and the increase invehicle control technology, the vehicle states information will be more required, especialautonomous driving vehicle. Then, It is impossible that all the states information is mea-sured by sensors. The one reason is easy to produce information redundancy, informationconfusion, information loss because of excessive vehicle sensors. On the other hand, thecost of production will be greatly increased due to the installation of a large number ofvehicle sensors. It limits the enhancement of vehicle safety performance because of theexistence of these problems.Autonomous driving vehicle control technologies require a large of vehicle states.While It’s the foundation of autonomous driving vehicle control that known accuratelyvehicle longitudinal velocity, lateral velocity and road grade, etc. But these states can notbe measured directly using a sensor. And though they can be measured, the instrumentsare much expensive. In response to these problems, combined with the R&D needs ofhighway vehicle intelligent driving key issues research major project which is supported bythe National Natural Science Foundation and is researched by our laboratory as well as.This paper proposes that the unknown and required states will be estimated by estimatorwith the existing vehicle states information. The vehicle system is a very strong nonlinearsystem and each state can’t exist independently due to the interaction between them witha strong coupling. Therefore, this paper presents a nonlinear full-order observer whichis used to estimate the longitudinal velocity, lateral velocity and road grade of HongqiHQ430vehicle with autonomous driving.In order to design the nonlinear full-order observer for the vehicle states, this paperdoes some work. First, establish a simplified vehicle dynamic model for Hongqi HQ430vehicle. The model can characterize the performance of autonomous driving vehicle on highway. Design a nonlinear full-order observer which is used to estimate the longitudinalvelocity, lateral velocity and road grade of Hongqi HQ430vehicle with autonomous drivingbased on this simplified model. Second, in order to verify the estimate efect of thenonlinear full-order observer, we do some simulation experiments for the autonomousdriving vehicle. This paper builds a simulation dynamic model of the HongqiHQ430withautonomous driving based on dynamic software veDYNA taking into account the securityand the economy. Thus, the validity of estimator will be verified through simulationwith the established simulation dynamic model instead of real vehicle. Then, verify theperformance through the data of the real vehicle tests which implemented accordancewith national laws and regulations. At last, validate the estimator’s validity. Beforesimulation, this paper give the vehicle’s simulation parameters and the selection progressof the gains about designing observer. Do simulate and research for the estimate of thevehicle states through diferent simulation conditions. And Analyze the estimation resultsfor nonlinear full-order observer of the vehicle states based this. The simulation resultsshow that the estimation method have certain validity and accuracy.