自动驾驶论文2364 Probabilistic Safety Assured Adaptive Merging Control for Autonomous Vehicles.pdfVIP

2021IEEEInternationalConferenceonRoboticsandAutomation(ICRA2021)

May31-June4,2021,Xian,China

ProbabilisticSafety-AssuredAdaptiveMergingControl

forAutonomousVehicles

YiweiLyu,WenhaoLuoandJohnM.Dolan

Abstract—Autonomousvehiclesfacetremendouschallenges

whileinteractingwithhumandriversindifferentkindsof

scenarios.Developingcontrolmethodswithsafetyguarantees

whileperforminginteractionswithuncertaintyisanongoing

researchgoal.Inthispaper,wepresentareal-timesafecontrol

frameworkusingbi-leveloptimizationwithControlBarrierFig.1:Rampmergingscenario(Dongetal.[15]).Theegovehicle

Function(CBF)thatenablesanautonomousegovehicleto(green)isanautonomousvehicle,runningonthemainroad;the

interactwithhuman-drivencarsinrampmergingscenariosmergingvehicle(red)isahumandrivencar,runningontheramp.

withaconsistentsafetyguarantee.InordertoexplicitlyTheegovehicleandthemergingvehicleareinterchangeable.

addressmotionuncertainty,weproposeanovelextensionof

controlbarrierfunctionstoaprobabilisticsettingwithprovable

chance-constrainedsafetyandanalyzethefeasibilityofour

controldesign.Theformulatedbi-leveloptimizationframeworkhighwayrampmergingscenario[4].Inaddition,vehicle

entailsfirstchoosingtheegovehicle’soptimaldrivingstyleincontrolwithACCintherealworldcanbedifficult,since

termsofsafetyandprimaryobjective,andthenminimallysometime