Development Of Emergency Steering Assistant Obstacle Avoidance System Under Urban Conditions

With the increasing emphasis on traffic safety,more and more active safety systems are installed on vehicle platforms to reduce accidents caused by human negligence or improper operation.At present,the research about forward anti-collision system mostly adopts braking intervention method to avoid collision,however,the method can not avoid the collision completely because of the long braking distance in some special scenarios such as high relative speed and poor road adhesion conditions.For the reason this paper designs an assistant system which adopts steering intervention methed to control vehicle to achieve lateral motion obstacle avoidance,as an supplement to active anti-collision technology.Emergency steering assistant obstacle avoidance system mainly involves three key technologies: environment perception,planning and vehicle lateral motion control.This paper focuses on the latter two aspects.Firstly,thorough the investigation of application scenarios,sensor assembly schemes and control algorithms for steering obstacle avoidance system the paper puts forward the technical route of the research and designs the overall structure of the system.Secondly,through the screening of natural datacase of avoidance to determine the typical application scenario system,and five-degree polynomial algorithm is used to complete obstacle avoidance path planning.Then a multi-objective constrained safety distance model including lateral stability constraints,road boundary constraints and safety obstacle avoidance constraints is established.At last,the system decision-making strategy is put forward according to the judgment conditions of safety time,collision risk coefficient and driver's steering behavior.Thirdly,the D2 F vehicle dynamic model and Pacejka tire model are established to describe the vehicle's motion during steering,and then a sample set of vehicle's motion is generated,which meets the condition of emergency collision avoidance.Vehicle inverse dynamics model is trained by radial basis function(RBF)neural network algorithm.The results show that the identification accuracy of inverse dynamics model is high.What's more,a fuzzy adaptive PID controller is designed to track vehicle lateral acceleration accurately by feedforward and feedback control method.Finally,the simulation test scenario is established by referring to C-NCAP(2018)VRU_Ped test standard,and the joint simulation test is completed by Carsim and Simulink.The results show that the emergency steering assistant obstacle avoidance system can avoid obstacles completely and keep a certain safe distance,and the lateral motion controller can accurately control the vehicle to track the planning path.What's more,based on a SUV vehicle platform,the sensors,controllers and actuators are integrated into a driving assistance system to complete the real vehicle test in typical test scenarios.The real vehicle test results show that the system can identify the target in front of the vehicle in time,then activate the system according to the decision strategy and control the vehicle to complete the avoidance task through the planned path,which verifies the effectiveness of the emergency steering assistant obstacle avoidance system.