Study Of Active Lane Changing Strategy Based On Vehicle Motion State Prediction

As one of the common driving behaviors,lane changing is also an important part of intelligent driving research.This paper takes the lane changing as the research object,analyzes and studies the vehicle motion state prediction in lane changing scenarios,lane changing decision making,trajectory planning and trajectory tracking.In this paper,the vehicle motion state is firstly modeled based on the car-following model of the long short term memory network,based on this,a lane changing decision-making model considering speed dissatisfaction expectation is established,and a selecting function of the feasible area of lane changing is established for the vehicle.Then,the trajectory of the lane changing process is planned based on the polynomial curve,finally the planned trajectory is followed based on model predictive control.Firstly,in the aspect of vehicle motion state modeling,a vehicle motion state model based on the car following model of the long short term memory network is proposed to determine the motion state of each vehicle in the lane changing scene.The theory-driven car-following model has good interpretability,the intelligent driver model can represent the vehicle-following state under various traffic flow conditions.However,since the vehicle’s following state is affected by various factors such as the driver’s personality and external environment,it is difficult for a car following model based on a fixed expression to truly reflect various car-following situations.Since the driving process of the vehicle is continuous in time dimension,this paper models the motion state of the vehicle based on the long short term memory network model in the data-driven car-following model,which can characterize the influence of the driving situation in the past period of time on the motion state of the vehicle.Secondly,in terms of lane changing decision making,this paper establishes a lane changing decision making model based on speed dissatisfaction expectation,and judges the feasibility of lane changing through the minimum safe distance model and time headway.This paper predicts the speed of the vehicle on the current road,calculates the cumulative value of speed dissatisfaction expectation to characterize the vehicle’s lane changing willingness,and selects the target lane based on indicators such as vehicle speed and relative distance.Since only the minimum safe distance model is considered,the distance between vehicles may be too small during lane changing in some specific scenarios,the variable time headway is introduced to correct the model.The influence of lane changing behavior is reduced while ensuring the safety of lane changing.Thirdly,in terms of trajectory planning,by comparing the characteristics of different types of lane changing trajectories,this paper selects the quintic polynomial as the form of lane changing trajectories.When the vehicle’s willingness to change lane is strong but the current road environment does not meet the lane changing conditions,this paper designs a selecting function for the feasible area of lane changing,that is,planning the vehicle’s speed for a period of time in the future,controls the vehicle to perform active acceleration and deceleration operations,thereby creating a feasible lane changing space for itself.When the lane changing safety is satisfied,considering the influence of road curvature,this paper establishes corresponding trajectory planning models for straight and curved roads respectively.An objective function that considers the comfort and efficiency of lane changing is established,and it is solved to complete the trajectory optimization.Finally,a lane changing trajectory tracking model is established based on model predictive control.Firstly,a three-degree-of-freedom vehicle model is established and linearized.By establishing the relevant motion constraints of the vehicle,taking into account the objective function of the following effect and the change of the control amount,the optimal control amount is finally obtained by the method of rolling optimization.On the condition that the output control amount is reasonable,the effect of trajectory following is effectively improved.This paper establishes different test scenarios in Prescan,and the control strategies proposed in this paper are simulated and verified through the form of Prescan/Carsim/Simulink co-simulation.