Research On Automatic Lane Change For Obstacle Avoidance Of Intelligent Vehicle Based On Nonlinear Model Predictive Control

As one of the key technologies of intelligent assistant driving,automatic lane change for obstacle avoidance control is very important for the safe driving of driverless vehicles.This paper discusses the control of "single vehicle lane change" and "multi vehicle lane change".In the specific thought,firstly,combine with the vehicle kinematics and dynamics principle,a three degree of freedom nonlinear dynamics model under the intelligent vehicle geodetic coordinate system is established,and the prediction model for controller design is obtained through discretization;then,the control system design of single vehicle obstacle avoidance and multi vehicle collaborative obstacle avoidance is carried out according to different obstacle avoidance scenarios and lane change requirements.Aiming at the single vehicle automatic lane change for obstacle avoidance control,this paper takes the steering wheel movement stability as the optimization objective,the vehicle lateral movement at the end of lane change as zero,and the lateral displacement reaches the desired lane change distance as the terminal constraint,and designs a lane change for obstacle avoidance controller based on model predictive control.Under the condition of left lane change and overtaking,the effectiveness of the proposed controller is verified by the joint simulation of Simulink and Car Sim.Aiming at the problem of multi vehicle cross lane change for obstacle avoidance,considering the collision probability of different positions of multi vehicle in the process of lane change,the collision threat model based on the vertical and horizontal distance is derived.The objective of the optimization is to reduce the tracking error and the collision probability,and ensure the stability of lane change.Under the lane change condition of multi lane and multi scene,the collision threat model of multi vehicle cross lane change for obstacle avoidance system is proposed.And the effectiveness of the proposed controller is verified.The two main research contents of this paper are interrelated and complementary,which reflect the essence of the problem from different aspects,and provide some solutions for active lane change control.However,some researches are not perfect.In the future,further efforts will be made in algorithm solving speed and real vehicle test.