Model Predictive Control Based Motion Planning Algorithm Of Intelligent Vehicles

With the development of intelligent technology,intelligent driving technology has gradually become one of the research hotspots.Motion planning algorithm is a key technology to ensure the safe driving of intelligent vehicles,and its quality directly affects the safety,efficiency,and ride comfort of vehicles.Most of the existing motion planning algorithms cannot effectively consider the complex constraints in the driving process.Therefore,the planned trajectory is quite different from the vehicle’s actual driving trajectory,and there are higher requirements for the trajectory tracking algorithm.Motion planning algorithms based on numerical optimization can handle mul-tiple constraints,but complex obstacle constraints will bring a computational burden.Therefore,because of the above problems,this paper designs a motion planning algorithm based on Model Predictive Control(MPC),and the constraint simplification algorithm is designed to simplify the constraints of complex obstacles linearly.Simultaneously,design a rule-based behavioral deci-sion algorithm to improve the algorithm’s ability to handle dynamic environments in structured and unstructured road scenarios.This article first proposes a model predictive motion planning algorithm based on linear con-straints.This method considers the vehicle’s geometric constraints by modeling the vehicle as three intersecting circles to ensure the collision-free safety of the vehicle in motion planning.For complex non-convex obstacle constraints,this paper proposes a simplified algorithm to generate simplified linear obstacle constraints through the geometric relationship between the vehicle ge-ometric representation circle and the obstacle.In the case of multiple obstacles,the constraint generator can generate two linear constraints at most,effectively reducing the number of con-straints and improving calculation efficiency.In the design of the MPC based motion planner,this algorithm adds the kinematics and dynamics constraints of the vehicle,and at the same time,adds simplified obstacle constraints through the predictive model to realize the dynamic generation of the moving obstacle constraints and ensure the safety of the motion planning algorithm.The algo-rithm treats all constraints as linear inequalities and the optimization problem of motion planning as quadratic programming to ensure the algorithm’s real-time performance.Through the simula-tion test in the MATLAB environment and compared with the existing algorithms,the proposed motion planning algorithm’s validity and real-time performance are verified.In order to enable the algorithm to be applied to different road conditions,including structured road scenes and unstructured road scenes.This paper designs a rule-based behavioral decision-making method.The designed decision-making algorithm processes driving behaviors into five types of typical behaviors,distinguish structured and unstructured road scenes according to the number of reference trajectories,and analyzes and processes their state transitions respectively so that the algorithm can adapt to different traffic environments.Different reference trajectories are generated to guide the motion planning algorithm to complete the motion planning according to different driving behaviors.Simultaneously,the decision-making algorithm also considers the scenes that need to stop so that the algorithm planning trajectory is more in line with human driving habits and can be applied to complex traffic scenes that need to avoid moving obstacles.Finally,simulations are performed in three different scenarios.The comparison between the same scenario and the algorithm without decision making and the light decision method verifies the proposed decision-making algorithm’s effectiveness and the improvement of the motion planning effect.