Wuhan University of Technology Undergraduate Graduation Design (Thesis)

Research on local path planning of intelligent vehicle considering speed planning

School: School of International Education

Specialty amp; Class: Automotive Engineering gj1603

Name: Xiaotian Zhou

Tutor: Houyu Yu

Statement of Degree Thesis originality

I solemnly declare：The thesis submitted is the research result independently obtained by myself under the guidance of my tutor. Except for the contents specifically quoted in this thesis, this thesis does not include any other works that have been published or written by individuals or groups. I fully understand that the legal consequences of this statement are borne by me.

Author’s Signature：Xiaotian Zhou

Date:2020.5.31

Using Authorization of Thesis copyright

The author of this thesis fully understands the provisions of the school regarding the protection and use of thesis, and agrees that the school retains and sends copies and electronic versions of the thesis to the relevant thesis management departments or institutions, allowing the thesis to be viewed and borrowed. I authorize the provincial excellent bachelor thesis selection institution to compile all or part of the thesis into relevant data for retrieval. The thesis can be saved and compiled by photocopying, printing or scanning.

The thesis belong to 1、Confidentiality，Apply this authorization after the year of declassification)

2、No Confidentiality囗 .

(Please type “√” in above relevant block)

Author’s Signature：Xiaotian Zhou Date: 2020.5.31

Tutor’s Signature： Date:

Abstract

Under the upsurge of artificial intelligence, the development of the smart car industry has also received widespread attention from society. Because it has made important contributions in reducing the incidence of traffic accidents, reducing traffic congestion, improving the atmospheric environment, and reducing energy consumption. Driverless vehicles is an advanced stage of the development of smart cars. As one of the core technologies of driverless vehicles,motion planning is an indispensable transfer of research on driverless technologies. It is of great significance to the research of driverless vehicles. Its basic task is Considering the geometric relationship between the vehicle and the obstacles, a displacement trajectory to avoid collisions is planned, and at the same time, it should have higher riding comfort, thereby improving travel efficiency.

Based on the research of motion planning algorithms at home and abroad, this paper proposes a decoupled decision, that is, to transform the trajectory planning problem in three-dimensional space into the path planning and velocity planning problem in two two-dimensional space. After studying the five major algorithms of current motion planning, a piecewise jerk path optimization algorithm based on quadratic programming is proposed based on the discrete optimization algorithm. The main idea of the path planning part is to discretize the road guidance line based on the Frenet coordinate system, and then set the value range of the lateral displacement of each discrete point according to the environmental information. Finally, the objective function is constructed by taking the lateral displacement and the velocity, acceleration and jerk of the lateral displacement as independent variables. In this way, the problem is transformed into a convex quadratic programming problem, which makes the algorithm very efficient. peed planning is based on the path - time obstacle diagram planning speed curve, the basic idea is similar to path planning. At the same time, this paper compares and analyzes this algorithm with A* and RRT* algorithm, and finds that the piecewise jerk path optimization algorithm has the best overall performance.In order to verify the practicability of the planning algorithm, this paper designs a single-lane obstacle avoidance scenario, a multi-lane lane change overtaking scenario and a car-following scenario based on the automatic driving toolbox. It also proposes to avoid pedestrians such as low-speed moving obstacles in the path planning module, which can make the algorithm more practical. At the same time, it constructs a decision function to judge whether the vehicle is overtaking or following, and a double-line shifting model. Finally, the actual simulation is carried out on the simulation platform, and the correctness of the programming algorithm is proved.

The results show that the motion trajectory generated by the piecewise jerk path optimization algorithm proposed in this paper is smooth, safe, and highly executable. At the same time, the solution efficiency of the algorithm is higher than that of the traditional A * algorithm and RRT * algorithm, and it has strong practicability. In the future, if the performance of sensor module and V2X technology is improved to a certain extent, this algorithm may be widely used.

Key Words：motion planning；path planning；speed planning；Piecewise jerk path optimization algorithm

Directory

Abstract I

Chapter 1 Introduction 1

1.1 Background and significance of the research 1

1.2 Research status of motion planning algorithm at home and abroad 3

1.3 Research content and technical route 6

Chapter 2 Motion planning algorithm design 8