1. 研究目的与意义（文献综述）
underthe upsurge of artificial intelligence, the development of intelligentautomobile industry has also been widely concerned by the society, because itwill make important contributions to reducing the traffic accident rate,reducing traffic congestion, improving the atmospheric environment and reducingenergy consumption[1]. unmanned driving is an advanced stage ofintelligent vehicle development. as one of the core technologies of unmanneddriving, local path planning is an indispensable link in the research ofunmanned driving. local path planning, also known as obstacle avoidance pathplanning, is to consider the geometric relationship between the vehicle and theobstacle to find a path to avoid collision with the obstacle, at the same timeto ensure the flexibility of the vehicle, the safety of passengers, ridecomfort, algorithm reliability, etc.
at present, themore mature local path planning algorithms include artificial potential fieldmethod, graph search method, sampling method, discrete optimization method andoptimization algorithm[2-3]. however, these algorithms still havetheir own limitations. for example, the artificial potential field method is avirtual force method proposed by khatib. this method is simple and has goodreal-time performance[4-5]. a*and d*algorithms are commonly used ingraph search methods. although these algorithms are widely used in the field ofrobotics, the planned path fails to meet the non-integrity constraints ofvehicles, and the path smoothness is poor[6]. sampling based methodis rapidly random tree (rrt). although rrt algorithm has probabilisticcompleteness, the random results of each search are relatively large. secondly,a large number of collision detection in the process of uniform random samplingby standard rrt algorithm seriously affects the efficiency. in addition, themeasured distance in vehicle configuration space has no closed form expression,so it is difficult to calculate[7]. the path planning method basedon discrete optimization is to describe the vehicle's motion by numericalintegral and differential equations, so as to produce a limited number ofcandidate paths, and select the optimal path from the candidate paths bydesigning the cost function [8]. optimization algorithms aregenerally based on biological intelligence or physical phenomena of the randomsearch algorithm, the common fuzzy logic algorithm, neural network algorithm,genetic algorithm, ant colony algorithm. however, there are many problems, suchas large amount of computation, difficulty in optimizing the high dimension andlong learning time, which still need to be improved. due to the traditionalsingle path planning algorithm using the unmanned vehicle to a lot oflimitation, today the mainstream of the study is combining two or morealgorithms, to improve the efficiency of algorithm by complementary[9].such as the literature[10] presents a based on improved fireworks -mixed ant colony algorithm of intelligent mobile obstacle avoidance pathplanning, put forward to increase spark "pioneer" and adopt"specular mapping rules in dealing with cross-border" pioneer sparkimprovement method; then, the shortest path obtained by the improvedfireworks algorithm is taken as the reference path and converted into theinitial pheromone distribution of ant colony algorithm, so as to solve theshortcomings of the ant colony algorithm such as slow convergence speed andinsufficient initial pheromone. in the literature[11],fuzzy logic and genetic algorithm are used to construct an intelligent vehicleobstacle avoidance path planning method. which uses genetic algorithm tooptimize the target fuzzy rule table, improving the computational efficiency ofthe algorithm. in order to avoid the local minimum problem of classicalartificial potential field method (apf), a new method based on modified apfalgorithm and fuzzy logic is proposed[12]. the algorithm overcomesthe local minimum problem and improves its effectiveness in complexenvironments. particle swarm optimization (pso) algorithm is used to optimizethe membership function of fuzzy logic algorithm. in the static and dynamicenvironment response speed, and can effectively avoid obstacles. documents willbe virtual force method combined with path planning based on rolling windows,can solve well to avoid static obstacles and low-speed dynamic obstacleproblem, control the vehicle to target location, and meet the vehicle dynamicsconstraint in the process of obstacle avoidance, without collision and roadboundary, ensure the safety and stability of the vehicle[13].
speed planning, as an important part ofintelligent vehicle movement planning, is a key factor affecting thelongitudinal driving safety and ride comfort of intelligent vehicle. speedplanning is to calculate the speed distribution on the future path according tothe planned path, environmental information and vehicle state, so as to realizesafe and smooth driving. the velocity planning methods can be divided intocoupling velocity planning and decoupling velocity planning[14]. theformer exists in a motion planning framework that uses optimization algorithms[15-16] or search algorithms [17] to simultaneously explore spatiotemporalspace. time parameterized trajectory planning based on optimal control mostlybelongs to this category. because it is solved in non-convex domain, thecomputational efficiency of this method is low. the latter is often found in theframework of hierarchical motion planning[18], which decouples themotion by planning a path and then reconstructing a velocity curve along thepath, which greatly improves the efficiency of calculation and is themainstream direction of velocity planning research.
2. 研究的基本内容与方案
2.1basic contents of the research (design) :(1)analyze the domestic and foreign research status of local path planning forintelligent vehicles;
(2)compare the advantages and disadvantages and application scenarios of theexisting local path planning methods for smart cars;
(3)design local path planning method for intelligent vehicles;
3. 研究计划与安排
| Weekly(time) | Jobcontent | Submission(end of phase) |
| 1-2 (weeks 18and 19 of semester 7) | Todetermine the graduation project topic, to improve the graduation projecttask book (related parameters), and to collect internal and external data | Graduationproject assignment |
| 3-4 (weeks 1-2of semester 8) | Plandesign, document search, complete proposal report | Literaturesearch, opening report |
| 5 (week 3 ofsemester 8) | Foreigntranslation, data collection | Foreignlanguage translation |
| 6 (week 4 of 8,March 19) | Openinga plea | PPTof proposal defense, record sheet of proposal defense |
| 7-8 (week 5-6of semester 8) | Designcalculation, sketch drawing | Designcalculation draft, sketch |
| 9-11 (week 7-9of 8th semester) | Drawingand compiling design and calculation specifications (thesis) | Design,paper draft |
| 12 ~ 15 (week10-13 of 8th semester) | Designand paper arrangement; Intermediate inspection | Formaldrawings, papers |
| Two weeks ofgraduation practice (weeks 11-12 of 8th semester) | Off-campusor online internship, data collection, completion of internship report | Theinternship report |
| 16 (week 14 ofsemester 8) | Studentsapply for defense, prepare for defense, and organize the materials.Examination of the defense qualification and reexamination; The teacherreviews drawings and specifications | Graduationproject information kit |
| 17 (15 weeksof semester 8) | Attendrejoin | ReplyPPT Graduationproject |
4. 参考文献（12篇以上）
[1].陈缘.谈“互联网 ”时代下智能汽车的发展[j].农机使用与维修,2019(01):19.
[2].david gonzález,joshué pérez,vicente milanés,fawzi nashashibi.a review of motion planning techniques forautomated vehicles[j]，ieee transactions onintelligent transportation systems，2016，17(4): 1135 –1145.
[3].彭晓燕,谢浩,黄晶.无人驾驶汽车局部路径规划算法研究[j].汽车工程,2020,42(01):1-10.
课题毕业论文、开题报告、任务书、外文翻译、程序设计、图纸设计等资料可联系客服协助查找。
