【摘要】：隨著社會(huì)不斷進(jìn)步,經(jīng)濟(jì)快速發(fā)展,科學(xué)技術(shù)也在突飛猛進(jìn),交通行業(yè)是典型的領(lǐng)域之一�，F(xiàn)階段的交通發(fā)展以實(shí)現(xiàn)智能交通系統(tǒng)為目標(biāo),正逐漸從信息化步入智能化,朝著智慧化邁進(jìn)。近年來,一系列先進(jìn)的理念和技術(shù)用來解決交通擁堵、道路安全、能源消耗和環(huán)境污染等問題,自動(dòng)駕駛汽車就是發(fā)展最迅速的代表。現(xiàn)階段針對(duì)無人駕駛汽車的研究多集中在車輛制造技術(shù),而對(duì)其在現(xiàn)實(shí)場(chǎng)景中的仿真研究,尤其是對(duì)高速公路下匝道過程的研究甚少。為此,本文依托國(guó)家自然科學(xué)基金項(xiàng)目“混入自動(dòng)駕駛汽車的高速公路交通流微觀建模與仿真方法研究”,以手動(dòng)駕駛汽車下匝道行為研究和已有自動(dòng)駕駛汽車實(shí)地試驗(yàn)研究為基礎(chǔ),建立自動(dòng)駕駛汽車下匝道的交通流模型；提出自動(dòng)駕駛汽車下匝道路徑優(yōu)化控制策略,運(yùn)用計(jì)算機(jī)仿真手段,確定其最優(yōu)的下匝道路徑控制策略,為面向自動(dòng)駕駛汽車的高速公路管理與控制提供前瞻性研究基礎(chǔ)。首先,在把握自動(dòng)駕駛汽車研發(fā)進(jìn)展和實(shí)地試驗(yàn)的基礎(chǔ)之上,對(duì)自動(dòng)駕駛汽車跟車模型以及常規(guī)換道模型進(jìn)行綜述,并且從宏觀和微觀兩個(gè)層面分析了出口匝道交通流特性,同時(shí)根據(jù)自動(dòng)駕駛汽車的環(huán)境感知系統(tǒng)確定其檢測(cè)半徑。其次,建立自動(dòng)駕駛汽車元胞自動(dòng)機(jī)換道模型。本文選取Jiang提出的自適應(yīng)巡航控制模型作為自動(dòng)駕駛汽車的跟車規(guī)則；在換道規(guī)則方面,分自由換道和強(qiáng)制換道。對(duì)于非下匝道車輛,引入冒險(xiǎn)因子,建立自動(dòng)駕駛汽車的自由換道模型；面向下匝道車輛的強(qiáng)制換道方面,將到達(dá)出口匝道前的強(qiáng)制換道全過程細(xì)分為自動(dòng)采集數(shù)據(jù)、確定理論安全間距、實(shí)測(cè)間距優(yōu)劣排序、實(shí)測(cè)間距安全分類和換道選擇與實(shí)行五個(gè)步驟,進(jìn)而完成強(qiáng)制換道模型的構(gòu)建。再次,以分級(jí)分段控制思想為指導(dǎo),提出自動(dòng)駕駛汽車下匝道路徑控制策略,分為階梯策略、全域策略和混合策略,建立靈活可變的仿真場(chǎng)景；同時(shí)確定策略評(píng)價(jià)的宏觀指標(biāo)和微觀指標(biāo),形成最終的綜合成本評(píng)價(jià)模型。其中該模型參數(shù)的標(biāo)定采用專家打分法和層次分析法相結(jié)合。最后,開展對(duì)自動(dòng)駕駛汽車下匝道路徑優(yōu)化控制策略的仿真工作,并對(duì)其評(píng)價(jià)分析,主要包括交通流基本參數(shù),平均行程時(shí)間以及包括換道時(shí)間、間距、換道率等在內(nèi)的換道行為特性分析。再者對(duì)比換道過程的七項(xiàng)成本,分析連續(xù)策略的綜合成本,最終確定自動(dòng)駕駛汽車下匝道的最優(yōu)路徑控制策略。
[Abstract]:With the progress of society and the rapid development of economy, science and technology are advancing by leaps and bounds. Transportation industry is one of the typical fields. At the present stage, the traffic development aims at realizing intelligent transportation system, and is gradually moving from informationization to intelligence. In recent years, a series of advanced concepts and technologies have been used to solve the problems of traffic congestion, road safety, energy consumption and environmental pollution. At present, the research on driverless vehicles is mainly focused on the vehicle manufacturing technology, but the simulation research on it in the real scene, especially on the off-ramp process of freeway is very little. Therefore, based on the project of National Natural Science Foundation of China, "study on microscopic modeling and simulation method of freeway traffic flow mixed with self-driving vehicle", Based on the research on the off-ramp behavior of manual driving vehicle and the field test of the existing self-driving vehicle, the traffic flow model of the downramp of the self-driving vehicle is established, and the optimal control strategy for the off-ramp path of the self-driving vehicle is proposed. By means of computer simulation, the optimal off-ramp path control strategy is determined, which provides a prospective research basis for the expressway management and control for self-driving vehicles. First of all, on the basis of grasping the research and development progress and field test of self-driving vehicle, the paper summarizes the model of self-driving vehicle and the conventional change model, and analyzes the traffic flow characteristics of exit ramp from the macro and micro levels. At the same time, the detection radius is determined according to the environment sensing system of the autonomous vehicle. Secondly, the model of automatic driving car cellular automata is established. In this paper, the adaptive cruise control model proposed by Jiang is chosen as the car-following rule for self-driving vehicles, and in the changing rules, it is divided into two parts: free change of track and forced change of course. For the off-ramp vehicle, risk factor is introduced to establish the free change model of the self-driving vehicle, and the whole process of forced lane change before the off-ramp is reached is subdivided into automatic data collection. The five steps of determining the theoretical safety distance, ranking the measured distance, classifying the measured distance and selecting and implementing the change of track are five steps to complete the construction of the forced change model. Thirdly, under the guidance of the idea of hierarchical and piecewise control, this paper puts forward the strategy of off-ramp path control for self-driving vehicles, which can be divided into ladder strategy, global strategy and hybrid strategy, and a flexible and variable simulation scenario is established. At the same time, the macro and micro indexes of strategy evaluation are determined to form the final comprehensive cost evaluation model. The calibration of the model parameters is based on the combination of expert scoring method and Analytic hierarchy process (AHP). Finally, the simulation work of the optimal control strategy for the off-ramp path of the self-driving vehicle is carried out, and its evaluation and analysis are carried out, including the basic parameters of the traffic flow, the average travel time, the changing time and the spacing. Analysis of the characteristics of the changing channel behavior including the rate of change of channel etc. Furthermore, by comparing the seven costs of changing lanes, the comprehensive cost of continuous strategy is analyzed, and the optimal path control strategy of the off-ramp of self-driving vehicles is finally determined.
【學(xué)位授予單位】：東南大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2016
【分類號(hào)】：U463.6

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 張翔;;自動(dòng)駕駛汽車技術(shù)的發(fā)展趨勢(shì)[J];汽車電器;2015年08期

2 夏n，

本文編號(hào)：2217686