基于速度障礙的結伴行人運動模型
發(fā)布時間：2018-09-11 21:42
【摘要】：在大型交通樞紐、購物中心、道路的設計和規(guī)劃中,行人的安全以及高效運行越來越得到重視。為了幫助設計師和工程師了解行人流在步行設施中的運動特性,需要有一個合理的行人運動模型體系,并在該體系上建立仿真軟件。結伴行人是人流的重要組成部分,但是對結伴效應的研究卻十分稀少。在微觀上結伴行人是怎樣采取避讓行為的,在宏觀上結伴效用對人流有怎樣的影響至今仍然沒有定論。正是出于這樣的考慮,本文希望建立一種全新的結伴行人模型,其能夠直觀并且行之有效的表現(xiàn)出結伴行人之間的社會聯(lián)系,并能夠采取合理的避讓策略,為建立大型的行人仿真模型提供一些幫助和思路。本文具體研究內容如下:(1)基礎理論框架確定。本文分析了現(xiàn)今在獨立行人建模中常用模型的優(yōu)缺點,總結了目前結伴行人的國內外研究現(xiàn)狀。由于速度障礙理論在機器人避讓運動中的成功運用,因此本文決定采用速度障礙理論作為模型的基礎框架,并在這個計算框架下構建結伴行人模型。(2)行人步行特性分析。在微觀建模前需要對模型產(chǎn)生的結果有一個基本的認識,這需要足夠多的基礎數(shù)據(jù),并詳細的了解行人運動的微觀特性以及宏觀特性。因此本文從行人速度、空間需求、行人間的交互作用、行人交通流特性、人流的自組織現(xiàn)象,結伴行人空間結構六個方面說明了現(xiàn)在的最新研究情況,提出了建模時需要考慮的因素。(3)模型建立。由于速度障礙理論能夠確保單個行人步行時不發(fā)生碰撞,因此本文在此基礎上構建結伴模型,提出用距離約束線的概念將分散的行人聯(lián)系起來。先用組內約束線構建雙人結伴和三人結伴組,再用組外約束線將獨立行人,二人組以及三人組連接起來構成大范圍的結伴行人組。隨后,本文提出了一種算法,將距離約束轉化成速度約束,并確保模型始終有解。(4)模型標定,仿真與驗證。首先利用視頻數(shù)據(jù)標定了模型中的系數(shù),其次設計了雙人結伴和三人結伴對向實驗,分別驗證了兩種大組組成單元的避讓特性。在確保低密度無碰撞后,設計了一種圓形的高密度交互實驗,雖然在中心區(qū)域難以避免的出現(xiàn)了重疊現(xiàn)象,但是在高密度下的避讓特性仍然是可接受的。最后的宏觀實驗中,模型生成了分層現(xiàn)象,并且有一個合理的基本圖結果,因此認為,本文所建立的多人結伴避讓模型是有效的。
[Abstract]:In the design and planning of large transportation hubs, shopping centers and roads, pedestrian safety and efficient operation have been paid more and more attention. In order to help designers and engineers understand the movement characteristics of pedestrian flow in walking facilities, a reasonable pedestrian movement model system is needed and a simulation software is established on the system. Pedestrian is an important part of pedestrian flow, but the study of the effect of pedestrian is very rare. At the micro level, how to avoid the behavior of companion pedestrian, and how the effect of partnership effect on the flow of people is still uncertain. It is out of this consideration that this paper hopes to establish a new pedestrian model, which can directly and effectively show the social relationship between pedestrians, and can take reasonable avoidance strategies. It provides some help and train of thought for the establishment of large pedestrian simulation model. The main contents of this paper are as follows: (1) the basic theoretical framework is determined. This paper analyzes the advantages and disadvantages of the models commonly used in independent pedestrian modeling, and summarizes the current research situation of companion pedestrians at home and abroad. Due to the successful application of the velocity obstacle theory in the robot avoiding movement, this paper decides to adopt the velocity obstacle theory as the basic framework of the model, and build a pedestrian model under this framework. (2) pedestrian walking characteristics analysis. Before modeling, we need to have a basic understanding of the results of the model, which requires enough basic data, and detailed understanding of the microscopic and macro characteristics of pedestrian movement. Therefore, this paper explains the latest research situation from six aspects: pedestrian speed, space demand, human interaction, pedestrian traffic flow characteristics, pedestrian flow self-organization phenomenon, pedestrian spatial structure. The factors that need to be considered in modeling are put forward. (3) Modeling. Because the theory of velocity obstacle can ensure that the pedestrian does not collide on foot, this paper constructs a joint model based on this theory, and puts forward the concept of distance constraint line to connect the scattered pedestrians. In this paper, a two-person and three-person group is constructed with the constraint line within the group, and then the independent pedestrian group, the two-person group and the three-person group are connected together to form a large range of pedestrian groups by using the constraint line outside the group. Then, an algorithm is proposed to convert the distance constraint into the velocity constraint, and to ensure that the model always has a solution. (4) Model calibration, simulation and verification. First, the coefficients of the model are calibrated with video data, and then the experiments of double and triple pairs are designed to verify the avoidance characteristics of two large groups of units. A circular high-density interactive experiment is designed to ensure the low density without collision. Although there is an inevitable overlap in the central region, the avoidance property under the high density is still acceptable. Finally, in the macroscopic experiment, the delamination phenomenon is generated and a reasonable basic graph result is obtained. Therefore, it is considered that the model is effective.
【學位授予單位】：南京理工大學
【學位級別】：碩士
【學位授予年份】：2017
【分類號】：TP391.9;U491.226
本文編號：2237940
[Abstract]:In the design and planning of large transportation hubs, shopping centers and roads, pedestrian safety and efficient operation have been paid more and more attention. In order to help designers and engineers understand the movement characteristics of pedestrian flow in walking facilities, a reasonable pedestrian movement model system is needed and a simulation software is established on the system. Pedestrian is an important part of pedestrian flow, but the study of the effect of pedestrian is very rare. At the micro level, how to avoid the behavior of companion pedestrian, and how the effect of partnership effect on the flow of people is still uncertain. It is out of this consideration that this paper hopes to establish a new pedestrian model, which can directly and effectively show the social relationship between pedestrians, and can take reasonable avoidance strategies. It provides some help and train of thought for the establishment of large pedestrian simulation model. The main contents of this paper are as follows: (1) the basic theoretical framework is determined. This paper analyzes the advantages and disadvantages of the models commonly used in independent pedestrian modeling, and summarizes the current research situation of companion pedestrians at home and abroad. Due to the successful application of the velocity obstacle theory in the robot avoiding movement, this paper decides to adopt the velocity obstacle theory as the basic framework of the model, and build a pedestrian model under this framework. (2) pedestrian walking characteristics analysis. Before modeling, we need to have a basic understanding of the results of the model, which requires enough basic data, and detailed understanding of the microscopic and macro characteristics of pedestrian movement. Therefore, this paper explains the latest research situation from six aspects: pedestrian speed, space demand, human interaction, pedestrian traffic flow characteristics, pedestrian flow self-organization phenomenon, pedestrian spatial structure. The factors that need to be considered in modeling are put forward. (3) Modeling. Because the theory of velocity obstacle can ensure that the pedestrian does not collide on foot, this paper constructs a joint model based on this theory, and puts forward the concept of distance constraint line to connect the scattered pedestrians. In this paper, a two-person and three-person group is constructed with the constraint line within the group, and then the independent pedestrian group, the two-person group and the three-person group are connected together to form a large range of pedestrian groups by using the constraint line outside the group. Then, an algorithm is proposed to convert the distance constraint into the velocity constraint, and to ensure that the model always has a solution. (4) Model calibration, simulation and verification. First, the coefficients of the model are calibrated with video data, and then the experiments of double and triple pairs are designed to verify the avoidance characteristics of two large groups of units. A circular high-density interactive experiment is designed to ensure the low density without collision. Although there is an inevitable overlap in the central region, the avoidance property under the high density is still acceptable. Finally, in the macroscopic experiment, the delamination phenomenon is generated and a reasonable basic graph result is obtained. Therefore, it is considered that the model is effective.
【學位授予單位】：南京理工大學
【學位級別】：碩士
【學位授予年份】：2017
【分類號】：TP391.9;U491.226
【參考文獻】
相關期刊論文 前7條
1 王璐;葉順強;謝能剛;;基于臟臉博弈模型的人車路口穿越行為研究[J];物理學報;2015年12期
2 童蔚蘋;程琳;;考慮結伴行為的行人流場域元胞自動機模型[J];系統(tǒng)工程理論與實踐;2014年09期
3 任剛;丁晨滋;陸麗麗;于晨;姚夢佳;;基于元胞自動機的結伴過街行人流建模與仿真[J];交通運輸系統(tǒng)工程與信息;2014年02期
4 任剛;陸麗麗;王煒;;基于元胞自動機和復雜網(wǎng)絡理論的雙向行人流建模[J];物理學報;2012年14期
5 孫澤;賈斌;李新剛;;基于元胞自動機的行人和機動車相互干擾機理研究[J];物理學報;2012年10期
6 賈洪飛;楊麗麗;唐明;;綜合交通樞紐內部行人流特性分析及仿真模型參數(shù)標定[J];交通運輸系統(tǒng)工程與信息;2009年05期
7 龔曉嵐;魏中華;;行人交通流自由速度模型研究[J];北京工業(yè)大學學報;2009年04期
，本文編號：2237940
本文鏈接：http://www.sikaile.net/kejilunwen/daoluqiaoliang/2237940.html
