【摘要】：人群運(yùn)動仿真技術(shù)和計算機(jī)技術(shù)的長足發(fā)展,大規(guī)模人群運(yùn)動仿真也逐漸成為新興的研究熱點。如何更加全面而準(zhǔn)確地研究并解決仿真中的問題,使仿真效果更具有真實感和高效性成為了研究者的重要課題。所以對大規(guī)模人群運(yùn)動仿真的研究具有深遠(yuǎn)的學(xué)術(shù)意義和極大的現(xiàn)實價值。為了實現(xiàn)高效、真實的人群運(yùn)動仿真,本文以大規(guī)模人群運(yùn)動為研究對象,深入分析流體動力學(xué),將流體仿真的方法應(yīng)用到大規(guī)模人群運(yùn)動的仿真中,實現(xiàn)兩者的結(jié)合,提出了一種簡潔靈活的人群運(yùn)動控制方法。本文詳細(xì)研究了基于徑向基函數(shù)(Radial basis function,RBF)的群組運(yùn)動仿真方法,研究其主要思想,完成實驗,并在其基礎(chǔ)上展開了本文的工作。本文采用了全局控制的框架,模擬在二維情景下的大規(guī)模人群行走。本文有別于RBF方法,通過求解納維-斯托克斯(Navier-Stokes,NS)方程構(gòu)建一個覆蓋全局的速度場來驅(qū)動群組運(yùn)動,在模擬過程中,允許用戶通過交互操作改變NS方程組的參數(shù)來改變速度場,快速地改變?nèi)巳旱倪\(yùn)動。另外,在局部(個體運(yùn)動)控制方面,提出了基于光滑粒子流體動力學(xué)(Smoothed Particle Hydrodynamics,SPH)的碰撞處理策略。利用SPH方法來模擬個體的碰撞檢測和碰撞避免等行為,避免了復(fù)雜包圍盒碰撞檢測技術(shù),大大提高了模擬效率。在實驗階段,我們實現(xiàn)了本文算法,并將其和RBF仿真方法、Treuille勢能場仿真方法、Narain仿真算法進(jìn)行了比較。實驗結(jié)果顯示了本文基于GPU計算的算法在大規(guī)模人群運(yùn)動仿真上有良好的效果,在復(fù)雜場景下將有較好的應(yīng)用前景。本文算法使人群運(yùn)動符合水流運(yùn)動的規(guī)律,能更好的表現(xiàn)人群的行走效果。
[Abstract]:With the rapid development of crowd motion simulation technology and computer technology, mass crowd motion simulation has gradually become a new research hotspot. How to study and solve the problems in simulation more comprehensively and accurately has become an important task for researchers to make the simulation effect more realistic and efficient. Therefore, the research of large-scale crowd motion simulation has profound academic significance and great practical value. In order to realize efficient and real crowd movement simulation, this paper takes large-scale crowd motion as the research object, deeply analyzes fluid dynamics, applies fluid simulation method to large-scale crowd movement simulation, and realizes the combination of them. A simple and flexible method for crowd movement control is proposed. In this paper, the method of group motion simulation based on radial basis function (Radial basis function,RBF) is studied in detail, its main idea is studied, the experiment is completed, and the work of this paper is carried out on the basis of it. In this paper, a global control framework is used to simulate large-scale walking in two-dimensional scenarios. In this paper, different from RBF method, we construct a global velocity field to drive group motion by solving the Navier-Stokes,NS equation. It allows the user to change the velocity field by changing the parameters of the NS equations and to change the movement of the crowd quickly. In addition, in the aspect of local (individual motion) control, a collision processing strategy based on smooth particle hydrodynamics (Smoothed Particle Hydrodynamics,SPH) is proposed. The SPH method is used to simulate the behaviors of individual collision detection and collision avoidance. The complex bounding box collision detection technique is avoided and the simulation efficiency is greatly improved. In the experiment stage, we implement the algorithm and compare it with RBF simulation method, Treuille potential energy field simulation method and Narain simulation algorithm. The experimental results show that the algorithm based on GPU algorithm has good results in large-scale crowd motion simulation, and will have a good application prospect in complex scenarios. The algorithm of this paper makes the movement of crowd accord with the law of flow motion, and can better represent the effect of walking of the crowd.
【學(xué)位授予單位】：天津大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2016
【分類號】：TP391.9

【相似文獻(xiàn)】

相關(guān)期刊論文 前10條

1 吳躍;馬小虎;;一種基于GPU的陰影映射改進(jìn)算法[J];微計算機(jī)信息;2009年36期

2 馬伯寧;王晨昊;湯曉安;匡綱要;;基于GPU的二維離散小波變換快速計算[J];國防科技大學(xué)學(xué)報;2011年03期

3 ZW;;3D游戲利器 主流嵌入式處理器GPU逐個看[J];電腦迷;2011年19期

4 王志國;王貴錦;施陳博;苗權(quán);林行剛;;積分圖像的快速GPU計算[J];計算機(jī)應(yīng)用研究;2011年10期

5 盧永菁;王東;;基于GPU的高速網(wǎng)絡(luò)入侵檢測系統(tǒng)設(shè)計[J];計算機(jī)工程與應(yīng)用;2011年33期

6 儲t熆，

本文編號：2426998