【作者】 杨洁；

【导师】 韩秀华；

【作者基本信息】 吉林大学 ， 交通运输规划与管理， 2007， 硕士

【摘要】 本文针对近年来高速公路上频繁发生的低能见度引发交通事故的问题,从驾驶员视觉特性等方面分析了能见度降低时,速度变化与行车安全之间的关系。并结合低能见度条件下速度控制的理论依据,与最近发展起来的自适应神经模糊推理系统原理,研究了一种速度控制方法,对高速公路上行驶车辆的速度控制研究和工程实践具有一定的现实意义。在此基础上,本文系统的阐述了自适应神经模糊推理系统的基本原理,并将该方法用于低能见度条件下高速公路的车速控制。建立了低能见度条件下高速公路车速控制的自适应神经模糊分析模型,并对模型进行了训练,对模型的拟和性进行了有效性检验。计算分析表明,所建立的模糊神经网络模型能够描述输入、输出的映射规律。该方法对于将来工程实践中的进一步研究应用能起到一定的参考作用。更多还原

【Abstract】 Rapid economic development is leading to the traffic congestion, increased accidents and worsening pollution problems. Fog made a further expand of the expressway, which has become the largest natural disasters. Therefore, during the highway infrastructure, it is an urgent task to study the advanced technologies of highway fog management for the highway managers and researchers.This paper mainly deals with the accidents that have frequently occurred in highway, which caused by poor visibility in recent years. From drivers’visual characteristics ,we analyzed the relationship between velocity and traffic safety when the visibility has reduced. Combined the Speed control of low visibility conditions theoretical basis with the recent development of the Adaptive Neural Fuzzy Inference System Theory, we also studied a method of speed control. For the speed control of vehicles traveling on the expressway and the project of engineering, this may have some practical significance.The highway speed impact can be attributed to many factors, so the idea of the precise control of speed is difficult to achieve. Therefore, researchers have tended to focus on the macro level. For example, fog zone to establish an early warning system to monitor traffic, traffic control, transportation organizations and so on. But seldom focused on a single vehicle speed control. This paper is based on this background, to try a new method, which can link visibility and the speed of vehicles to achieve a relatively safe speed value through an adaptive analysis of real-time data on traffic.Based on the above ideas, this article made VI chapters to Explained. The main content of the chapters are below:Chapter1 Introduction. In this chapter we introduce the purpose and significance of the research, and introduce the current research status in domestic and foreign countries, study contents and so on. Chapter2 Impact of changing speed on the highway traffic safety in the fog. Introduce the characteristics of expressway accident in fog, analyze the bumping accident and the relationship between speed and accidents.Chapter3 The speed control under low visibility. Mainly expounded the theoretical basis of speed control under low visibility. And classification at home and abroad.Chapter4 Adaptive Neuro-Fuzzy Inference System. I introduced this chapter for the follow-up to pave the way for the content. We introduced the system’s structure, learning algorithm and its approximation.Chapter5 The model of Adaptive Neuro-Fuzzy Inference System. I explained the general steps of building the model of adaptive neural networks and fuzzy inference system. Its membership function types of input variables, determination of the number of input variables and space division.Chapter6 ANFIS in highway regulating control applied research. Before the union states the theoretical analysis and under this article aim visibility condition highway speed control concrete question, establishes the initialization model, and carries on the training and the validity check to the model.Finally, summary. In this chapter, we summarized the place which the present paper main research content, the main conclusion and waits for further studies.The main viewpoint and the conclusion have following several points:1. The auto-adapted nerve network fuzzy inference method uses in the highway regulating control, has the quicker network convergence rate.2. Has very good drawing up after the training ANFIS model with the ability.3. In the ANFIS model which actual center may use take massive is accurate, has the representative training sample collection as the basic premise.4. When the consideration influence speed factor integer compares the youth, the input variable space division method should use the grid division law. But works as the input variable the integer big increase, also has when the training sample collection very are also many, then should use the subtraction to gather a kind of law.Waits for question which further studies:1. The highway speed control question is a tendency, the non- definite question, also on the influence highway the vehicles speed factor has very much. So it is unable to obtain the corresponding sample data through the investigation and study. Under this condition, should through to the many kinds of theories method research (theory and knowledge, computation and actual) gradually consummates the speed control question the analysis and the design process.2. This article theoretically was only aiming at the speed control question to establish the model, and has carried on the model validity check. Has not carried on the simulation experiment to it, its practical application research also needs massively to control the theory and the control engineering aspect knowledge.3. Because the auto-adapted nerve fuzzy inference system is only suitable for the multi- inputs list output situation, therefore, should study can realize inputs the multi- outputs the auto-adapted nerve fuzzy inference method, opens up its application scope.4. This article studies the speed control model mainly or, in the actual information which says in view of the quota data also has the massive qualitative data. How but realizes these two kind of information pale pinkish purple modeling, but also waits for further studies. In recent years appeared the thick collection theory, the statistics studied and supports information processing method and soon the vector machine has the possibility to provide some new mentalities and the key to the situation.更多还原