CASIA OpenIR  > 毕业生  > 博士学位论文
Abstract本论文结合人工智能和智能控制方面的理论和方法,以网络控制和移动Agent技术为基础,针对交通信号控制问题的特点,从交通控制的实际情况和需要出发,从底层路口的控制算法开始进行研究,然后提出基于Agent的路口信号控制系统,该控制系统上层的Agent调度采用基于CBR的策略,最后对上层需要的优化算法等进行了研究。主要内容如下: 1.设计单个路口多相位模糊逻辑控制算法。该算法与普通的模糊控制方法不同,不仅考虑了当前绿灯相位和下一通行相位的车辆排队情况,还考虑了其它相位的车辆排队情况,使控制更接近人的思维。由于输入量较多,为了避免模糊规则出现维数爆炸,算法采用了分级实现。 2.对干线交通控制进行了研究,提出了模糊补偿控制的算法及其实现。该算法结合了传统数学模型和智能控制技术两方面的优点,首先对确定性因素建立基于线性规划的数学模型,以绿波带的宽度作为目标函数进行求解,然后对实际控制中出现的不确定性因素进行补偿控制,对每个路口的相位差和绿信比进行调节。仿真结果表明控制效果要好于感应控制。 3.提出了基于网络控制和移动Agent的城市路口信号系统以及控制的实现策略。利用Agent技术把不同的交叉口信号控制方案包装成可在广域网上分布转移的Agent。并由位于网络调度中心的网络统一调度器根据交叉口实际交通情况的变化进行相应移动控制Agent的派遣或回收,以达到更好的控制效果, 同时也为各种移动Agent的学习和参数优化提供必要的条件和保障。这样就不需要底层交叉口信号控制器同时支持其工作所需要的所有的移动控制Agent,可大大降低交叉口信号控制器所需要的内存和计算能力,特别适合当前流行的基于嵌入式操作系统的交叉口信号控制器。 4.对网络调度中心的调度策略进行了研究,提出了基于CBR的Agent调度策略。CBR(Case Based Reasoning)通常称为基于案例的推理,是人工智能领域中的一个重要范畴,旨在利用已有的案例去解决新问题。通过基于CBR的Agent调度策略可以对道路交叉口进行有效的控制。采用TSIS、VC和MATLAB组成的仿真平台,结合前面的控制算法对控制策略实现进行了研究,仿真的结果表明本文的调度算法无论是单个路口还是路网均取得了良好的控制效果,确实改善了路口的性能,并且能够对诸如交通事故等意外情况采取迅速有效的应对方案。 5.对控制系统的网络调度中心的支撑算法进行了研究。调度中心不仅要根据实际的交通状况对Agent进行调度,还需要相应的支撑算法对交通控制进行相应的辅助和决策。主要研究了两种算法:交通量预测和基于有序样聚类的TOD参数优化算法。
Other AbstractStudy is done from low-level intersection control, and then the intersection control system based on agent is presented. The strategy of Case Based Reasoning is adopted to schedule agent in high-level. At last, the optimal algorithms in high-level are studied. The main contents are as follow: 1.The multi-phase fuzzy control algorithm for a single intersection is designed. This algorithm is different from general fuzzy control. Not only queues of current and next phases but also those of other phases are considered. So the control is more close to the human thinking. Since more variables are inputted, algorithm is realized by two stages in order to avoid the dimension explosion of fuzzy rules. 2.The arterial control is studied, and the algorithm of fuzzy compensation control method is presented. It combines advantages of traditional mathematic model and intelligent technology. Determining facts in traffic are modeled by traditional linear programming at first, and the initial timing scheme is available. Then the fuzzy control is used to compensate uncertain and random facts, and the phase difference and split of every intersection are adjusted. The simulation result illuminates that its effect is better than that of actuated control. 3.Urban intersection signal system based on networked control and mobile agent is presented, and the realization strategy of signal control is studied. By using agent technology, various intersection control algorithms are packaged as agents which can be transferred at WAN. Dispatcher which is at network dispatch center sends and retracts the mobile control agent according to the traffic status of the intersection. In this way, the intersection controller doesn’t need all of the mobile control agents to support its mission, the requirement for memory and computing capacity of intersection controller will be decreased, and it is suitable to the current intersection controller based on the embedded operation system. 4.The dispatch strategy of network dispatch center is studied, and the dispatch strategy based on the CBR (Case Based Reasoning) is presented. CBR is an important content in artificial intelligence, and it solves the new problem by using the existed cases. The agent dispatch strategy based on CBR can effectively control the intersection signal. The simulation result shows that it improves the performance of traffic in either single intersection or road network, and it can quickly and effectively deal with the unexpected situation, such as traffic incident. 5.The support algorithms of the network dispatch center of control system are studied. The dispatch center not only sends agent according to the real traffic status, but also assumes responsibility of agents’ study and optimization. Two algorithms are mainly studied: traffic flow prediction and TOD parameter optimization based on sequence sample cluster.
Keyword交通控制 网络控制 基于agent控制 网络调度中心 当地简单 交通流预测 远程复杂 智能控制 交通仿真 Traffic Control Networked Control Agent-based Control Network Dispatch Center Local Simple Traffic Flow Prediction Remote Complex Intelligent Control Traffic Simulation
Document Type学位论文
Recommended Citation
GB/T 7714
李伟. 基于网络的智能路口信号控制系统的算法及实现[D]. 中国科学院自动化研究所. 中国科学院研究生院,2006.
Files in This Item:
File Name/Size DocType Version Access License
CASIA_20021801460316(861KB) 暂不开放CC BY-NC-SA
Related Services
Recommend this item
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[李伟]'s Articles
Baidu academic
Similar articles in Baidu academic
[李伟]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[李伟]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.