CN102842232B

CN102842232B - FPGA (Field Programmable Gate Array) online predication control method based on Kerner-Konhauser macroscopic traffic flow model

Info

Publication number: CN102842232B
Application number: CN201210316185.8A
Authority: CN
Inventors: 史忠科; 刘通
Original assignee: Northwestern Polytechnical University
Current assignee: Northwestern Polytechnical University
Priority date: 2012-08-30
Filing date: 2012-08-30
Publication date: 2015-07-15
Anticipated expiration: 2032-08-30
Also published as: CN102842232A

Abstract

The invention discloses an FPGA online predictive control method based on a Kerner-Konhauser macro traffic flow model, which is used to solve the technical problem of poor real-time performance of the existing FPGA predictive control method. The technical solution is to establish a parallel processing flow through the approximate discretization of the model, design a dynamic data storage scheme, and use FPGA to realize the predictive control of closed road turns and variable information signs based on the Kerner-Konhauser macroscopic traffic flow model. Real-time and effective control of the traffic flow density and driving speed of the expressway is realized.

Description

FPGA Online Predictive Control Method Based on Kerner-Konhauser Macro Traffic Flow Model

技术领域 technical field

本发明涉及一种FPGA预测控制方法，特别涉及一种基于Kerner-Konhauser宏观交通流模型的FPGA在线预测控制方法。The invention relates to an FPGA predictive control method, in particular to an FPGA online predictive control method based on a Kerner-Konhauser macro traffic flow model.

背景技术 Background technique

随着经济的快速发展，汽车保有量的不断增加，交通拥挤已成为世界各国共同关注的焦点和急需解决的重要问题，交通拥挤同时也造成了严重的环境污染，在9种主要的空气污染物质中，6种直接或间接地与汽车尾气排放有关，堵车状态下汽车排出的有害物质浓度比正常行驶时高出5～6倍；此外，交通拥挤与交通事故是城市交通共生的两大问题。一方面，城市交通高峰时期密集的交通流，使得交通事故频发，极易引发严重的交通拥挤；另一方面，当堵车发生时，车辆驾驶员因为过度等待，容易失去耐心，使得交通事故发生的几率大大增加；可见交通拥挤已经成为影响全球城市可持续发展的一个全局性问题。With the rapid development of the economy and the continuous increase of car ownership, traffic congestion has become the focus of common attention and an important problem that needs to be solved urgently in all countries in the world. Traffic congestion has also caused serious environmental pollution. Among the nine major air pollutants Among them, six are directly or indirectly related to automobile exhaust emissions. The concentration of harmful substances emitted by automobiles in traffic jams is 5-6 times higher than that in normal driving. In addition, traffic congestion and traffic accidents are two major problems that coexist in urban traffic. On the one hand, the dense traffic flow during the peak period of urban traffic makes traffic accidents occur frequently, which can easily lead to serious traffic congestion; The chances of increasing significantly; it can be seen that traffic congestion has become an overall problem affecting the sustainable development of cities around the world.

为了有效疏导交通、提高高速公路的使用效率，常常使用信息显示牌作为交通信息发布和控制的手段；通常，信息显示牌及可变限速标志作为智能交通系统的重要信息发布，由监控中心计算机通过通讯网络实行远程控制、传送并显示各种图文信息、向司机及时发布不同路段的不同路面情况及各类交通信息、进行交通法规，交通知识的宣传、达到减少高速公路重现性阻塞、减少高速公路非重现性事故的影响，提高行车安全；如文献“海依拉提·巴拉提，高速公路信息显示牌设置技术探讨，大陆桥视野，2010年10月，139-140”所述，信息显示牌系统的设置机理为：(1)检测器信息收集和处理系统、(2)信息显示牌信息提供、(3)通信系统、(4)中央控制系统；信息显示牌的设置应从整个交通导行系统建设的角度出发，充分考虑导行与控制的关联，结合考虑地面道路与高架道路的综合效益，制定整体性、合理性、高效性的导行方案；信息显示牌依据设置的地点和目的的不同而采用不同的形式；一种是安装在主线上，进行主线诱导和出口诱导，以字符形式显示前方路段的交通状况如畅通、拥挤、延误等，从而使驾驶员可以转向地面道路，避开拥挤区；另一种安装在匝道入口附近，把匝道入口处的排队长度及拥挤预测情况报告给驾驶员，也可把邻近主线上的交通情况显示给匝道入口上的驾驶员，从而为他们提供合理地诱导；然而，这些方案，将高速道路入口诱导、道路主线诱导、道路出口诱导仅仅按照信息需求划分开，没有有机相结合，特别是信息显示牌的显示信息没有按照宏观交通模型预测输出自动设定，难以对高速公路的交通流密度、行车速度进行有效控制。In order to effectively guide traffic and improve the efficiency of highway use, information display boards are often used as a means of traffic information release and control; usually, information display boards and variable speed limit signs are released as important information of intelligent transportation systems, and are controlled by the monitoring center computer. Implement remote control through the communication network, transmit and display various graphic information, timely release different road conditions and various traffic information on different road sections to drivers, and publicize traffic laws and traffic knowledge to reduce recurring congestion on expressways. Reduce the impact of non-recurring accidents on the expressway and improve driving safety; as mentioned in the document "Hairati Balati, Discussion on the Setting Technology of Expressway Information Display Boards, Vision of the Land Bridge, October 2010, 139-140" As mentioned above, the setting mechanism of the information display board system is: (1) detector information collection and processing system, (2) information display board information provision, (3) communication system, (4) central control system; the information display board should be set from From the perspective of the construction of the entire traffic guidance system, fully consider the relationship between guidance and control, and consider the comprehensive benefits of ground roads and elevated roads to formulate a holistic, rational and efficient guidance plan; information display boards are based on the set Different forms are adopted depending on the location and purpose; one is installed on the main line for main line guidance and exit guidance, and displays the traffic conditions of the road ahead in the form of characters, such as smooth flow, congestion, delay, etc., so that the driver can turn to the ground Roads, avoiding congested areas; the other is installed near the entrance of the ramp, reporting the queue length and congestion forecast at the entrance of the ramp to the driver, and can also display the traffic conditions on the adjacent main line to the driver at the entrance of the ramp. To provide them with reasonable guidance; however, in these schemes, the highway entrance guidance, road mainline guidance, and road exit guidance are only divided according to information needs, and there is no organic combination, especially the information displayed on the information display board. The model prediction output is automatically set, which makes it difficult to effectively control the traffic flow density and driving speed of the expressway.

为了深入分析交通系统，国内外大量学者研究交通流模型，其中采用流体力学的观点建立的宏观和微观模型分析交通特性者居多；在宏观交通流模型中，交通流被视为由大量车辆组成的可压缩连续流体介质，研究车辆集体的平均行为、单个车辆的个体特性并不凸显；宏观交通流模型以车辆的平均密度ρ、平均速度v和流量q刻画交通流，研究它们所满足的方程；宏观模型可以更好地刻画交通流的集体行为，从而为设计有效的交通控制策略、模拟及估计道路几何改造的效果等交通工程问题提供依据；在数值计算方面，模拟宏观交通流所需时间与所研究交通系统中车辆数目无关，只与所研究道路、数值方法的选取及其中空间x、时间t的离散步长Δx和Δt有关。故此，宏观交通流模型较适合于处理大量车辆组成的交通系统的交通流问题；这类模型被国际上大多数学者用来讨论封闭道路的交通现象。In order to deeply analyze the traffic system, a large number of scholars at home and abroad have studied the traffic flow model, among which the macroscopic and microcosmic models established from the perspective of fluid mechanics are mostly used to analyze the traffic characteristics; in the macroscopic traffic flow model, the traffic flow is regarded as a system composed of a large number of vehicles The continuous fluid medium can be compressed, and the average behavior of the vehicle collective is studied, and the individual characteristics of a single vehicle are not prominent; the macroscopic traffic flow model describes the traffic flow with the average density ρ, average speed v, and flow q of vehicles, and studies the equations they satisfy; The macro model can better describe the collective behavior of traffic flow, thus providing a basis for traffic engineering problems such as designing effective traffic control strategies, simulating and estimating the effect of road geometric transformation; in terms of numerical calculation, the time required for simulating macro traffic flow and The number of vehicles in the researched traffic system has nothing to do with it, but only with the researched road, the choice of numerical method, and the discrete steps Δx and Δt of space x and time t. Therefore, the macro-traffic flow model is more suitable for dealing with traffic flow problems in a traffic system composed of a large number of vehicles; this type of model is used by most scholars in the world to discuss the traffic phenomenon of closed roads.

然而，宏观交通流模型大多数采用偏微分方程描述，即使离散形式的宏观交通流模型也很复杂，这些模型的处理通常在台式机以上的系统处理，很难使用宏观模型对封闭道路匝口和可变信息牌进行在线预测控制。However, most of the macroscopic traffic flow models are described by partial differential equations. Even the discrete macroscopic traffic flow models are very complicated. These models are usually processed by systems above desktop computers. It is difficult to use macroscopic models to analyze closed road turns and Variable information board for online predictive control.

发明内容 Contents of the invention

为了克服现有FPGA预测控制方法实时性差的不足，本发明提供一种基于Kerner-Konhauser宏观交通流模型的FPGA在线预测控制方法。该方法通过对模型近似离散化处理，建立了并行处理流程，设计了动态数据存储方案，用FPGA实现了基于Kerner-Konhauser宏观交通流模型的封闭道路匝口和可变信息牌的预测控制。可以使得高速公路的交通流密度、行车速度实现实时有效控制。In order to overcome the disadvantage of poor real-time performance of the existing FPGA predictive control method, the present invention provides an FPGA online predictive control method based on the Kerner-Konhauser macroscopic traffic flow model. In this method, the parallel processing flow is established through the approximate discretization of the model, the dynamic data storage scheme is designed, and the predictive control of closed road turns and variable information signs based on the Kerner-Konhauser macroscopic traffic flow model is realized by FPGA. It can realize real-time and effective control of the traffic flow density and driving speed of the expressway.

本发明解决其技术问题所采用的技术方案：一种基于Kerner-Konhauser宏观交通流模型的FPGA在线预测控制方法，其特点是包括以下步骤：The technical scheme that the present invention solves its technical problem adopts: a kind of FPGA online predictive control method based on Kerner-Konhauser macro-traffic flow model, it is characterized in comprising the following steps:

步骤一、根据Kerner-Konhauser宏观交通流模型：Step 1. According to the Kerner-Konhauser macroscopic traffic flow model:

$\{\begin{matrix} \frac{&PartialD; &PartialD; ρ ρ}{&PartialD; &PartialD; t t} + + \frac{&PartialD; &PartialD; ((ρv ρv))}{&PartialD; &PartialD; x x} = = π π [[r r ((x x,, t t)),, s the s ((x x,, t t))]] \\ \frac{&PartialD; &PartialD; v v}{&PartialD; &PartialD; t t} + + v v \frac{&PartialD; &PartialD; v v}{&PartialD; &PartialD; x x} = = \frac{{V V}_{e e} ((ρ ρ)) - - v v}{T T} - - \frac{{c c}_{00}^{22}}{ρ ρ} \frac{&PartialD; &PartialD; v v}{&PartialD; &PartialD; x x} + + \frac{μ μ}{ρ ρ} \frac{{&PartialD; &PartialD;}^{22} v v}{{&PartialD; &PartialD; x x}^{22}} \end{matrix}$

式中，ρ为车辆的平均密度、v为平均速度、t为时间，x为与仿真道路起始点的距离，π[r(x，t)，s(x，t)]为由于匝口进入或驶出的车流量引起的密度变化率函数，r(x，t)为由匝口进入的车流量，s(x，t)=s₀(x，t)+s_q(x，t)为由匝口驶出的车流量，s₀(x，t)为由匝口驶出的正常车流量、s_q(x，t)为信息显示牌强制驶出车辆造成的流量增量，V_e(ρ)为等价速度，T，μ均为常数；In the formula, ρ is the average density of vehicles, v is the average speed, t is time, x is the distance from the starting point of the simulated road, π[r(x, t), s(x, t)] is the Or the density change rate function caused by the exiting traffic flow, r(x, t) is the traffic flow entering from the turn, s(x, t)=s ₀ (x, t)+s _q (x, t) is the traffic flow out of the turn, s ₀ (x, t) is the normal traffic flow out of the turn, s _q (x, t) is the flow increase caused by the forced exit of the vehicle on the information display board, V _e (ρ) is the equivalent velocity, T, μ is a constant;

用差分格式表示微分项并略去高阶项，得到：Representing the differential terms in difference form and omitting higher-order terms yields:

$\frac{&PartialD; &PartialD; ρ ρ}{&PartialD; &PartialD; t t} = = \frac{ρ ρ ((x x,, t t + + ξ ξ)) - - ρ ρ ((x x,, t t))}{ξ ξ} + + o o ((ξ ξ)) = = \frac{{ρ ρ}_{i i}^{n no + + 11} - - {ρ ρ}_{i i}^{n no}}{ξ ξ}$

$\frac{&PartialD; &PartialD; ρ ρ}{&PartialD; &PartialD; x x} = = \frac{ρ ρ ((x x + + h h,, t t)) - - ρ ρ ((x x,, t t))}{h h} + + o o ((h h)) = = \frac{{ρ ρ}_{i i + + 11}^{n no} - - {ρ ρ}_{i i}^{n no}}{h h}$

$\frac{&PartialD; &PartialD; v v}{&PartialD; &PartialD; t t} = = \frac{v v ((x x,, t t + + ξ ξ)) - - v v ((x x,, t t))}{ξ ξ} + + o o ((ξ ξ)) = = \frac{{v v}_{i i}^{n no + + 11} - - {v v}_{i i}^{n no}}{ξ ξ}$

$\frac{&PartialD; &PartialD; v v}{&PartialD; &PartialD; x x} = = \frac{v v ((x x + + h h,, t t)) - - v v ((x x,, t t))}{h h} + + o o ((h h)) = = \frac{{v v}_{i i + + 11}^{n no} - - {v v}_{i i}^{n no}}{h h}$

$\frac{{&PartialD; &PartialD;}^{22} v v}{{&PartialD; &PartialD; x x}^{22}} = = \frac{v v ((x x + + h h,, t t)) - - 22 v v ((x x,, t t)) + + v v ((x x - - h h,, t t))}{{h h}^{22}} + + o o (({h h}^{22})) = = \frac{{v v}_{i i + + 11}^{n no} - - 22 {v v}_{i i}^{n no} + + {v v}_{i i - - 11}^{n no}}{{h h}^{22}}$

式中，ξ为t的微分，h为x的微分，o(ξ)为ξ的高阶无穷小，o(h)为h的高阶无穷小，o(h²)为h²的高阶无穷小，ρ(x，t)为t时刻x处车辆的平均密度，v(x，t)为t时刻x处车辆的平均速度，把道路分成多个路段，每个路段长度为h，采样周期为ξ，为第i个路段在[nξ，(n+1)ξ]内车辆的平均密度，为第i个路段在[nξ，(n+1)ξ]内车辆的平均速度；In the formula, ξ is the differential of t, h is the differential of x, o(ξ) is the higher order infinitesimal of ξ, o(h) is the higher order infinitesimal of h, o(h ² ) is the higher order infinitesimal of h ² , ρ(x, t) is the average density of vehicles at x at time t, v(x, t) is the average speed of vehicles at x at time t, the road is divided into multiple sections, the length of each section is h, and the sampling period is ξ , is the average density of vehicles in [nξ, (n+1)ξ] of the i-th road segment, is the average speed of vehicles in [nξ, (n+1)ξ] of the i-th road section;

得到Kerner-Konhauser宏观交通流模型的差分形式为：The differential form of the Kerner-Konhauser macroscopic traffic flow model is obtained as:

$\{\begin{matrix} {ρ ρ}_{i i}^{n no + + 11} = = ξπ ξπ [[r r ((i i,, n no)),, s the s ((i i,, n no))]] - - \frac{ξ ξ}{h h} [[{v v}_{i i}^{n no} (({ρ ρ}_{i i + + 11}^{n no} - - {ρ ρ}_{i i}^{n no})) + + {ρ ρ}_{i i}^{n no} (({v v}_{i i + + 11}^{n no} - - {v v}_{i i}^{n no}))]] + + {ρ ρ}_{i i}^{n no} \\ {v v}_{i i}^{n no + + 11} = = {v v}_{i i}^{n no} + + ξ ξ [[\frac{{V V}_{e e} (({ρ ρ}_{i i}^{n no})) - - {v v}_{i i}^{n no}}{T T} - - \frac{{c c}_{00}^{22} (({v v}_{i i + + 11}^{n no} - - {v v}_{i i}^{n no}))}{{ρ ρ}_{i i}^{n no} h h} - - \frac{{v v}_{i i}^{n no} (({v v}_{i i + + 11}^{n no} - - {v v}_{i i}^{n no}))}{h h} + + \frac{μ μ (({v v}_{i i + + 11}^{n no} - - 22 {v v}_{i i}^{n no} + + {v v}_{i i - - 11}^{n no}))}{{ρ ρ}_{i i}^{n no} {h h}^{22}}]] \end{matrix}$

式中，r(i，n)表示第i个路段在[nξ，(n+1)ξ]内由匝口进入的车流量，s(i，n)表示第i个路段在[nξ，(n+1)ξ]内由匝口驶出的车流量；In the formula, r(i, n) represents the traffic flow of the i-th road section entering from the turn in [nξ, (n+1)ξ], and s(i, n) represents the i-th road section in [nξ, (n+1)ξ]. n+1)ξ], the traffic flow out of the turn;

步骤二、建立等价速度模型为：Step 2: Establish an equivalent velocity model as:

式中，v₀，E，均为常数，v_ea为可变信息显示牌指定速度；In the formula, v ₀ , E, Both are constants, v _ea is the specified speed of the variable information display board;

第i路段： Section i:

式中，v₀，E，均为常数，v_ea(i)为第i个路段可变信息显示牌指定速度；In the formula, v ₀ , E, Both are constants, v _ea (i) is the specified speed of the i-th section variable information display board;

步骤三、结合Kerner-Konhauser模型的差分形式和等价速度模型，在FPGA中设计包含车辆平均密度ρ和平均速度v的计算模块，根据实际道路的长度和匝口信息把高速公路分成多个路段，每个路段对应一个计算模块，根据初始信息和调控信息，在FPGA中同时并行运行这些计算模块，预测出各个路段下一时间段的车辆平均密度和平均速度，然后把车辆平均密度和平均速度存入寄存器，在所有计算模块完成计算后，输出车辆平均密度和平均速度，同时把这些数据回传给计算模块进行下一步的计算；Step 3. Combining the differential form of the Kerner-Konhauser model and the equivalent speed model, design a calculation module including the average vehicle density ρ and the average speed v in the FPGA, and divide the expressway into multiple road sections according to the actual road length and turn information , each road section corresponds to a calculation module, according to the initial information and control information, run these calculation modules in parallel in the FPGA at the same time, predict the average vehicle density and average speed of each road section in the next time period, and then calculate the average vehicle density and average speed Stored in the register, after all calculation modules complete the calculation, output the average density and average speed of the vehicle, and send these data back to the calculation module for the next step of calculation;

步骤四、以匝口进入封闭道路流量作为模型输入，可变信息牌作为强制速度和匝口强制输出调节量，对于给定控制输入预测每个路段的平均交通流密度和车辆平均速度，如果每个路段都满足最低速度、最大密度要求，则选择该方案以控制封闭道路匝口及可变信息牌，否则调整控制方案。Step 4. Take the turn entering closed road flow as the model input, and the variable information board as the mandatory speed and turn forced output adjustment. For the given control input, predict the average traffic flow density and the average vehicle speed of each road section. If each If all road sections meet the minimum speed and maximum density requirements, then choose this scheme to control closed road turns and variable information signs, otherwise adjust the control scheme.

所述计算模块采用浮点数运算，自定义浮点数结构如下表所示：The calculation module uses floating-point calculations, and the structure of the custom floating-point numbers is shown in the following table:

1符号S 1 Symbol S 6阶码e 6th order code e 17尾数M 17 Mantissa M

共24位，其中符号1位，阶码6位，尾数17位，代表的数据大小为F=(-1)^s×1.M×2^e-31。A total of 24 bits, including 1 bit for the sign, 6 bits for the exponent code, and 17 bits for the mantissa. The data size represented is F=(-1) ^s ×1.M×2 ^e-31 .

本发明的有益效果是：由于通过对模型近似离散化处理，建立了并行处理流程，设计了动态数据存储方案，用FPGA实现了基于Kerner-Konhauser宏观交通流模型的封闭道路匝口和可变信息牌的预测控制。使得高速公路的交通流密度、行车速度实现了实时有效控制。The beneficial effects of the present invention are: due to the approximate discretization of the model, a parallel processing flow has been established, a dynamic data storage scheme has been designed, and the closed road turn and variable information based on the Kerner-Konhauser macroscopic traffic flow model have been realized with FPGA Card predictive control. Real-time and effective control of the traffic flow density and driving speed of the expressway is realized.

下面结合附图和实施例对本发明作详细说明。The present invention will be described in detail below in conjunction with the accompanying drawings and embodiments.

附图说明Description of drawings

图1是本发明基于Kerner-Konhauser宏观交通流模型的FPGA在线预测控制方法的计算结构图。FIG. 1 is a calculation structure diagram of the FPGA online predictive control method based on the Kerner-Konhauser macroscopic traffic flow model of the present invention.

图2是本发明基于Kerner-Konhauser宏观交通流模型的FPGA在线预测控制方法的FPGA实现框图。Fig. 2 is the FPGA implementation block diagram of the FPGA online predictive control method based on the Kerner-Konhauser macroscopic traffic flow model of the present invention.

具体实施方式 Detailed ways

参照图1、2详细说明本发明。The present invention will be described in detail with reference to FIGS. 1 and 2 .

1、根据Kerner-Konhauser宏观交通流模型：1. According to the Kerner-Konhauser macro traffic flow model:

式中：ρ为车辆的平均密度、v为平均速度、t为时间，x为与仿真道路起始点的距离，π[r(x，t)，s(x，t)]为由于匝口进入或驶出的车流量引起的密度变化率函数，r(x，t)为由匝口进入的车流量，s(x，t)=s₀(x，t)+s_q(x，t)为由匝口驶出的车流量，s₀(x，t)为由匝口驶出的正常车流量、s_q(x，t)为信息显示牌强制驶出车辆造成的流量增量，V_e(ρ)为等价速度，T，μ均为常数，全申请书符号定义相同；In the formula: ρ is the average density of vehicles, v is the average speed, t is time, x is the distance from the starting point of the simulated road, π[r(x, t), s(x, t)] is the Or the density change rate function caused by the exiting traffic flow, r(x, t) is the traffic flow entering from the turn, s(x, t)=s ₀ (x, t)+s _q (x, t) is the traffic flow out of the turn, s ₀ (x, t) is the normal traffic flow out of the turn, s _q (x, t) is the flow increase caused by the forced exit of the vehicle on the information display board, V _e (ρ) is the equivalent velocity, T, μ are constants, and the definitions of the symbols in the whole application form are the same;

式中：ξ为t的微分，h为x的微分，o(ξ)为ξ的高阶无穷小，o(h)为h的高阶无穷小，o(h²)为h²的高阶无穷小，ρ(x，t)为t时刻x处车辆的平均密度，v(x，t)为t时刻x处车辆的平均速度，把道路分成多个路段，每个路段长度为h，采样周期为ξ，为第i个路段在[nξ，(n+1)ξ]内车辆的平均密度，为第i个路段在[nξ，(n+1)ξ]内车辆的平均速度；In the formula: ξ is the differential of t, h is the differential of x, o(ξ) is the higher order infinitesimal of ξ, o(h) is the higher order infinitesimal of h, o(h ² ) is the higher order infinitesimal of h ² , ρ(x, t) is the average density of vehicles at x at time t, v(x, t) is the average speed of vehicles at x at time t, the road is divided into multiple sections, the length of each section is h, and the sampling period is ξ , is the average density of vehicles in [nξ, (n+1)ξ] of the i-th road segment, is the average speed of vehicles in [nξ, (n+1)ξ] of the i-th road section;

式中：r(i，n)表示第i个路段在[nξ，(n+1)ξ]内由匝口进入的车流量，s(i，n)表示第i个路段在[nξ，(n+1)ξ]内由匝口驶出的车流量；In the formula: r(i, n) represents the traffic flow of the i-th road segment entering from the turn in [nξ, (n+1)ξ], s(i, n) represents the i-th road segment in [nξ, (n+1)ξ] n+1)ξ], the traffic flow out of the turn;

2、建立等价速度模型为：2. Establish the equivalent velocity model as:

第i路段： Section i:

3、在本实施例中，FPGA芯片选用Altera公司的EP3C120F484C6芯片，与上位机通信采用RS-232协议，电平转换芯片选用MAX3232芯片；然后在FPGA中按附图1所示计算结构对各个路段进行仿真计算。本实施例中把道路分成10个路段，附图2中计算模块1-计算模块10为按照前述偏微分方程组的差分解法使用浮点数运算器组合而成的路段仿真计算模块，具体的数据流向为：数据接收模块接收上位机传来的各个路段的交通流密度、平均速度的初始数据以及调控数据（包括各路段匝口进入车流量、驶出车流量和等价速度），然后传给数据分配模块，数据分配模块将使能信号和这些初始数据传给各个计算模块，各个计算模块接收到使能信号后同时对车辆平均密度和平均速度进行仿真计算并把结果存入寄存器，各个模块计算结束后把各自的计算结束信号传给同步模块，同步模块在所有计算模块完成计算后发送信号通知数据分配模块和数据输出模块接收车辆平均密度和平均速度的仿真结果，数据分配模块再把各路段的仿真结果和调控信息分配给计算模块进行下一步计算，同时数据输出模块输出仿真结果；3, in the present embodiment, FPGA chip selects the EP3C120F484C6 chip of Altera Company for use, and adopts RS-232 agreement with host computer communication, and level conversion chip selects MAX3232 chip for use; Perform simulation calculations. In the present embodiment, the road is divided into 10 road sections. Calculation module 1-calculation module 10 is a road section simulation calculation module combined with floating-point arithmetic unit according to the difference decomposition method of the aforementioned partial differential equations in the accompanying drawing 2, and the specific data flow direction It is: the data receiving module receives the initial data of traffic flow density and average speed of each road section and the control data (including the entering traffic flow, exiting traffic flow and equivalent speed) of each road section from the upper computer, and then transmits the data Distribution module, the data distribution module transmits the enable signal and these initial data to each calculation module, and each calculation module performs simulation calculation on the average vehicle density and average speed after receiving the enable signal and stores the result in the register, and each module calculates After the completion, the respective calculation end signals are sent to the synchronization module. After all the calculation modules complete the calculation, the synchronization module sends a signal to notify the data distribution module and the data output module to receive the simulation results of the average vehicle density and average speed. The simulation results and control information are assigned to the calculation module for the next calculation, and the data output module outputs the simulation results;

4、所述浮点数运算器采用自定义浮点数格式，浮点数结构如下表所示：4. The floating-point arithmetic unit adopts a custom floating-point format, and the structure of the floating-point number is shown in the following table:

1符号S 1 Symbol S 6阶码e 6th order code e 17尾数M 17 Mantissa M

共24位，其中符号1位，阶码6位，尾数17位，代表的数据大小为F=(-1)^s×1.M×2^e-31；A total of 24 bits, including 1 bit for the symbol, 6 bits for the exponent code, and 17 bits for the mantissa. The data size represented is F=(-1) ^s ×1.M×2 ^e-31 ;

所述数据接收模块接收上位机传来的8位的数据，并把连续三个8位的数据转化为24位数据传给数据分配模块；The data receiving module receives the 8-bit data from the upper computer, and converts three consecutive 8-bit data into 24-bit data and passes it to the data distribution module;

所述数据输出模块接收计算模块传来的24位计算结果，把它们拆分成8位的数据进行输出，在输出计算结果之前先输出有效数据开始识别码0XFF，0XF1，0XF1，计算结果输出完毕之后输出有效数据结束识别码0XFF，0XF2，0XF2；The data output module receives the 24-bit calculation results from the calculation module, splits them into 8-bit data for output, and outputs valid data start identification codes 0XFF, 0XF1, 0XF1 before outputting the calculation results, and the output of the calculation results is completed Then output valid data end identification code 0XFF, 0XF2, 0XF2;

5、以匝口进入封闭道路流量作为模型输入，可变信息牌作为强制速度和匝口强制输出调节量，对于给定控制输入预测各个路段的交通密度和车辆平均速度，如果每个路段都满足最低速度、最大密度要求，则选择该方案以控制封闭道路匝口及可变信息牌，否则调整控制方案。5. Take the flow of the turn entering the closed road as the model input, and the variable information board as the mandatory speed and the output adjustment of the turn. For the given control input, predict the traffic density and average vehicle speed of each road section. If each road section satisfies If minimum speed and maximum density are required, then select this scheme to control closed road turns and variable information signs, otherwise adjust the control scheme.

Claims

1., based on a FPGA on-line prediction control method for Kerner-Konhauser macroscopic traffic flow, it is characterized in that comprising the following steps:

Step one, according to Kerner-Konhauser macroscopic traffic flow:

\{\begin{matrix} \frac{&PartialD; ρ}{&PartialD; t} + \frac{&PartialD; (ρv)}{&PartialD; x} = π [r (x, t), s (x, t)] \\ \frac{&PartialD; v}{&PartialD; t} + v \frac{&PartialD; v}{&PartialD; x} = \frac{V_{e} (ρ) - v}{T} - \frac{c_{0}^{2}}{ρ} \frac{&PartialD; v}{&PartialD; x} + \frac{μ}{ρ} \frac{{&PartialD;}^{2} v}{&PartialD; x^{2}} \end{matrix}

In formula, ρ is the average density of vehicle, v is average velocity, t is the time, x is current location and the distance emulating road starting point, π [r (x, t), s (x, t) be] the rate of change of the density function that the vehicle flowrate entering due to circle mouth or roll away from causes, r (x, t) is the vehicle flowrate entered by circle mouth, s (x, t)=s ₀(x, t)+s _q(x, t) vehicle flowrate for being rolled away from by circle mouth, s ₀(x, t) normal vehicle flowrate, s for being rolled away from by circle mouth _qthe flow increment that (x, t) forces outgoing vehicles to cause for information displaying board, V _e(ρ) be equivalent speed, T, μ is constant;

Represent differential term by difference scheme and omit higher order term, obtaining:

\frac{&PartialD; ρ}{&PartialD; t} = \frac{ρ (x, t + ξ) - ρ (x, t)}{ξ} + o (ξ) = \frac{ρ_{i}^{n + 1} - ρ_{i}^{n}}{ξ}

\frac{&PartialD; ρ}{&PartialD; x} = \frac{ρ (x + h, t) - ρ (x, t)}{h} + o (h) = \frac{ρ_{i + 1}^{n} - ρ_{i}^{n}}{h}

\frac{&PartialD; v}{&PartialD; t} = \frac{v (x, t + ξ) - v (x, t)}{ξ} + o (ξ) = \frac{v_{i}^{n + 1} - v_{i}^{n}}{ξ}

\frac{&PartialD; v}{&PartialD; x} = \frac{v (x + h, t) - v (x, t)}{h} + o (h) = \frac{v_{i + 1}^{n} - v_{i}^{n}}{h}

\frac{{&PartialD;}^{2} v}{&PartialD; x^{2}} = \frac{v (x + h, t) - 2 v (x, t) + v (x - h, t)}{h^{2}} + o (h^{2}) = \frac{v_{i + 1}^{n} - 2 v_{i}^{n} + v_{i - 1}^{n}}{h^{2}}

In formula, ξ is the step-length to t differential, i.e. sampling period; H is the differential step size to x, i.e. each road section length; The higher-order shear deformation that o (ξ) is ξ, the higher-order shear deformation that o (h) is h, o (h ²) be h ²higher-order shear deformation, ρ (x, t) is the average density of t x place vehicle, and v (x, t) is the average velocity of t x place vehicle, and road is divided into multiple section, be the average density of i-th section at [n ξ, (n+1) ξ] interior vehicle, be the average velocity of i-th section at [n ξ, (n+1) ξ] interior vehicle;

The difference form obtaining Kerner-Konhauser macroscopic traffic flow is:

\{\begin{matrix} ρ_{i}^{n + 1} = ξπ [r (i, n), s (i, n)] - \frac{ξ}{h} [v_{i}^{n} (ρ_{i + 1}^{n} - ρ_{i}^{n}) + ρ_{i}^{n} (v_{i + 1}^{n} - v_{i}^{n})] + ρ_{i}^{n} \\ v_{i}^{n + 1} = v_{i}^{n} + ξ [\frac{V_{e} (ρ_{i}^{n}) - v_{i}^{n}}{T} - \frac{c_{0}^{2} (v_{i + 1}^{n} - v_{i}^{n})}{ρ_{i}^{n} h} - \frac{v_{i}^{n} (v_{i + 1}^{n} - v_{i}^{n})}{h} + \frac{μ (v_{i + 1}^{n} - 2 v_{i}^{n} + v_{i - 1}^{n})}{ρ_{i}^{n} h^{2}}] \end{matrix}

In formula, r (i, n) represents that i-th section is at [n ξ, (n+1) ξ] the interior vehicle flowrate entered by circle mouth, s (i, n) represents the vehicle flowrate that i-th section is rolled away from by circle mouth in [n ξ, (n+1) ξ];

Step 2, set up equivalent speed model and be:

In formula, v ₀, E, be constant, v _eafor variable information display board command speed;

I-th section:

In formula, v ₀, E, be constant, v _eai () is i-th section variable information display board command speed;

Step 3, in conjunction with difference form and the equivalent speed model of Kerner-Konhauser model, in FPGA, design comprises the computing module of vehicle average density ρ and average velocity v, according to the length of real road and circle message breath, highway is divided into multiple section, the corresponding computing module in each section, according to initial information and regulation and controlling of information, these computing modules of parallel running simultaneously in FPGA, dope vehicle average density and the average velocity of each section subsequent time period, then vehicle average density and average velocity stored in register, after all computing modules complete calculating, export vehicle average density and average velocity, these data back are carried out next step calculating to computing module simultaneously,

Step 4, enter blocked road flow as mode input using circle mouth, changeable message signs force Drazin inverse amount as pressure speed and circle mouth, given control inputs is predicted to average traffic current density and the vehicle average velocity in each section, if each section meets minimum speed, maximal density requirement, then select the program to control blocked road circle mouth and changeable message signs, otherwise adjustment control program.