传感器网络在智能交通的应用

2023-09-28 来源：好走旅游网

An Intelligent Guiding and Controlling System for Transportation Network

Based on Wireless Sensor Network Technology

CHEN Wenjie, GAO Liqiang, CHAI ZhileiˈCHEN Zhanglong, TU Shiliang

IDepartment of Computer Science and Engineering, Fudan University

220 Handan Rd., Shanghai, 200433, China

{wenjiechen, gaoliqiang, zlchai, chenzl, sltu}@fudan.edu.cn

Abstract

This paper proposes architecture based on Wireless Sensor Network (WSN) technology for Intelligent Transportation System (ITS) of a transportation network. With the help of WSN technology, the traffic info of the network can be accurately measured in real time. Based on this architecture, an optimization algorithm is proposed to minimize the average travel time for the vehicles in the network. Compared to randomly-chosen algorithm, simulation results show that the average speed of the road network is significantly improved by our algorithm, and thus improve the efficiency of the road network. Some extended applications of the proposed WSN system are discussed as well.

traffic flows based on the measurements from the surveillance sub-system. Various algorithms are proposed for this purpose, some typical examples follow. Papageorgiou et al. summaries some implementations on fixed-time strategies and traffic-responsive strategies for isolated strategies and coordinated strategies in [4]; In [5], Shimizu et al. propose a balance control algorithm to optimize the congestion length of the whole transportation network; in [6], Di Febbraro presents a hybrid Petri Net module to address the problem of intersection signal lights coordination.

The control sub-system controls the signal lights on the intersection. The guiding sub-system provides the real-time traffic information for the drivers to select the best route. The navigation sub-system uses satellite signal such as GPS to locate the specific vehicle, and with the help of electronic map, select the optimal route for the vehicle.

One shortage of the systems mentioned above is that the sensors can only detect the vehicles in a fixed spot. They can not track the vehicles out of the spot. Clearly, if we can monitor and measure the traffic status dynamically in real time, an efficient traffic control will be easier to realize.

With the development of microelectronic and computer technologies, the low-power-consumption, low-cost and relatively powerful wireless sensor network (WSN) technology has been applied in many areas[7-9]. However, the application of WSN in the traffic control system is rarely documented. In [10], we proposed a WSN-based system for an efficient traffic control in an isolated road intersection. This paper extends our previous work to a transportation network. A WSN-based traffic control, guiding, and navigation system is proposed to optimize the traffic in a transportation network.

The rest of this paper is organized as follows:

1. Introduction

Transportation plays an important role in our modern society. How to efficiently exploit the transportation capacity of the existing transportation infrastructure receives a lot of attention from the researchers across the world. The ntelligent Transportation System (ITS) has been proposed by many researchers to solve the problem.

ITS comprises of three main sub-systems. They are surveillance sub-system, analysis and strategy sub-system and execution sub-system. The execution sub-system can be a traffic control sub-system, a vehicle guiding sub-system, or a navigation sub-system.

The surveillance sub-system measures the traffic information such as the vehicle's location, speed, number of the vehicles on the road, etc., using certain type of sensor, such as inductive loops [1] or ultrasonic sensor [2]. A new method based on video analysis is now under development [1;3].

The analysis and strategy sub-system optimizes the

Section 2 describes the structure of the proposed WSN-based traffic control system. Section 3 describes the optimization algorithm for the traffic network. The simulation results and some discussions are presented in Section 4. Finally, Section 5 concludes this paper.

2. System Structure

2.1. WSN Module

To perform the traffic control, below, we shall first have a look at the configuration of the transportation network. Then, some parameters are introduced to describe traffic information in the network. By optimizing these parameters, the proposed optimization algorithm is expected to optimize the traffic in the transportation network.

As a example of a real-life traffic network, Fig. 2 illustrates the road net of Fukuyama city [11]. On the figure some parameters such as the link length, lane numbers, and legal speed are marked on it.

Figure 1. Module structure of a WSN node used in this

paper

WSN module is a basic component in our traffic control system. As illustrated in Fig. 1, a WSN module comprises of 3 main components, i.e., RF (Radio Frequency), MCU (Micro Control Unit) and Power Supply. The RF encodes, modulates and sends the signal. Also it receives, decodes and demodulates the signal as well. MCU integrates processor and memories, where the programs resides and executes. The Power Supply supplies the power to entire module.

In the proposed system, WSN modules are widely distributed on vehicles, roadsides and intersections to collect, transfer and analyze the traffic information. See section 2.3 for details.

Figure 2 . Traffic network around Fukuyama Station

(cited from [11])

2.2. Urban Traffic Network

Several different facilities are installed in the urban traffic network to perform their specific functions. For example, the Signal Lights are installed in the road intersection to directly control the vehicle through the intersection; the Variable Message Sign (VMS) is set up along the road side to help drivers to select the optimal route; the Navigation system (electronic-map and satellite-based positioning system) is installed in the vehicle for vehicle locating and navigation.

The target of an ITS is to optimize the traffic in a transportation network by controlling the signal lights in the intersections, by providing the accurate traffic information in the VMS, or by selecting the best route in the e-map.

In this paper, we consider the traffic system that contains 3 types of basic elements, i.e., intersection (N), Link (L) and Vehicle (V). An Intersection can be described by 2 parameters: 1) the phase type (the type of the vehicles on different lanes passing through the intersection simultaneously); 2) the duration of every phase. A Link can be described by 4 parameters, i.e., the link length, lane numbers (include every turning-direction), mean speed, vehicle number.󰀃A Vehicle can be described by 5 parameters. They are: 1) the location of the vehicle, 2) the vehicle velocity, 3) the origin, 4) the destination, 5) the length of the route, 6) the total time and, 7) the average speed on the route. Among these parameters, 1) some are fixed, such as the lane numbers and link length; 2) some are measured by the surveillance sub-system, such as the mean speed, the number of the vehicles on a link; 3) some are set by an optimization algorithm, such as the intersection signal light and the next link selected by a vehicle.

The vehicle velocity, direction, and the number of the vehicles are the basic variables of the whole system. It is the main task of our algorithm to optimize these parameters.

2.3. Data Collection and Transferring

As illustrated in Fig. 3, there are 3 types of WSN nodes installed in our system, i.e., the vehicle unit on the individual vehicle; the roadside unit along both sides of the road; and the intersection unit on the intersection.

directions, the intersection unit analyzes the information and makes the decision to control the signal light, or to send navigate information to the vehicle.

3. Optimization Alorithm for Traffic Network

3.1 Optimization Target

From the point view of the whole transportation network, the objective of the proposed TS is to improve the use efficiency of the network, maximize the mean speed of the whole road network, and reduce the traffic congestions and accidents. From the view of an individual driver or passenger, the objective is to arrive at the destination safely with a minimum cost. The cost may be route length, fuel used, payment for taxi, or time spent. Clearly, the minimum length from the origination to the destination is a static problem, and is out of our discussion. In this paper, we only consider the minimum-travel-time algorithm. That is, the purpose of our optimization algorithm is to minimize the travel time that a vehicle drives from the origination to the destination.

Figure 3. Intersection unit (A), roadside unit (B) and vehicle unit (C) and their distribution on the road

network

The main function of intersection unit is to receive and analyze the information from other units to control the signal light. The main function of roadside unit is to gather the information of the vehicles around, and transfer it to the intersection unit. (Roadside units are installed on the lamp posts along both sides of the road every 50~200m according to the wireless cover range.) The main function of the vehicle unit is to measure the vehicle parameters and transfer them to the roadside units. (Vehicle unit is installed in every vehicle.) The intersection unit, roadside units and vehicle units are denoted as A, B and C in Fig. 2.

Roadside units broadcast messages every second. A message includes the ID of the roadside unit and its relative location to the intersection (xB, yB). Normally, vehicle unit is in the listening state. When a vehicle comes into the broadcast range of the roadside units and receives the broadcasted message, the vehicle unit switches to the active state. According to the wireless locating method [12;13], if a vehicle unit receives messages from more than three nodes, it can calculate its location (x, y) and velocity v. After that, the vehicle unit sends the information (x, y, v) to the roadside unit nearby.

Based on the (x, y, v) from the vehicles, the roadside unit can calculate the mean speed of the vehicles in its scope. The roadside then transfers the calculated information to the intersection unit.

After receiving the messages from the four

3.2 Minimum Travel Time Optimization Algorithm

gThe travel time of a vehicle comprises the running

time on the road and the waiting time for the green light at the intersection. For the ease of discussion, the following a few denotations are defined.

Node: The intersection. It is denoted as Ni.(i=0,1,2 …) Link: the road from an intersection Ni to a successive intersection Nj. It’s denoted as Li,j. Link is one-way. Say, Li,MĮLM,i.

Total Travel Time (TTT):The total time spent while a vehicle travels from the origination to the destination along a specified route.

Link Travel Time (LTT): the time spent while a vehicle travels from a node to the other node along the link.

Link Average Velocity (LAV): the average velocity of all the running vehicles in the link.

Waiting Green-light Time (WGT): The time elapsed when a vehicle or a queue waits the right-to-go phase in the front of an intersection. The parameter of WGT includes node, incoming link, outgoing link, and the time when the vehicle reach the intersection. So it can be denoted as WGT(Node, Lin, Lout, Time).

Total Travel Length (TTL): the total route length that a vehicle traveled.

The basic idea of the optimization algorithm is that: Before we choose the next link to ride, we firstly predict the time cost of the candidate routes. The route with the minimum cost is then chosen as the best route. In order to predict the total time cost, we should know the travel time in all links to pass and the waiting time before every intersection.

Let’s see a simple situation. As shown in Fig.3, the current time is Ĳ; a vehicle C is running on link L1,4with velocity v; and the destination is N8. Then, there are two routes with the approximate length: Į: N4Æ N7Æ N8;

ȕ: N4Æ N5Æ N8.

4. Simulation Result and Discussions

To demonstrate the effect of the proposed algorithm, some simulations are conducted in the PC using the data of a real urban road network which is reported in [11].

The road network is illustrated in Fig. 2. Vehicles appear in this network in a random origination to a random destination. The incoming vehicles of the entire network are recorded every 15 minutes, which are illustrated in Fig. 5(a).

In our algorithm, the mean speed (MS) of the entire road network is calculated, which is defined as follows:

MS =

/¦TTT(V) (2) ¦TTL(V)

The total travel time of Į(TTT(Į))can be calculated

as follow:

t1 = LTT(d,v) = d/v; t2 = t1+WGT(N4, L1,4, L4,7 , Ĳ+t1); t3 = t2+LTT(L4,7, LAV(L4,7 ,t+t2)) = t2 + L4,7.Length / LAV(L4,7 ,Ĳ+t2) t4 = t3+WGT(N21, L4,7, L7,8, t+t3); t5 = t4+LTT(L7,8,LAV(L7,8,t+t4)) =TTT(Į)

(1.a) (1.b) (1.c) (1.d) (1.e)

where, V is the vehicles that reach the destination in the time period.

Fig.5 (b) presents the result, curve A indicates the optimized route. As a contrast, curve B represents the results of a randomly-chosen route among several routes with approximately equal length.

TTT(ȕ) can be calculated similarly. After that, the path with the minimum TTT is selected.

From above algorithm, we can see that TTT is related to link length,d,v and LAV(Ĳ+ t2). Link length is fixed; d and v can be detected by the method presented in section 2.3. Now, the question is how can we get LAV(Ĳ+ t2)?

In [11], The author uses legal velocity to estimate the link average velocity. In [14], the author assumes that if the link is not congested, then the velocity is a constant (say, the legal velocity), otherwise, the velocity is zero.

In fact, the average velocity of a link is also related to the number of vehicles running on it, or the congestion grade since the vehicle should keep a safe distance between each other. We can construction a function between the average velocity and the vehicle number (VN) based on surveillance. Thus, if we know the vehicle number on a link, we can get the LAV of it. Since the system know the target and previous chosen route, it can compute the vehicle number in the special link at time Ĳ+ t2 -1, i.e., VN(L,Ĳ+ t2 -1). VN(L,Ĳ+ t2)󰀃Ĭ VN(L,Ĳ+ t2 -1);

Then , we can get LAV(L, ,Ĳ+ t2). So the TTT of a special route can be calculated.

(a) Incoming vehicle number in the network in every

15 minutes

(b) Mean speed of the entire network with optimization

(curve A) and without optimization (curve B)

Figure 5. Simulation Result

The proposed WSN system can also be used for many other transportation applications to improve their efficiency. For examples: 1) a “green wave” along the route of important emergent cars will be easier to implement; 2) parking management will be smarter; 3)Electronic Toll Collection (ETC) system can be improved from multilane [15] to free lane, without any tollgate to limit the vehicle stream. Some more complicated functions, such as Asymmetric signal phase control and automatic “Tide wave” control.The WSN system can also be used as a dual communication network. t can be used for the management center to track and schedule the vehicles such as taxis, buses and freight carriers.

[3] F. Woelk, S. Gehrig, and R. Koch, “A monocular image based intersection assistant”. Intelligent Vehicles Symposium 2004 , IEEE, 2004, pp.286-291

[4] M. Papageorgiou, C. Diakaki, V. Dinopoulou, A. Kotsialos, and W. Yibing, \"Review of road traffic control strategies,\" Proceedings of the IEEE, vol. 91, no. 12, pp. 2043-2067, 2003.

[5] H. Shimizu, T. Nanba, and A. Narumi, \"Analysis of structure parameters for urban traffic networks\Proceedings of the 37th SICE Annual Conference. 1998, pp. 1031-1036. [6] A. Di Febbraro, D. Giglio, and N. Sacco, \"Urban traffic control structure based on hybrid Petri nets,\" Intelligent Transportation Systems, IEEE Transactions on, vol. 5, no. 4, pp. 224-237, 2004.

[7] Anna Hac, Wireless sensor network designs, John Wiley & Sons Ltd, 2003.

[8] Edgar H.Callaway, Wireless Sensor Networks: Archi-tectures and Protocols, CRC Press, 2004.

[9]I. F. Akyildiz, S. Weilian, Y. Sankarasubramaniam, and E. Cayirci, \"A survey on sensor networks,\" Communications Magazine, IEEE, vol. 40, no. 8, pp. 102-114, 2002.

[10]CHEN Wenjie, CHEN Lifeng, CHEN Zhanglong, and TU Shiliang, \"A Realtime Dynamic Traffic Control System Based on Wireless Sensor Network\Proceedings of the 2005 International Conference on Parallel Processing - Workshops, IEEE, 2005.

[11]H. Shimizu and H. Kita, \"Signal control algorithms of congestion length\Proceedings of the 36th SICE Annual Conference. 1997, pp. 1195-1200.

[12]W. Chin-Der and L. Ming-Hui, \"Data fusion methods for accuracy improvement in wireless location systems\Wireless Communications and Networking Conference, 1 ed 2004, pp. 471-476.

[13]Zhao Yilin, \"Mobile phone location determination and its impact on intelligent transportation systems,\" Intelligent Transportation Systems, IEEE Transactions on, vol. 1, no. 1, pp. 55-64, 2000.

[14] L. Jia and U. Ozguner, \"Integration of dynamic routing and intersection control in intelligent transportation system\Proceedings of Intelligent Transportation Systems, IEEE , 2000, pp. 137-142.

[15] A. S. Mustafa, M. Pitsiava-Latinopoulou, and G. A. Giannopoulos, \"The multilane electronic toll collection system in Thessaloniki: Evaluation of its first 6 months of operation\Proceedings of Vehicle Navigation and Information Systems Conference ,1994, pp. 699-703.

5. Conclusion

n This paper, a WSN-based architecture is Ipresented for ITS of a transportation network. With the help of WSN technology, the traffic info of the network can be accurately measured in real time. Based on this architecture, an optimization algorithm is proposed to minimize the average travel time for the I

vehicles in the network. Compared to the randomly-chosen algorithm, simulation results show that the average speed of the road network is significantly improved by our algorithm, and thus improve the efficiency of the road network. Some extended applications of the proposed WSN system is discussed as well.

6. Acknowledgment

The authors are grateful to Mr. LI Ping from Philips Electronics Singapore Pte Ltd and Mr. GU Yongfeng from University of Boston, for their insightful comments and suggestions.

References

[1] L.A. Klein, “Traffic parameter measurement technology evaluation”. Proceedings of the IEEE-IEE Vehicle Navigation and Information Systems Conference, 1993, pp.529-533

[2] T. Matsuo, Y. Kaneko, and M. Matano, “Introduction of intelligent vehicle detection sensors”. Proceedings of 1999 IEEE/IEEJ/JSAI International Conference on Intelligent Transportation Systems, 1999, pp.709-713

因篇幅问题不能全部显示，请点此查看更多更全内容

查看全文

全部栏目

传感器网络在智能交通的应用