IEEE-International Conference on Recent Trends in Information Technology, ICRTIT 2011 MIT, Anna University, Chennai.

June 3-5, 2011

Dynamic Road Traffic Management based on Krushkals Algorithm

Mr.V.Shashikiran#1, T.T Sampath Kumar#2,N.Sathish Kumar #3, V.Venkateswaran#4, S Balaji#5
1 2, 4, 5

Assistant Professor, Department Of Computer Science, Sri Venkateswara College of Engineering, Chennai, India M.tech (pursuing), Department of Computer Science, Sri Manakula Vinayagar Engineering College, Pondy, India
1 2

[email protected] [email protected]

Abstract The dynamic road traffic management is based on the dynamic vehicle routing during peak hour traffic. The real time traffic management has become a hectic problem in daily life due to the increasing traffic, sudden accidents, and bottle necks due to various reasons. In Dynamic Vehicle Navigation System (DVNS), the real time traffic junctions are mapped as nodes and the traffic rate between the signals is considered as the link weight for the selection of routes from source to destination. The selection of the route depends on various parameters such as traffic rate, speed of the vehicle, shortest path etc. The dynamic route selection is implemented using the krushkals algorithm based on different parameters. The end-user selects the source and destination and sends the information to our system using Google maps; our system suggests the optimal route to the enduser based on the dynamic routing table. The traffic in-charge personnel have to update the traffic rates using their mobile PDA. The Dynamic routing table is updated periodically to accomplish the optimal solution. Keywords Traffic Management, Dijkstras algorithm Krushkals algorithm,

I. INTRODUCTION Nowadays, there is an urgent need for the robust and reliable traffic assistance system to improve traffic control and management to solve the problem of increasing urban traffic. Vehicle detection technique appears to be the weakest in traffic guidance and control. Many traffic state parameters can be detected through traffic guidance system, including traffic flow density, the length of queue, average traffic speed and total vehicle in fixed time interval. To achieve these goals, in past decades, there have been many approaches proposed for handling the related problems. Numerous research projects aimed to detect traffic flow have been carried out in terms of measuring traffic performance during the past decades. There are already several kinds of traffic flow detection methods such as, background subtraction method, frame subtraction method, land mark based method and edge detection method. But in our approach, the system is concerned about the dynamic route selection for the vehicles. The major challenge presented in dynamic traffic management is the optimal utilization of all the routes according to their limitation in number and capacity. In most cases, it is inconvenient to construct new routes or increase roadway capacities. So identifying ways to maximize the

utilization of the existing transportation network becomes more important. For this traffic management application, it is particularly important to establish a dynamic model which is based on dynamic network modeling techniques by considering the traffic network. This model should also have a simple structure that can be solved efficiently and quickly so that the optimal solution can be obtained soon after the disaster happens. The scope of dynamic navigation goes beyond real-time traffic information services for the public. A more important ability is to provide predictive traffic information services for a certain future period based on the processing of real-time traffic information [1]. Research on the time dependent shortest path algorithm is a hot issue in solving the dynamic navigation problem [2, 3, 4, 5, and 6]. Although there are several different algorithms in dynamic routing, traffic data organization is always one of the most important issues which affects the route speed and system performance in DVNS. In this paper, we propose a system for the dynamic vehicle routing. The dynamic route is based on various parameters such as road size, traffic across the path, speed of the vehicle which the user intended to travel with, the distance between the source and destination, peak hour traffic etc. The remainder of this paper is organized as follows. The Section II describes the architecture of the system and the Krushkals algorithm [7] which is used for the dynamic routing. Section III gives the implementation details of the dynamic vehicle navigation system. Section III describes the complete implementation details. Section IV describes the implementation issues. Section V describes the survey about the real time traffic. Finally, section VI gives conclusion and directions for future research. II. ARCHITECTURE OF THE SYSTEM In order to develop the requirements of traffic data organization and deployment, it is necessary to analyze the navigation data components and route processes of DVNS. The traffic guidance and routing is implemented using the krushkals algorithm and the route discovery is implemented using the dijkstras algorithm. A. Data Component of Vehicle Navigation System Navigation data is the foundation for a vehicle navigation system. There are several file formats for navigation maps,

978-1-4577-0590-8/11/$26.00 2011 IEEE

200

IEEE-ICRTIT 2011
including KIWI, GDF, SDAL and etc. Following the development of dynamic navigation systems, current navigation data can be divided into two parts in general. Figure 1 shows the data components in vehicle navigation systems. According to the data type, navigation data can be classified as map data and traffic data. Traffic data are composed of traffic facility and traffic information, and most of the time, traffic facility data can be found in map data. Hence, the difference between map data and traffic information data lies in the frequency of data update. Although map data updates quite frequently, traffic information has the fastest rate of change in vehicle navigation systems. In this article, we consider the map data as static in contrast to traffic information. Because of these, we limit our concern to the organization and deployment of traffic information data. C. Architecture The Dynamic Vehicle Navigation System consists of the traffic database, dynamic route generator, the traffic event updates and the end user. Figure 2 shows the architecture of the Dynamic Vehicle Navigation System.

Fig. 3 Dynamic Vehicle Navigation System Architecture

Fig. 1 Data component of vehicle navigation system

B. Route process on dynamic vehicle navigation system The Dynamic Navigation System is a dynamic traffic route generation and route assistance system. The route decision process can be divided into two phases, before-driving and ondriving, while before-driving is the first route planning. Figure 2 gives the route process of dynamic navigation system.

Fig. 2 Route process of dynamic navigation system

Both the phases are using traffic data while routing. The difference is that the first phase uses historical data to predict the driving path and the second phase uses the real time data to calculate the path in time. The requirements of the traffic database are listed below: Store traffic flow and traffic event data and build the relationship between these two data. Support quickly the query current traffic information with time and spatial attribute Use travel time attribute instead of travel speed in traffic flow data.

In this DVNS X represents set of nodes. {x1, x2,. xi } represents the individual traffic node. For each set of traffic nodes there exists a traffic rate between them. The traffic rate is represented by {r11, r12,rnn}. X = {x1, x2,, xi }, where 0< i < n, where n is the total no. of signals (nodes) in the traffic management system. R= {r11, r12,, rnn}, where r11, r12, , rnn are the traffic rates between the nodes. D= {d11, d12,,dnn}, where d11, d12,, dnn are the distance between the nodes. SP= {sp11, sp12,, spnn}, where s11,s12,, snn are the speed in which a vehicle can travel in between the nodes. W= {w11,w12, . wnn}, where s11,s12,, snn are the link weights between the nodes. For any given traffic junction xij, the traffic rate rij depends on the inflow and the outflow of the traffic from those junctions. If the outflow of any node xi is lesser than the inflow, the bottleneck arises in that specific node and this bottleneck will propagate to the nodes which are directly connected to the node xi. So there is a dependency of traffic flow between the signals in which each node is directly or indirectly dependent on the traffic flow of the other nodes. The traffic rate rij between the nodes xi and xj is dependent on the nodes which are directly linked to the xi and xj. If A and B are dependent events, then the probability of A happening AND the probability of B happening, given A, is P (A) P (B after A). Let A be the event of the increase or decrease in the traffic rate rij between the nodes xi and xj. The event of increase or decrease in the traffic rate depends on the event B increase or decrease in the traffic rates of other nodes which are directly connected. S = {Si}, Si Sj =, where S is a Collection of disjoint sets. Need to support following operations [8] MakeSet(x): S = S U {{x}} Union(Si, Sj): S = S - {Si, Sj} U {Si U Sj} FindSet(X): return Si S such that x Si

201

Dynamic Road Traffic Management based on Krushkals Algorithm

Steps in Krushkals Algorithm: Step1 : T = ; Step 2 : perform until v V MakeSet(v); Step 3: sort E by increasing edge weight w Step 4 : perform until (u,v) E (in sorted order) Step 6: if FindSet(u) FindSet(v) goto step 7 Step 7: T = T U {{u,v}}; Step 8: Union(FindSet(u), FindSet(v)); Here the u, v are two nodes from X= {x1, x2,, xi} which uses the weight W= { w11,w12, . wnn}, SP= { sp11,sp12, . spnn} and the traffic rate R= {r11,r12, ., rnn} are used to generate the map which will be shown to the traffic incharge. With the help of the map which is generated by the krushkals algorithm, the Traffic In-Charge has to update the dynamic traffic information like traffic rates, traffic diversions, alternate route when accidents, The traffic database consists of the history of the traffic data and the current traffic data in the dynamic routing table. The dynamic routing table consists of the nodes and the weight between the nodes and the distance between the nodes, the rate at which the vehicle can travel during peak hours and the non-peak hours and the traffic rates. In our implementation we assume that value 1 corresponds to low traffic rate, 3 corresponds to average traffic rate, 5 refers to high traffic rate. D. Dijkstra's algorithm This dijkstras algorithm [9] is used to generate the shortest path between any two nodes. This shortest path may be based on various parameters such as traffic rate, distance or both. Let the node at which we are starting be called the initial node. Let the distance of node Y is the distance from the initial node to Y. Dijkstra's algorithm will assign some initial distance values and will try to improve them step by step. 1. 2. 3. Assign to every node a distance value: set it to zero for our initial node and to infinity for all other nodes. Mark all nodes as unvisited. Set initial node as current. For current node, consider all its unvisited neighbors and calculate their tentative distance. For example, if current node (A) has distance of 6, and an edge connecting it with another node (B) is 2, the distance to B through A will be 6+2=8. If this distance is less than the previously recorded distance, overwrite the distance. When we are done considering all neighbors of the current node, mark it as visited. A visited node will not be checked ever again; its distance recorded now is final and minimal. If all nodes have been visited, finish. Otherwise, set the unvisited node with the smallest distance (from the initial node, considering all nodes in graph) as the next "current node" and continue from step 3. route is planned and presented to the end user. The end user can plan his route based on traffic between the places, speed and distance between the two places. III. IMPLEMENTATION The Dynamic Vehicle Navigation System is implemented using the Google maps API. The DVNS has two users, traffic in-charge and the normal end-user. The traffic in-charge can update the traffic rates. Figure 4 represents the Dynamic routing table, which has the vertex, distance between the nodes, speed of the vehicle, traffic rate as fields. Figure 5 gives the view of how the junctions are named in the Google maps. Figure 6 represents the junctions as the graphical nodes with edges connecting two nodes.

Fig. 4 Dynamic Routing table

Fig. 5 Representation of the junctions

4.

5.

The DVNS analyzes the dynamic routing table and using the Dijkstra's algorithm and the shortest path or the optimal

Fig. 6 Representation of the junctions as nodes

202

IEEE-ICRTIT 2011
In figure 7, the end user requests the shortest path from the node a to node g based on traffic. In Figure 10, The end user requests the route between node a and node g based on the distance and the DVNS proposes the new route which was updates using the krushkals algorithm.

Fig. 7 End User Request the Route from Node A to Node G based on traffic

Figure 8 represents the user interface for the traffic personnel, where he can update the traffic rates between any two junctions based on the map which is generated by the krushkals algorithm. This map shows the traffic congestion between any two nodes.

Fig. 10 End user Requests route between a and g based on distance

Figure 11 gives the representation of the junctions as the graph nodes. The Shortest path using the krushkals algorithm is represented as graph nodes. This is used by the traffic personnel for the purpose of the diverting the traffic based on the traffic rated between the nodes.

Fig. 8 Traffic In-charge updates the traffic between the nodes h and e to high.

In Figure 9, the end user request the route between the same node a and g but the DVNS system suggests the alternate route because the traffic between the nodes h and e has been increased and updated by the traffic personnel.

Fig. 11 Shortest path using krushkals algoithm as graph nodes

IV.

IMPLEMENTATION ISSUES

Fig. 9 New Route After the updation of the rate between the node h and e to high

As the system is implemented using the PDAs the performance of the system depends on the signal strength, servicng time for the end-users, update speed of the traffic rates by the traffic personnel, computational capacity of the system when the load increases and the traffic diversion has to be efficient in such a way that it should not create more congestions.

203

Dynamic Road Traffic Management based on Krushkals Algorithm

V. SURVEY OF REAL TIME TRAFFIC We have made a detailed survey on the traffic conditions across cities in daily life which is presented in Figure 12. In this survey, college students, IT employees and employees from various sectors have participated. In which most of the people who are travelling more is college students and the IT employees. On an average people are travelling nearly 40 kilometers everyday and everyone is spending up to 3 hours for their travelling. Another interesting thing from the survey is the average speed of the vehicle in which they travel is only 20 km/hour. The very interesting fact about the survey result is that most of the people like travelling for a long time.

Fig. 12 Survey on time spend on travelling

VI. CONCLUSIONS AND FUTURE WORK The Dynamic Vehicle Navigation System is implemented successfully. This provides a way for the endusers to find the route between any two places across cities at anytime. Our future work is to create the traffic management systems for the entire country. REFERENCES
[1] [2] [3] [4] [5] Michalopoulos, P.G., Vehicle detection video through image processing: the Autoscope system, IEEE Transactions on Vehicular Technology, vol. 40, pp. 21-29, Feb. 1991. Y. Liu, J. Zheng, L. Yan and Y. Xu, "Study on the real time navigation data model for dynamic navigation, in Proc. IGARSS'05, 2005, pp. 784-787. D. E. Kaufman and R. L. Smith, "Fastest paths in time-dependent networks for intelligent vehicle-highway systems application, Journal of Intelligent Transportation Systems, vol.1, no.1, pp. 1-11, 1993. H. D. Sherali, A. G. Hobeika and S. Kangwalklai, "Time-dependent, label-constrained shortest path problems with applications, Transport Science, vol.37, no.3, pp. 278-293, Aug. 2003. Q. Wu, Hartley, J K, Al-Dabass, D, "Time-dependent stochastic shortest path(s) algorithms for a scheduled transportation network", International Journal of Simulation Systems, Science and Technology, vol 6, no.7-8, pp. 53-60, 2005. T. O. L. Zhao, H. Nagamochi, "A* algorithm for the time-dependent shortest path problem", in Proc. WAAC08, 2008, pp. 36-43. Joseph. B. Kruskal, On the Shortest Spanning Subtree of a Graph and the Traveling Salesman Problem. Proceedings of the American Mathematical Society, vol 7, no. 1, pp. 4850, 1956. M. A. Weiss, Data Structures and Algorithm Analysis in C, AddisonWesley, 1996. Dijkstra's algorithm - Wikipedia, the free encyclopedia. [Online]. Available: http://en.wikipedia.org/wiki/Dijkstra's_algorithm.

[6] [7] [8] [9]

204