Author Topic: A Critical Review on the Development of Urban Traffic Models & Control Systems  (Read 3090 times)

0 Members and 1 Guest are viewing this topic.

IJSER Content Writer

  • Sr. Member
  • ****
  • Posts: 327
  • Karma: +0/-1
    • View Profile
Author : Rishi Asthana, Neelu Jyoti Ahuja, Manuj Darbari, Praveen Kumar Shukla
International Journal of Scientific & Engineering Research Volume 3, Issue 1, January-2012
ISSN 2229-5518
Download Full Paper : PDF

Abstract— Modeling and development of control systems to deal with the congestion at intersection in urban traffic is a critical research issue. Several approaches have been used to develop the modeling and controlling phenomenon in the said problem. These approaches include, Petri net, Fuzzy Logic, Neural Network, Genetic Algorithms, Activity Theory, Multi Agent Systems and many more. This paper is a survey on the development of Urban Traffic Control Systems using techniques discussed above in the last decade.   

Index Terms— Fuzzy Logic, Neural Network, Genetic Algorithms, Multi Agent Systems, Activity Theory, Petri Nets.

1   INTRODUCTION                                                                      
The development of control systems to deal with the con-gestion at the intersection in urban traffic is a critical re-search issue. The prime requirements of the developed system are, the signal must not allow the ambiguous move-ment to the traffic and it must be clear that how/when the indication of signal shown to be changed. Two other aspects to be handled are to take decisions about signal indication sequence in the control system to make the system well opti-mized and development of control logic for signal generation. 
This paper has been divided into 6 subsections. In section 2, Petri net based modeling has been studied and revised. Section 3 contains the review of multi-agent systems in urban traffic control systems. Neural Network based approaches are discussed in Section 4. Section 5 contains the fuzzy logic based approaches in Traffic Control Systems. Several hybrid approaches of fuzzy logic, neural network, petri nets are discussed in section 6. Section 7 contains various other approaches for the traffic control systems, like activity theory, complex network theory, incident and real time traffic control etc.
Petri Nets [1] are also known as a place/Transition Net or P/T net. It is the mathematical modeling language for the description of Discrete Event Systems (DES). PN theory is developed in 1962 by Carie Asam Petri. These are highly applicable in graphical modeling, Mathematical modeling, simulation and real time control by the use of places and transitions.

Different variations of the Petri Nets are applied in the modeling and control of traffic systems.
A Colored Timed Petri net (CTPN) model has been used for validating a Urban Traffic Network in [2].
A model for real time control of urban traffic networks is proposed in [3]. A modular framework based on first order hybrid Petri nets model is developed. The vehicle flows by a first order fluid approximation, in this approach. The lane interruptions and the signal timing plan controlling the area are developed by the discrete event dynamics integrated with timed Petri nets.

A new hybrid Petri net model for modeling the traffic behavior at intersection is developed in [4]. The important aspects of the flow dynamics in urban networks are interpreted very well.
A new approach of continuous Petri nets with variable speed (VCPN) is proposed in [5]. The analysis and control design in urban and interurban networks is done.

A network model via hybrid Petri nets [6] is used to dem-onstrate and implement the solution of the problem of coordinating several traffic lights. It aims the improvement in the performance of some classes of special vehicles, like public and emergency vehicles.

A model of TCPN (Timed Control Petri Nets) is used to demonstrate and solve the problem of coordinating sever-al traffic lights in [7]. The analysis of the control TCPN mod-els is done by Occurrence Graphs (OG) techniques.

A Colored Petri Net Model of an urban traffic network for the purpose of performance evaluation is demonstrated in [8]. The subnets for the network, the intersections, the external traffic inputs and control are discussed.

A Urban Traffic Simulation has been done using petri net in [9]. This approach is based on generating producer consumer network and grid simulation of petri nets.

A multi-agent system (MAS) [10] is a system consists of multiple interacting intelligent agents. Multi-agent systems can be used to solve problems that are possible to be difficult or impossible for an individual agent or a monolithic sys-tem to solve. Intelligence may include few methodic, functional, procedur-al and algorithmic searching, finding and processing techniques.
A multi agent system approach to develop distributed un-supervised traffic responsive signal control models, has been developed in [11]. Each agent in the system is a lo-cal traffic signal controller for one intersection in the traffic network. The first multi agent system is identified using hybrid soft computing techniques. Each agent employs a multistage online learning process to update and adapt its knowledge base and decision-making procedure. The second multi agent system is produced by integrating the simultaneous perturbation stochastic approximation theorem in fuzzy extended neural networks (NN).

An approach to model the traffic of an important crossroad in Mashhad city using intelligent elements in a multi-agent environment and a large amount of real data, has been developed in [12]. The overall traffic behavior at the intersection was first modeled by the Bayesian networks structures. Also, the probabilistic causal networks are used to model the effective factors.
Among the several ITS applications is the notion of Dynamic Traffic Routing (DTR), which involves generating “optimal” routing recommendations to drivers with the aim of maximizing network utilizing. In [13], it has been presented that the feasibility of using a self-learning intelligent agent to solve the DTR problem to achieve traffic user equilibrium in a transportation network. The agent then learns by itself by interacting with the simulation model. Once the agent reaches a satisfactory level of performance, it can then be deployed to the real-world, where it would continue to learn how to refine its control policies over time.

The integration of cooperative, distributed multi-agent system to improve urban traffic control system is proposed in [14]. Real-time control over the urban traffic network is done through an agent-based distributed hierarchy traffic control system. This system cooperates with dynamic route guidance system. Cooperative system framework and agent structure are discussed in this work.

A new framework of hybrid control system for UTC is presented in [15], in which any optimal control strategy can be adopted. By the interface D-S and interface C-S namely cooperation model, the hybrid system of UTC is divided into three layers including digital control loop, discrete event module, and Group Decision-making Support System (GDSS). By integrating GDSS consisted of central agent and intersection agents, real time control and coordinate control with the characteristics of self-decision, cooperation, and intelligence are implemented.
An approach of modeling the urban traffic flow system is discovered for combining the global and local model information for the whole city net in [16]. It is assumed that traffic digraph consists of several nodes and those nodes are linked with routes lines. The proposed system uses the random walk theory. Vehicle flow density and driver strategy independence are also the important factors in this approach.

An agent-based approach to model the individual driver behavior under the influence of real-time traffic information is proposed in [17]. The driver behaviour models developed in this work are based on a behavioural survey of drivers. This survey was conducted on a congested commuting corridor in Brisbane, Australia. Based on the results obtained from the behavioural survey, the agent behaviour parameters which define driver characteristics, knowledge and preferences were identified and their values determined.

Read More: Click here...