Design of an Intelligent Traffic-Control System

by István Varga, Balázs Kulcsár and Péter Tamás

Through the project ‘Advanced Vehicles and Vehicle-Control Knowledge Centre’, the Hungarian National Office for Research and Technology supports the design of intelligent traffic-control systems. The main goal of the project is to interlace technological transfer by connecting universities, research centres and leading industrial partners.

Automotive technologies are gaining ground in modern road traffic-control systems, since the number of road vehicles and passengers is rapidly growing. There is a perpetual need for safety-critical traffic automation, and traffic engineering makes the dynamic or static analysis and the synthesis of automotive vehicle technologies possible. The main goal of engineering is the planning and management of traffic systems.

The project supports the development of reliable and optimal control structures for urban traffic and for motorway systems. The intelligent and cooperative set-up of actuation and its linkage to the central control system is vital for avoiding traffic jams and accidents. Moreover, environmental costs (eg pollution) can be decreased. The control architecture of systemis shown in Figure 1.

Figure 1
Figure 1: General traffic management and control structure.

One aspect of the project aims at developing a traffic control algorithm for future technology. The design of the traffic control system can be evaluated in two steps – synthesis and analysis. Several models and multiple control strategies exist, and engineers must decide between them using a priori knowledge of the real system. Previously collected information can help to choose the appropriate model, parameters, measurement and control methodologies to create the optimal solution.

In many cases, control-related variables are almost inaccessible for design unless estimation techniques are applied. In a situation like this, the approximation, computer-based estimation of the variables could be useful. Traffic simulations can be classified in several ways, including the division between microscopic, mesoscopic and macroscopic, and between continuous and discrete time approaches. The methodologies of static and dynamic analysis of traffic systems are known. Several state variables, derived from the description of the dynamic system, can be used for operational and planning aspects.

A newly emerged area is demand estimation through microscopic traffic modelling. The dynamic aspect of traffic simulation requires previously measured or estimated volumes of traffic. Since the measurement of certain variables in the dynamic description is rather costly, one tries to estimate them. For instance, the observation of constantly varying turning rates at a simple intersection is fairly costly. However, the number of turning vehicles could be applied to traffic light harmonization, or generally speaking to traffic light control.

Figure 1
Figure 2: Traffic jam.

Some time has passed since traffic issues were first addressed; in fact, traffic lights apparently existed a century ago. When they were first introduced, their purpose was to ensure the safety of people in the traffic, but as time has passed and traffic has become denser, flow-control issues have become important. Nowadays, control algorithms are extremely complicated, and take traffic-dependent light-control architectures into account. Nevertheless, complementary functions exist to augment the intelligence of such systems.

The behaviour of traffic is influenced by two main factors: the control inputs, and the disturbances incurred. The control inputs are directly related to corresponding control devices such as traffic lights and variable message signs. The manipulation of disturbance values is not possible, but in some cases they are measurable (eg demand), detectable (eg incident) or predictable over a certain time horizon. The most challenging issues relate to automatic incident detection, the modelling of uncertainties, providing a solution that offers robustness under external disturbances, the use of variable message signs in order to avoid traffic jams, and finding an optimal itinerary.

The development of an intelligent control structure ensures an optimal solution for all participants in the transportation and road traffic system.

After the first year, the most significant result of the project is the structural analysis of the references. Several comparative studies have been elaborated to create a basis for further research on estimation and optimal light-control systems. In particular, this will include:


Please contact:
Balázs Kulcsár, István Varga, SZTAKI, Hungary
Tel: +36 1 463 3089, +36 1 279 6227