Сайт Информационных Технологий


M.S.Koupriyanov, N.I. Matvienko, A.A. Kakunin, A.A. Kokhanenko

Saint Petersburg Electrotechnical University

Аннотация - В настоящем докладе рассматриваются вопросы создания симулятора интеллектуальных подвижных объектов на примере моделирования движения автомобилей на перекрестке. Использование методов мягких вычислений обеспечивает простоту создания модели автомобиля, адекватно реагирующего на изменение внешних условий, таких как изменение расстояния до светофора, сигнала светофора изменением скорости движения. В докладе содержится информация о моделях, методах и структуре симулятора.


Simulator of Intelligent Mobile Objects is intended to simulate vehicles behavior in the crossroad. Vehicles have intelligent properties. These properties allow simplifying creation of vehicle model, adequately reacting on changing of external conditions, such as changing distance to the crossroad, changing of speed and Traffic Light signal. Simulator is an independent system, consisting of the following blocks: Simulator Control Block (SCB), User Interface and Visualization System (UI&VS) interacting by TCP/IP protocol. Simulator can be used also as a part of the system, which simulates several mobile objects (vehicles) of various types on the roads, having crossroads. Manipulation with Simulator by TCP/IP protocol provides compatibility of modules and their functioning on remote computers (computers under Linux and Windows).

Advantages of Simulator:


The structure of Simulator has hierarchical property. The principles of its construction and construction of Multi-Agent structures in general are the same. Multi-Agent structure is based on the following definition: agent - is a substance which is in some environment, from which it receives data and which reflect events, occurring in environment, interpret events and execute instructions which influence on environment (thus the international organization FIPA represents essence of Multi-Agent systems).

Simulator consists of the following blocks: block based on the principles of artificial intelligence method, that is systems based on the knowledge (deliberative agent architecture); block based on the simple reaction on the external events (reactive architecture); block working separately from external events, except of its directive correction.


The crossroad in Simulator is implemented according to its standard definition. Crossroad is considered as noncontrollable crossroad of equivalent roads if the Traffic Light signal is flash-yellow or if the Traffic Light doesn't work.

Simulating is implemented in two zones: crossroad zone and adjacent zone. Inside adjacent zone the vehicle reacts on the changing of Traffic Light signal by changing of speed. Maximum vehicle speed is 16.66 m/s (60 km/h) by default, minimal vehicle speed is 0 m/s by default.

During simulating in the crossroad zone, one of the following actions is possible: movement ahead, turn to the right, and turn to the left. The speed before vehicle’s turn will be decreased, if the speed of vehicle is greater than 9 m/s. If the speed of vehicle is smaller than 9 m/s, the speed of the vehicle will be grown.

Traffic Light

The model of the Traffic Light with green, yellow or red signals is used in Simulator. The model of the Traffic Light can be constructed from three sections (3s model) or two sections (2s model).

The following Traffic Light signals are used in 3s model: green - permits movement; flash-green - permits movement and informs, that its time is finished and prohibiting signal will be switched on soon; yellow - forbids movement, except the cases when a driver cannot stop without emergency braking, and warns about forthcoming change of signals; red - forbids movement; red and yellow - forbid movement and inform about green signal will be switched on; flash-yellow - permits movement and informs about noncontrollable crossroad, warns about danger; doesn't work - permits movement and informs about noncontrollable crossroad, warns about danger.

The following Traffic Light signals are used in 2s model: green - permits movement; flash-green - permits movement and informs that its time is finished and prohibiting signal will be switched on soon; red - forbids movement; flash-red - forbids movement; doesn't work - permits movement and informs about non-controllable crossroad, warns about danger.


Vehicles have intelligent properties. These properties allow simplifying creation of vehicle model, adequately reacting on changing of external conditions, such as changing distance to the crossroad, changing of speed and Traffic Light signal.

The vehicle model includes the following blocks:

"If......then .... " and have the following format:

IF <in.variable> IS <value> [AND <in.variable> IS <value> [...]] THEN <out.variable> IS <value > [AND <out.variable> [...]]


<in.variable > - input linguistic variable.

<out.variavle > - output linguistic variable.

<value> - value of linguistic variable.

If "Position" Is "Near" And "Speed" Is "Fast" Then "Acceleration" Is "Brake"

If "Tlcolor" Is "Green" And "Position" Is "Close" And "Speed" Is "Fast" Then "Acceleration" Is "Brake"

Functioning Fuzzy Regulator can be represented by the following steps:

1) Fuzzification. Crisp input values are translated into linguistic variables, which are represented by fuzzy rules. Degrees of membership for all values are assigned.

2) Fuzzy Rule Inference. IF-THEN rules, which define the relationship between the linguistic variables, are set. The rules determine the course of action. The inference is a calculus consisting of two main steps: aggregation and consequence/composition.

The first step (aggregation) determines degree to which the complete IF-part of rule is fulfilled. MIN fuzzy operator is used to aggregate the degrees of validity of the various preconditions.

The second calculation step of each production rule (composition) uses the validity of the condition to determine the validity of consequence. In used standard MAX-MIN inference method, the consequence of rule is considered equally as true as the condition.

3) Defuzzification. The result of fuzzy inference is retranslated from a linguistic concept to a crisp output value. The Center of Gravity for Singletons does computing of crisp output.

Speed Converter based on the kinematics model is described by the equations connected input and output variables.


The purpose of this testing is to confirm asserting for vehicle behavior and also TCP/IP interaction. The important question is also the dependence of the time of one calculation step and computer performance. To estimate and to correct turning this dependency SimStep variable is used as a parameter of simulation. SimStep is the time of calculation of new position and speed in milliseconds.


It should be emphasized that all properties of Simulator are in strong correlation and can determine some perspectives in development:

Interacting Vehicles. Simulating vehicles interacting on the crossroad is possible by extension of intelligent properties of the vehicle as a complex deliberative agent. It is planning to provide creating different vehicle models.

Dynamical Modification of Fuzzy Knowledge Bases. The reason why the one of basic directions in research and developing Simulator is developing intelligent agents, is connected with desire to improve the performance of transportation systems on the coordination and the integration between vehicles. The improvement can be implemented by application of optimization techniques. It is planning to use dynamic syntactic and semantic adaptation of fuzzy rule base with proposed recursive procedure of decomposition of linguistic term by basic linguistic terms constructing linguistic pair.

Extension of Simulator up to Intelligent Transportation System. Intelligent transportation system can be considered as a system consisting of sections (road segments) which are connected by critical points (crossroads). Intelligent transportation system should collect information about characteristics of sections and critical points to decide transportation tasks by generating control actions applied to objects of transportation system (vehicles and traffic lights).



  1. Zadeh L.A. Fuzzy Sets. -Inf. Contr., 1965,8,p.338-353
  2. Fuzzy computing; Theory, Hardware and Applications / Fd. by M. Gupta and T. Yamanawa Amsterdam: North-Holland, 1989.
  3. Koupriyanov M.S., Yarygin O.N. Construction of relations and measure of similarity of fuzzy objects, Technical cybernetics, 1988, N3.
  4. Walker A., Woodridge M. Understanding the emergence of Conventions in Multi-Agent Systems. – 1995
  5. Feng Weidong, He Guoguang, Liu Bao, Study on ITS Self-organization Theory, IFAC 1999
  6. Stotsky A., Ioannou P., Roadway Controller for Automated Highway System, IFAC 1999

Site of Information Technologies
Designed by