Skip to main content
main-content

About this book

The book describes an approach to the multi-agent systems (MAS) design for applications of robotic soccer in the MiroSot category. The described MAS is designed for dynamic, quickly changing environments, in which not only the actions of our MAS are observed, but also those of the opposing MAS. It actively tries to affect the environment to score goals faster than the opposing MAS. Multi-agent systems (MAS) are mostly applied in the environments in which they exist and act without an opposing system. The book also describes strategies based on a supervisor that makes decisions depending on behavior prediction of the opposing MAS and the ball movement in the working place. A sophisticated distribution of tasks was designed for each agent to cooperate in order score goals as fast as possible. Simultaneously, these agents try, by permitted means, to prevent the enemy agents from scoring goals. The approach described is an excellent guide to the constantly evolving abilities of mobile robotics, both for real-world applications, such as cooperation of multiple robots in life-saving activities, and for the steadily developing applications of mobile robots in various robotic competitions (e.g. Robocup, etc.). The book provides readers with high-level knowledge on how to design strategies and how to implement such systems, and the ideas presented enable them to further refine the approach utilizing the latest hardware and use it in new systems implementations of sophisticated intelligent engineering.

Table of Contents

Frontmatter

Chapter 1. Multi-agent Systems—Terminology and Definitions

The field of distributed computing has been in existence for such a long time as the solution of one computing problem might be divided into more processors. Originally these processors used data linked to problem solving and their fundamental goal was to facilitate their parallelism and synchronization.

Mikuláš Hajduk, Marek Sukop, Matthias Haun

Chapter 2. MAS Properties and Classification

The ability of agents to change their inner organization and to modify their interventions to environment is related to the following qualities: adaptability and flexibility. These abilities of agents depend on learning or mechanical (environmental) reasons that might affect the success of an agent at task solving.

Mikuláš Hajduk, Marek Sukop, Matthias Haun

Chapter 3. Agent Approach to Multi-agent Systems

Multi-robot systems are becoming more and more popular and useful in the field of industrial production, lifesaving operations, the Earth and the Space research military operations, agriculture, entertainment industry, etc.

Mikuláš Hajduk, Marek Sukop, Matthias Haun

Chapter 4. Multi-agent System Test Domain—Robosoccer

In 1990’s two robotic competitions were created independently. In both competitions robots compete against each other in disciplines by imitating soccer game. They were formed because of possibilities of testing designed and applied solutions within robotics and artificial intelligence in general.

Mikuláš Hajduk, Marek Sukop, Matthias Haun

Chapter 5. Multi-agent System Application in Robosoccer

SjF TUKE Robotics team which attends both European and International competitions organized by the international organization FIRA (Federation of International Robot-soccer association) has described the approach towards the problem-solving process of robot cooperating group in this chapter.

Mikuláš Hajduk, Marek Sukop, Matthias Haun

Chapter 6. Hierarchical Arrangement of the Agents in Developed MAS and Its Strategic Concept

This chapter deals with the descriptions of hierarchical arrangement of the agents in designed system and relations among agents within their strategic approach towards general strategy of system behaviour.

Mikuláš Hajduk, Marek Sukop, Matthias Haun

Chapter 7. Elementary Agents Players and Their Assignment to Robots

It is obvious that the master decides on the basic strategic positions of the robots on the field based on the combination of hierarchical arrangement and above mentioned decision-making process of the strategic action. Set of all elementary agents which were crated is in the Fig. 7.1.

Mikuláš Hajduk, Marek Sukop, Matthias Haun

Chapter 8. The Methodology of Creating Overall Strategy for Well-Developed MAS

In the stage of strategy s_highn development it was necessary to define the way the system has to behave and which agents from AH set are to be used. The behavior of the system is defined by the threshold parameters setting for switching and hystereses which determine the transitions between strategic actions s_mid and set of used elementary agents for strategy being formed.

Mikuláš Hajduk, Marek Sukop, Matthias Haun

Chapter 9. Analysis of Strategic Behavior of MAS Depending on the Basic Strategy Adjustment (Examples of Designed and Applied Strategies)

This chapter deals with the detailed description of system’s behavior for selected basic strategies chosen by the user. Settings for the individual strategies from the user’s point of view are listed at the beginning of the description of each strategy. There are sets of elementary agents described, from which the agent master chooses the elementary agents and the rules of ID assignment to each agent with reference to the sequence of assignment.

Mikuláš Hajduk, Marek Sukop, Matthias Haun

Chapter 10. The Driver and Its GUI (Graphical User Interface)

Software for the applied control of the robosoccer team was written in Borland Builder language C++6.0. The principle the whole system control is based on is created by the classes. The basic agents of the system were created by the appropriate interaction of these classes. Auxiliary classes were created for the correct system function in order to set the attributes of image, image process.

Mikuláš Hajduk, Marek Sukop, Matthias Haun

Backmatter

Additional information

Premium Partner

Neuer Inhalt
image credits