Communication And Multi-sensor Information Fusion Technology Of Autonomous Robot Soccer System

Robot Soccer Tournament System is a distributed system that is composed of multi-robot, and is a system integration of multi-subject and multi-domain. Its aim is to develop some interrelated subjects such as Distributed Artificial Intelligence, Robotics, etc. The Robot Soccer Tournament must solve real-time knowledge processing of some dispersive multi-agent via communication and cooperation one another in complicated dynamic environment.Autonomous Robot Soccer Tournament System is a typical distributed Multi-Agent Systems. Each robot is an independent agent, and is composed of vision and perceiving subsystem, wireless communication subsystem, decision-making subsystem, and electromechanics control subsystem. And the cooperation is achieved by exchange information among robots. Thus, it is a furthest defiant research item to exploit autonomous robot soccer tournament system. It has very high theoretic value and applied value.In this dissertation, wireless communication of robot soccer tournament system and vision and perceiving of autonomous soccer robot are studied. The primary studying content is follows.(1) Design and implementation of communication subsystem of semiautono-mous robot soccer system. The design requests of the subsystem are analyzed firstly. Its interfere sources are analyzed too. A single direction and broadcast wireless communication system is achieved. Its communication frequency is programmable to select. It has higher credible and Baud rate and ratio of capability and price. It is shown that the design is effective, and can satisfy all requirements of communication subsystem of semiautonomous robot soccer system.(2) Design and implementation of communication subsystem of autonomous robot soccer system. The multipath interfere of multi-robot communication is studied firstly. A communication platform of the subsystem is set up with aWLAN based on CSMA/CA, and its dialog management is set up with C/S model, and a viable dialog protocol for cooperation problem solving of Multi-robot System is put forward. The protocol provides "Catechetical Mechanism" and "Captain ID Mechanism" to ensure credible order and self-relive of the system and dynamic captain change. It is shown that the design is effective, and can satisfy real-time information exchange requirements of communication subsystem of autonomous robot soccer system.(3)The self-localization for autonomous soccer robot based on monocular vision. There is a precondition that is some color landmarks in the field have been recognized with cyber-vision. A method of depth measurement with monocular vision is used in the geometrical self-localization method. And three possible factor bring ranging error are analyzed in detail. And the corresponding resolventes have been put forward. PLSR is used in vidicon calibration for reducing the calibration error owing to the relativity among variables. It's shown that the geometrical self-localization method is effective, efficient robust.(4) Design of non-vision multi-sensor subsystem of autonomous soccer robot based on SOPC. A SOC of non-vision multi-sensor subsystem is set up with FPGA chip Cyclone? EP1C20. The system capability such as centralized manage, distributed control, extensible etc are realized. The error and interfere of multi ultrasonic sensor are analysed. And its resolventes based on CDMA are put forward. It is shown that the design is effective, and can satisfy the environment perceptive requirements of autonomous robot soccer system.(5) Cooperative map building of multi-robot of autonomous robot soccer system based on grey fusion. A sensor information express method with grey number is introduced, and a sub grid map building of single robot is introduced too. There is a precondition that the autonomous mobile robot have localized by its cyber-vision. To focus on an uncertainty variational environment, a grey fusion method of several sub grid map is designed to gain whole grid map which describes whole environment via communication cooperation among multi-robot based on grey system theory. It's shown that the whole cooperative map building method is effective, efficient robust.We make a summary of this dissertation lastly, and the more studying works of...