The Research Of A Multi-Agent Robot System For Central Air-Conditioning Ducts Cleaning

In recent years, the robot technology is more and more mature, its practicalapplication has extended into the field of family from the industrial field, also extendedfrom the production industry to the service occupation, along with the people concerningabout the health of breathing, central air-conditioning duct cleaning robot is valued bypeople increasingly. Therefore, how to make the robot clean the central air conditionerducts more efficiently and successfully has become an important research topic. In addition,the practical significance is great. This paper is exactly put forward under this background.Its purpose is to construct a complete ducts cleaning system to solve the problem that thecleaning robot is single function and low efficiency, and enable the robot to perform aducts cleaning task with time consuming less and more intelligent.Combining of central air-conditioning ducts’ specific features, this paper designs amulti-agent robot system for central air conditioning ducts cleaning. First of all, on thebasis of analyzing the existing multi-agent systems, and combining with the actualcharacteristics of the central air-conditioning ducts cleaning’s tasks, it designs themulti-robot system as a three levels hierarchical architecture. According to the demand ofcleaning the dusts’ task, this paper designs three different functions of the robot, andanalyzes the functional structure of the robot’s shape, capacity and the sensor, which laythe foundation for the later chapters.Secondly, with a detailed survey of the actual shape of the air conditioning ducts, byusing multilevel hierarchical thought in the region segmentation, the topology,characteristics and the knowledge utilization, combining with the2.5D description, theduct is abstracted into seven categories member, to establish the geometry of the modeland the connectivity model. This modeling approach improves the ability of adapting to themodeling of complex ducts, and reducing the amount of data, so that the duct modeling ismore concise, quick and accurate, and the system’s practicability is improved.Finally, this paper researches the multi-agent robot system’s task planning methods.Task planning is divided into regional responsibilities segmentation layer and job coordination layer, and researches designing the consultation coordinated approachesbased on the bidding mechanism, by this method, the system can effectively solvetask-level coordination between the robot planning problem to enhance the system’sefficiency and the flexibility.