Development of an intelligent robotic system in construction automation.

摘要

The development of an intelligent robotic system, employing vision measurement, optimal stock allocation, and process determination and sequencing techniques, to perform automatic activities in construction is presented.; The incorporation of the characteristics of robots with the advantages of machine intelligence, supported by sensory perception, artificial intelligence, optimization technique, etc., is a promising and challenging topic. This topic emphasizes software development as a fundamental requirement in order to achieve the establishment of an intelligent robotic system. The U.S. construction industry productivity has been declining for the past two decades. In order to compete with Japan in the development of construction robotization, the automation techniques recently developed in manufacturing should be transferred to construction industry by utilizing the superiority of U.S. software development facilities. The driving momentum of this research is to develop an intelligent robotic system for construction. Based on the equipment of the Robotics Laboratory, University of Maryland, stone-cutting is selected as a test bed because of its feasibility of operation.; By integrating a vision system, a robotic system, and a host computer, the geometrical data of models and stone stock can be automatically and accurately obtained through the use of a vision measurement technique. Based on these data, an optimal stock allocation technique is used to provide a complete plan to make use of material both economically and efficiently, and determines the associated dividing sequences. By utilizing the extraction of the boundary features of the models, a process determination and sequencing technique first determines the cutting processes to form the boundary faces of the models, then optimally sequence these processes required in order to fulfill the task of modeling duplication with minimum cost.; Experimental and computational results from the applications of the three proposed techniques show that an intelligent robotic system performing construction activities is achievable and would serve as a prototype for future construction robots. It is anticipated that the principles developed in this research will be applicable to other batch processes encountered in construction.