Automation and Robotics in Construction XVI © 1999 by UC3M Automated Data Acquisition for On - Site Control Ronie Navon and Eitan Goldschinidt Faculty of Civil Engineering and National Building Research Institute, Technion City, 32000 Haifa, Israel. evronie tx.technionac it Abstract: The purpose of this paper is to highlight the need for automated real-time project control, as well as to present a model for such control, based on Automated Data Acquisition of (ADA). The idea behind the present development is to determine if measuring the I rovides workers, or other mobile agents on-site, at constant time intervals, using remote sensing, p the required control data. The ideas developed here can also be implemented for navigating and controlling construction robots. 1. INTRODUCTION Real-time control of on-site construction is essential to identify discrepancies between the plan and the actual performance, in order to take immediate corrective measures and to reduce to a minimum the damages caused by the deviations. The later the deviations are discovered, the more serious the potential damage is, and the more complex and costly the corrective measures must be. High quality data is needed not only for real- time control of current projects, but also to update historic databases. Such an update will enable better planning of future projects in terms of costs, schedules, manpower allocation, etc. Based on today's practice, in order to collect control data in a building project accurately and in a timely manner, one needs to employ controllers to measure the time it takes to construct each element, record the number of workers in each crew, and calculate the respective quantities. The inputs can then be calculated on the basis of this data. For the control, one must compare the measured ( actual ) inputs to the planned ones and check their effects on the cost estimate, on the schedule, on the work methods, etc. Because such data collection is very expensive and time consuming, many construction companies do not perform much control and even less so in real -time. At best they use crude control methods which are normally based on accounting data These methods are only capable of giving data quite some time after the controlled events took place. Consequently, they do not permit an analysis of the causes for deviation, nor do they enable corrective measures to be taken for the current project to reduce the damage. Sometimes project managers and/or foremen do perform some control on-site, but this is normally not done systematically, it is done at very long time intervals and, in many cases, it is based on intuitive data. Construction has no theory of process control [1]. I i Most of the control efforts to date have been made to develop cost control models [ e.g. 6]. These models do not take advantage of the emerging new technologies, such as project modeling and Automated Data Acquisition (ADA). Very limited work has been published relating to using the latter in constriction [141. The present paper demonstrates the potential of the new technologies for real-time on- site control. 2. OBJECTIVES AND METHODOLOGY The main objective of the present research is to develop a model for on-site control based on ADA. As this is a new idea, this research attempts to check its feasibility; determine time needs for information flow among the pertinent construction management functions; identify enabling technologies for ADA needed to support the real-time nature of the model; develop an ADA model, which is capable of capturing and feeding the control model with the needed data in real-time; develop a model for real-time control; and try it in a construction project. The research methodology consists of two major activities. The first involves the understanding of concepts and principles relevant to the model and its enabling technologies. The second involves the development of the models and their implementation. The research methodology involves the following steps: 1. A literature review, pertaining to control methods, and to ADA technologies. 2. Determination of the operating environment of the model. 3. Development of a conceptual control model, including the following steps: 681