Conventional methods of data collection rely on textual data and subjective interpretations of the employees. With advancements in mobile robots’ capabilities in recent years, the interest in deploying robots on construction sites is increasing since they can navigate autonomously and acquire data. The robots can gather new kinds of on-site data that can be used for various applications. Having autonomous robots integrated with BIM and GIS is more intuitive for non-expert users and is more efficient in terms of navigation and data collection.
High quality data that is collected from spatial structures, environments, and their surrounding sites play a key role in developing applications such as as-built modelling, progress monitoring, and quality control. Due to the evolving nature of construction projects, this data should be collected frequently and accurately, thereby making this essential task repetitive, and monotonous. Autonomous robots can be employed to automate this task. However, both the robotic and construction industries lack integration in terms of semantic data translation.
BIM and GIS provide the holistic building-related data for indoor structure and outdoor environment respectively. This data can be used for robot navigation and autonomous data collection. In this current master’s research project, BIM and GIS semantics are leveraged for the following:
Since there is no standard format for data exchange between the construction and robotic domains, the semantics of BIM-GIS cannot be automatically integrated by any robotic platform. Additionally, there is a lack of infrastructure to leverage building-related information for smarter, safer, and more precise robot navigation during the construction phase. As a result, the lack of semantic information from the construction is a barrier for non-expert users to deploy robotic data acquisition.
To address the above-mentioned solutions, the current master’s thesis adopts an interdisciplinary approach requiring a holistic methodology for successful implementation. As illustrated in Figure 1 (below), the problem is industry-oriented and builds on the existing body of knowledge. The knowledge is acquired through a comprehensive literature review to identify the state-of-the-art in this domain. Most of the research is carried out in the “Design Cycle” phase where many iterations between development and the artifact are made to obtain a valid result.
To enable cross-domain data exchange for robotic data collection, the semantics and geometry of the building information are extracted by Dynamo (visual programming tool). The extracted data are stored in an XML database. A Python script was developed to query the XML database to provide the building semantics to the robot. Having all the building information ready, a ROS service is then developed to provide the robot with the information of its mission. The information is fed to the robot where and when the data of the corresponding location is being collected. The building information also complements the robot’s sensor limitation for a better, safer, and more precise robot navigation.
Additionally, an optimal path planner was developed to provide the robot with the locations of the rooms and doors for both robot navigation and autonomous data collection. The optimal path planner is also able to help the robot not to navigate through the locations where construction activity is present. This improves the safety of robot deployment on construction sites. The collected data integrated with BIM-GIS is more efficient to be used for applications such as 3D reconstruction, progress monitoring, quality control, and equipment tracking, thus increasing the project’s productivity.
Autonomous robotic data collection integrated with BIM-GIS provides semantic data for further applications. This improves productivity by reducing the unproductive hours of surveyors, eliminating commute hours, and providing data of high quality. For a repetitive task of data collection, autonomous robots can increase the productivity of construction projects. Moreover, since the robotic platform is integrated with BIM, users can simply deploy robots and intuitively send the robot on an autonomous mission using BIM-based data. This brings even more productivity to construction projects.
To make data-driven decisions in construction projects, accurate and frequent data collection is needed during the construction phase. Manual data acquisition is time-consuming, inaccurate, subjective, and cost-ineffective. To address this issue, autonomous robots integrated with different payloads can be employed to provide the construction stakeholders with different and new kinds of information helping them in achieving a higher rate of productivity. In addition, robot navigation and autonomous data collection integrated with building data (namely BIM and GIS) improve the efficiency and facilitate robot deployment on construction sites. The current master’s thesis adopts an interdisciplinary approach to enable cross-domain data exchanges between construction and robotics as it develops an optimal BIM-based path planner to facilitate intuitive robotic data collection for non-expert users.
