BIM based robot assisted object detection

Abstract

In the 21st Century, attributing to the rapid advancements in technology, automation is a rather prevalent area of focus. The key idea behind automation is to make a situation where simple and mundane jobs can be carried out without humans and hence, allowing this limited workforce to focus on higher skilled activities that can only be done by humans. Currently, in the context of building and construction, various types of robots are being used to achieve the push towards automation. Not only do these robots free up certain portions of the workforce to focus on higher skilled activities, they also help to boost efficiency of the process as well as ensure safety for the workers. For example, humans would no longer be required to go into certain hazardous or hard to reach areas to carry out these manual inspections. Similarly, the scope of this project would be to implement a BIM-based Navigation algorithm and a YOLOv3 Object Detection algorithm onto an Autonomous Mobile Robot (AMR) and test it for its reliability and effectiveness. In particular, it would tackle an ever prevalent issue of a dynamic working environment (ie. initially supplied goal pose is blocked by an obstacle) and implement a reactive action to this issue. In addition, it would also propose a solution which would enable multiple robots to be controlled on a single platform, as there is an increasing need for interoperability between various different types of robots in a shared workspace. The proposed algorithms have all been tested for their effectiveness in the EEE Robotics Lab.