A Fuzzy Logic Based System for Cloud-based Building Information Modelling Rendering Optimization in Augmented Reality

摘要

In recent years, Building Information Modelling (BIM) has become the standard for managing the lifecycle of a building. The metadata embedded to BIM models can be used onsite to enhance field worker’s view with relevant information about the task. Augmented Reality (AR) is a natural candidate for enabling onsite interactions with BIM models and data, due to its ability for overlaying digital information on top of real-world objects through a digital display. However, the complexity of BIM models and the limited hardware capabilities of AR-ready devices (e.g. head mounted and mobile devices), makes it difficult to provide reliable visualizations, decreasing considerably application’s performance. This paper presents a type-1 Fuzzy Logic System (T1FLS) that helps optimizing loading of BIM 3D models in an AR application. Experiments results show that the proposed T1FLS has an average frame per second (FPS) rate of 16.33 in the selected AR headset (Microsoft HoloLens). The FPS rate when using the proposed T1FLS is in average 2.33 times better than using a fixed batch size when loading BIM 3D objects.