Visualization of indoor thermal environment on mobile devices based on augmented reality and computational fluid dynamics

Augmented reality (AR) based visualization of computational fluid dynamics (CFD) simulation on mobile devices is important for the understanding and discussion of indoor thermal environments in the design process. However, utilizing AR-based mobile device for indoor thermal environment understanding still encounters problems due to limited computational power and lack of efficient interaction methods. To improve the performance of AR-based CFD visualization on a mobile device and provide the users with intuitive interaction with indoor environment, an integrated approach based on client-server framework is established. Within the approach, a new mobile friendly data format (cfd4a) with low computational complexity is proposed for CFD simulation results representation. Server-side data pre-processing method is also introduced to reduce the computational power and time needed on the client side. Furthermore, interactive section view selection and time-step animation methods are proposed for intuitive interaction with the AR environment. Then, a prototype system is developed with Unity3D engine and Tango Tablet. Demonstration in some typical scenarios shows that the proposed method and prototype system provide an intuitive and fluent AR-based environment for interactive indoor thermal environment visualization. Comparing to the widely used vtk format, a data compression ratio of 63.4% and a loading time saving ratio of 89.3% are achieved in the performance test with the proposed method. The proposed approach also shows good flexibility and extensibility in supporting different AR devices through exposing standard web services and utilizing the widely used Unity3D engine. Source code of the developed prototype and relevant testing data are also shared through github, enabling other researchers to compare our work with theirs. It is also suggested that stability of AR-based mobile devices, new interaction methods and integration with cloud computing still need further investigation and improvement.