|Summary||Develop a method for visualizing simulations involving uncertainty in 3D.|
|Author||Henry Andersson, Amadeusz Pawlik|
|Supervisor||Adam Duracz, Walid Taha|
Three dimensional models are an indispensable tool for visualising simulation results, particularly for models comprising a large number of variables. Due to the close correspondence between the 3D model and the system under design, the mapping of simulation results to 3D animations normally is straightforward. When the model includes uncertain parameter values, or the computational method used to simulate the model produces such values (e.g. simulation based on validated numerics), the mapping to a 3D animation becomes a challenge. The goal of this project is to develop a method to visualize results of such simulations.
- How can geometric objects defined by uncertain real-valued variables (intervals) be visualized effectively?
- How can such objects be visualized in a fashion that exactly captures the visualized data?
- How can uncertain discrete variables, such as strings, be visualized?
- A method for mapping simulation results involving uncertain values to 3D animations, based on sampling.
- A method for mapping simulation results involving uncertain values to 3D animations, based on box hull operations (if there is time).
- An implementation, integrated into the Acumen language.
- Firm grasp of Java, Scala or a similar functional or object oriented language.
- Experience with geometry, linear algebra and 3D visualization will all be helpful.
- Yingfu Zeng, Lightweight Three-Dimensional Visualization for Hybrid Systems Simulation, Masters thesis in Embedded and Intelligent Systems, Halmstad University, 2012. Archive includes thesis document, presentation, and Acumen models and snapshot of a Acumen pdf.
- Modeling Basic Aspects of Cyber-Physical Systems. DSLRob'12, pdf.
- Tucker, Warwick. "Validated numerics for pedestrians." European Congress of Mathematics, Eur. Math. Soc., Zrich, Austria. 2005 pdf.