I have been intrigued by the move towards more realistic gaming engines for a long time now. More and more game developers are shifting focus to the physics side of the game engines that are being developed. As this area is quite big, there are a lot of entry points into the world of physics-engine development.
Specifically, I will be focussing on object deformation in a game world, by representing objects using cellular automata. The idea being that representing an object as a collection of pieces (particles), will give way to more realistic deformation patterns.
Seeing as this model could become very computationally expensive, depending on the number of automata that an object will be made up of, simulations could potentially by performed in a real-time when applying General-purpose graphics processing units (using graphics processors to perform the simulations).
Thus, after developing a system to represent objects with cellular automata, incorporating all the essential Newtonian mechanics, the next focus of my research will be to apply Graphics Processing Units (GPUs) to perform simulations, comparing the results with simulations performed using Central Processing Units (CPUs).
The ultimate goal is to have a realistic system that is simulated in real-time.