Optimization of an autostereoscopic display for a driving simulator

In this paper, we present an algorithm to optimize a 3D stereoscopic display based on parallax barriers for a driving simulator. The main purpose of the simulator is to enable user studies in reproducible laboratory conditions to test and evaluate driving assistance systems. The main idea of our optimization approach is to determine by numerical analysis the best pattern for an autostereoscopic display with the best image separation for each eye, integrated into a virtual reality environment. Our implementation uses a differential evolution algorithm, which is a parallel, direct search method based on evolution strategies, because it converges fast and is inherently parallel. This allows an execution on a network of computers. The resulting algorithm allows optimizing the display and its corresponding pattern, such that a single user in the simulator environment sees a stereoscopic image without being supported by special eye-wear.