Efficient geometric computation for LADAR simulation

LADAR (Laser Detection and Ranging) is widely used for reconnaissance or target detection by being mounted on various moving vehicles in the defense field. During the design and development process of a LADAR system, system simulation is typically performed to assess its performance and to provide test data for the real applications. In order to generate simulated LADAR data with a high degree of reliability and accuracy, it is required to derive the precise geometric model of the sensors and to calculate the locations where the rays (laser pulses) reflected using the geometric model. As ten thousands of laser beams are transmitted to the targets every second during the real operation, a LADAR simulator should perform a tremendous amount of geometric computations to determine the intersections between the rays and the targets. In this study, we present an attempt to develop an efficient method for such geometric computation for LADAR simulation. In the computational process, we first search for the candidate facets which possess a high possibility to intersect with a ray then determine the actual intersecting facet, and further compute the intersection. To reduce the computational time, we employ an incremental algorithm and parallel processing based on a CUDA enabled GPU. We expect that our proposed approaches will enhance LADAR simulator software to be able to run in near realtime.