Fast Self-Adaptive Algorithms for Generating Hardbody Thermal Histories

To reduce the cost of new hardware development, the Missile Defense Agency (MDA) is developing fast new computational tools that enable in-the-loop hardware testing. The MDA and contractors have developed high-fidelity scene modeling tools such as the Fast Line-of-sight Imagery for Target and Exhaust-plume Signatures (FLITES) to test optical signature trackers. Currently, these high-speed tools are not capable of computing thermal histories for hardbody targets, an important requirement for infra-red signature prediction. Stellar Science and the University of New Mexico (UNM) will leverage their extensive expertise in thermal modeling, synthetic image generation, software engineering, and multi-processing to provide a method for generating hardbody thermal histories in real time. The innovative self-adaptive algorithm will enable engineers with workstations, specialized graphics processing units (GPUs), or clusters to use the same validated code base across computing platforms by automatically selecting the best optimizations for the particular architecture. The algorithm will then adjust the fidelity of the simulation to meet the real-time requirement. Within Phase I, we will implement a prototype of the self-adaptation and a prototype API that can be used to begin integration of the real-time thermal solver into scene-modeling tools.