High-performance computing using low SWAP devices


Large or real-time computing tasks currently require a significant sized computing resource, requiring large power supply, lots of thermal management, and accompanying large form factors. Recent trends in computing have investigated the likes of FPGA and Epiphany multi-core systems that provide for massively parallel computation in a low size-weight and power (SWAP) solution. Such require not only innovative algorithms to be implemented at high speed, but also methods for moving large amounts of data between processing nodes - this will be the focus of this study. However they offer the capability in credit card sized packages consuming only several watts of power. This project will investigate such capability both for SWAP applications on Earth (e.g. UAV image processing) and for on-orbit computation in small-satellites. As satellites such as cubesats are destined for deep space mission support of larger spacecraft they will require computation on-board for decision making and reduction in telemetry back to home. This is a unique opportunity for a student to shape what might be the future of computation in space, but with firmly grounded applications on the ground.