Environmentally Adaptive Framework for Resilient AI-enabled Edge Computing for Space Applications with the Xilinx Deep Learning Processing Unit (SpaceCube VESPR DPU)

This objective of this project will be to develop an environmentally adaptive framework to enable dependable, high-performance artificial-intelligence (AI) and machine-learning (ML) processing for spaceborne edge-computing applications by leveraging the Xilinx Deep Learning Processing Unit (DPU), a general-purpose convolutional neural network (CNN) accelerator, on the Xilinx Versal Adaptive Compute Acceleration Platform (ACAP). The Xilinx DPU, with its AI inference development stack, Vitis AI, provides the framework for rapidly prototyping and deploying hardware-accelerated, next-generation communication, navigation, and AI applications on the Versal ACAP architecture. The primary goals of this project are to deploy the Xilinx DPU on the Versal ACAP architecture, compare key metrics (i.e., accuracy, performance, power consumption) of various CNN models and DPU configurations, develop a reliable framework to enable the Versal ACAP system to autonomously adapt and change DPU models based on current environmental or operational conditions, and finally, to investigate DPU and reliability configurations for the SpaceCube VESPR single-board computer (SBC), the next-generation SpaceCube family card featuring the Versal ACAP architecture.

SpaceCube VESPR will be a next-generation addition to the SpaceCube family featuring the Xilinx Versal ACAP device and conforming to the CubeSat Card Specification (CS2), a preliminary architecture-design feasibility analysis is an anticipated end result for this project. The goal of the SpaceCube program is to provide substantial improvements in onboard computing capability while lowering relative power consumption and cost compared to traditional radiation-hardened single-board computers. The addition of the SpaceCube VESPR will further this goal by exponentially enhancing the capabilities of these devices for communication, navigation, and AI applications.

This is early/mid-stage (TRL 2) — the most realistic path in is NASA SBIR/STTR, which funds small businesses and research institutions to develop technology aligned with NASA's needs (equity-free, phased funding). Check whether a current SBIR/STTR solicitation topic overlaps with this project's technology area, or contact the project directly (above) to ask.

None of these are guaranteed paths for this specific project — TechPort itself doesn't have an "apply" button. Reaching out to the contact(s) above with a specific question is usually the fastest way to find out what's actually open.