SWaP-C Optimized Visual Sensor System for Computer Vision Assisted RPO

The growing interest in on-orbit servicing and active debris removal drives the need for cost-effective and robust space systems capable of performing autonomous proximity operations. During these missions, relative pose estimates between non-cooperative agents are required for reliable autonomous navigation. Current state-of-the-art solutions rely on cooperative targets engaged in active telemetry sharing or are limited to GPS-bounding them to low-earth-orbit. To address this issue, the University of Texas at Austin (UT-A) has developed and flight demonstrated a SWAP-C minimized visual sensor package utilizing a convolution neural network. This on-orbit mission demonstrated autonomous proximity operations only using edge-computing and optical sensors of the chaser spacecraft. Turion Space Corp (TSC) and UT-A propose a Phase I STTR to conduct research with the Department of Defense to assess the feasibility of adapting this technology for commercialization. During this research study, TSC and UT-A will identify defense needs and evaluate the feasibility using existing systems-based engineering models of the solution. The results of this study will enable solution adaptation, prototyping, and demonstration during Phase II.