Autonomous Networking for Robotic Deep Space Exploration. Wyatt, E. J., Belov, K., Castillo-Rogez, J., Chien, S., Fraeman, A., Gao, J., Herzig, S., Lazio, T. J. W., Troesch, M., & Vaquero, T. In International Symposium on Artificial Intelligence, Robotics, and Automation for Space (ISAIRAS 2018), Madrid, Spain, July, 2018.
Autonomous Networking for Robotic Deep Space Exploration [pdf]Paper  abstract   bibtex   19 downloads  
Networked constellations of small spacecraft are emerging as novel ways to perform entirely new types of science observations that would not otherwise be possible [1], enable exploration of regions of high scientific value and that also could potentially be occupied by future human explorers (i.e., caves) [2], and demonstrate capabilities that will be useful for eventual human-robotic teams on the surface of the Moon or Mars [3]. In this paper, three mission concepts are presented and the resulting mission architectures are described. The first is a low radio frequency observatory involving tens of small spacecraft; the second is a multi-vehicle surface armada involving heterogeneous rovers (scouts, science rovers); and the third is a Lunar or Mars cave exploration scenario. Spacecraft networking architectures are determined by a unique combination of factors, including mission design constraints, mission objectives, autonomy capabilities, and networking capabilities. The combination of two technologies in particular, Disruption Tolerant Networking (DTN) [4, 5] and coordinated autonomy algorithms [6] can be enabling to these types of missions and are a focus for this paper. DTN can be thought of as the internet protocol for space and other critical applications where reliable and automated store-and-forward communications are needed. While particularly useful for long-haul links with large light time delays, DTN is also powerful for automating communication and maximizing throughput even when the communication delays are relatively short between the networked nodes. At the application layer, the ability to plan, replan, and coordinate autonomously among the nodes of the network can be important to achieve mission objectives, lower operations cost, and maximum data return.

Downloads: 19