Space-based multi-hop networking

摘要

Many emerging applications will incorporate multiple spacecraft that form communications networks necessary to achieve coverage, latency and throughput requirements. Such networks may arise in the context of a space science mission consisting of distributed spacecraft performing multi-point sensing, or a global surveillance system serving military needs. Constellations of sensor spacecraft significantly benefit from incorporation of cross-link communications capabilities, thereby forming networks, by enabling continuous access to any/all spacecraft via a single ground contact, realtime coordinated observations, and autonomous in situ processing within a spatial neighborhood of spacecraft. Space-based networks may also be employed as relay infrastructure supporting (Earth, Moon, other planet) surface users engaged in various applications, such as space exploration. In this paper, we present an "L2 mesh" protocol for space-based sensor networks. Because of the large inter-spacecraft distances, directional antennas are used, with a single transceiver per spacecraft to achieve low cost. Orbital motion induces a dynamic albeit predictable geometry (and topology) among the spacecraft. One or more ground (base) stations are used; multiple ground stations are often required to ensure continuous network connectivity. Offered traffic patterns are general anycast to ground stations, and reverse-direction dissemination (for spacecraft commanding or "forward" relay). We present a technique that derives the link activation schedule (transmit/receive mode and communications neighbor selection) and route paths used for multi-hop relay through the network, leveraging the Florens and McEliece algorithm for tree networks. Highly efficient communications are achieved; in particular, the inherent tree structure enables accommodation of propagation delays that otherwise degrade the large delay-bandwidth links comprising space networks. An illustrative example is presented. Simulations demonstrate that the algorithm provides high throughput and low latency performance over general network configurations. An extension to the networking method is described that is traffic adaptive.