SELF ORGANIZING DISTRIBUTED SATELLITE NETWORKS

OVERVIEW: 

Small satellites are allowing for easier development of distributed satellite networks: groups of spacecraft working together towards common goals that may be networked through both ground data exchange and on-orbit satellite-to-satellite data exchange. These distributed satellites systems can range from a few to hundreds of spacecraft and can be spatially close together or separated over multiple orbit planes. As these networks grow, traditional human-in-the-loop operations becomes cumbersome, if not impossible. The US government is seeking technologies to create autonomous, self-organizing distributed satellite networks; are there current or emerging capabilities in the space, or other industries that have similar capabilities that can be leveraged to solve this problem? Solutions should be capable of re-configuring the distributed satellite network autonomously based on evolving operational user defined mission set(s) and can involve new sensors, new applications of existing sensors, methods/algorithms for controlling dynamic networked systems, or methods/algorithms for planning behaviors of independent but networked nodes.

BACKGROUND:

The CubeSat standard has enabled a whole new opportunity to space access by driving node cost of a single spacecraft down significantly.  There are now many new capabilities which can be made by utilizing small spacecraft including but not limited to: shared-, simultaneous-, and self-sampling. Potential for these mission sets can only be realized by driving autonomy to the spacecraft (node) and the network.

USE CASE EXAMPLE: WHY IT’S IMPORTANT

We need satellites to do network and to do navigation/control. For example, the Amazon delivery of packages in cities through UAVs (node is aware of where it needs to go and learns surroundings to properly achieve high-level goal of reaching a home), or self-driving cars (node is aware of other nodes but operates on its own), or cellular phone networks (nodes can enter/exit various networks but it is seamless to the user).  The ability to autonomously reconfigure the network to perform different missions, the ability to maintain a configuration (control, sensing/telemetering, etc. of the nodes), and the ability to react based upon network data collection are all areas of interest

PROBLEM SPONSORS:

Lee Jasper, AFRL