Abstract: The disclosed system provides advanced mission control predictive systems for low earth orbit semi-autonomous satellites. In operation, the system may include a satellite transit behavior prediction module configured to train a neural network based on current and past satellite orbital transit paths so as to predict future transit paths. Knowing precise future satellite transit paths enables a ground-based satellite control system to more efficiently (and with greater accuracy) control certain operations of the satellites. For example, a ground-based satellite control system can prepare for communications to commence at a particular time. Legacy systems can only predict one or two days in advance However, using a neural network that not only takes into consideration historical transit data, but also learned details such as solar wind, cloud patterns, etc., the neural network predicts future satellite transit paths with statistically-certain accuracy leading to a much narrower cone of uncertainty.

Abstract: The disclosed system provides distributed management and control of satellite constellations. Such distributed management and control that involves a plurality of satellites, a plurality of ground stations, a mission control station, and a data center that hosts a portion of an Internet cloud. The system resolves issues that have plagued legacy attempts. Specifically the disclosed system includes optimized communication with satellites that leads to more efficient transfer of data and payloads to or from or by or between satellites and ground control systems. This is achieved through use of micro-batching of communications. In this manner, data may be processed and more effectively transmitted to and from more than one satellite and back to one or more ground stations, including coordinated transmissions of portions of a micro-batch between two or more ground stations.

Abstract: Satellite optimization management systems based on natural language input and artificial intelligence methods are provided. Conventional satellite management systems require understanding how to control and program satellites. As the number of satellites are deployed, the need to control such satellites by natural language instruction increases. The systems and methods disclosed herein include a natural language processing module configured to receive and interpret user input expressed in natural language to determine a user's intent and map it to specific tasks. In order to correctly determine a user's intent, a neural network with artificial intelligence models may be used. Such user's intent may be used to generate and execute satellite command sequences based on operational tasks derived from the user's natural language inputs. Such natural language input and derived intents can be translated into command sequences to that are applicable across an entire constellation of a large number of satellites.

Abstract: The techniques disclosed herein include formulation and communication of intent-based instructions that include predetermined desired outcomes associated with satellites and/or corresponding ground stations even under dynamically-changing conditions, that allow for desired satellite behaviors and facilitate high-performance payload delivery. An intent processor, located on each satellite and/or on each corresponding ground station, is used to generate and execute low-level command sequences in order to fulfill the semantics of the intent-based instructions. The intent-based instructions and/or the low-level command sequences relate to dynamic operation of the satellite under changing conditions, and/or dynamic operation of corresponding ground stations so as to achieve on-going desired behaviors of the satellite with respect to the corresponding ground stations.