Session

Technical Session 1: Mission Operations and Autonomy

Location

Utah State University, Logan, UT

Abstract

As satellites and spacecraft grow in number and operate farther from Earth, there is an emerging need for increased autonomy via onboard decision making that is independent of ground stations but allows for collaboration between teams of assets. Such autonomy will relieve the burden on human operators, enable faster responses to dynamic events, and reduce communications between orbital assets and ground stations. Orbit Logic’s Autonomous Planning System (APS) is flexible and customizable onboard software that enables teamed autonomy through the use of Tasking, Collection, Processing, Exploitation, and Dissemination (TCPED) pipelines onboard the satellites. Its small computational/memory footprint makes it especially suitable for small satellites: APS has been successfully demonstrated on constrained platforms such as the Raspberry Pi and the Unibap e2100. While APS is employed to create, plan and orchestrate TCPED pipelines, its flexible architecture allows it to interface with other satellite or software components that can provide states or events to inform or trigger planning, and to integrate with satellite resources that can execute those plans. For example, in an Earth-imaging satellite mission, APS tasks the satellite to perform collections, facilitates delivery of the collected data to onboard processing/analysis modules, and uses the results to inform future tasking, e.g., following-up with additional collection or processing. APS on a given asset employs one or more Specialized Autonomous Planning Agents (SAPAs), software modules that plan onboard activities for a specialized need. Through configurable plugins, they can be customized to the capabilities and mission roles of the host asset. Each SAPA is dedicated to a general mission-or system-level need (e.g., separate SAPAs may focus on collection planning, contact scheduling, and fault management) and issue one or more high-level activities to fulfill that need. These activities are fielded by the Master Autonomous Planning Agent (MAPA), which performs intelligent deconfliction of the onboard resources that activity execution requires. The resource execution timeline is composed to maximize the “goodness” of all competing activities using a configurable multi-factor figure of merit (FOM). APS’s modular architecture and well-defined interfaces facilitate rapid development and deployment of novel or enhanced capabilities. The level of autonomy is customizable and can be tuned over the course of the mission to allow the satellite more autonomy as it gains trust. These features allow APS to be easily deployed for complex satellite missions with multiple competing mission objectives. APS’s constellation-level collaborative autonomy seamlessly extends its asset-level autonomy. Multiple APS-enabled satellites equipped with inter-satellite links or access to a space network can coordinate without ground station communications, e.g., a constellation of imaging satellites can perform load balancing among themselves to ensure coverage and limit redundancy. Such autonomous collaboration is especially important in scenarios where evolving conditions change mission parameters, e.g., if one satellite collects imagery from a region, and processing of that imagery identifies signatures warranting follow-up tasking, a different satellite overflying the location in the near future can perform the collection. APS has been developed and extended for multi-domain, multi-asset mission applications through multiple programs sponsored by AFRL, DARPA, NASA, and ONR.

Available for download on Saturday, August 07, 2021

Share

COinS
 
Aug 9th, 10:00 AM

Autonomous Planning System (APS) for an Onboard TCPED Pipeline

Utah State University, Logan, UT

As satellites and spacecraft grow in number and operate farther from Earth, there is an emerging need for increased autonomy via onboard decision making that is independent of ground stations but allows for collaboration between teams of assets. Such autonomy will relieve the burden on human operators, enable faster responses to dynamic events, and reduce communications between orbital assets and ground stations. Orbit Logic’s Autonomous Planning System (APS) is flexible and customizable onboard software that enables teamed autonomy through the use of Tasking, Collection, Processing, Exploitation, and Dissemination (TCPED) pipelines onboard the satellites. Its small computational/memory footprint makes it especially suitable for small satellites: APS has been successfully demonstrated on constrained platforms such as the Raspberry Pi and the Unibap e2100. While APS is employed to create, plan and orchestrate TCPED pipelines, its flexible architecture allows it to interface with other satellite or software components that can provide states or events to inform or trigger planning, and to integrate with satellite resources that can execute those plans. For example, in an Earth-imaging satellite mission, APS tasks the satellite to perform collections, facilitates delivery of the collected data to onboard processing/analysis modules, and uses the results to inform future tasking, e.g., following-up with additional collection or processing. APS on a given asset employs one or more Specialized Autonomous Planning Agents (SAPAs), software modules that plan onboard activities for a specialized need. Through configurable plugins, they can be customized to the capabilities and mission roles of the host asset. Each SAPA is dedicated to a general mission-or system-level need (e.g., separate SAPAs may focus on collection planning, contact scheduling, and fault management) and issue one or more high-level activities to fulfill that need. These activities are fielded by the Master Autonomous Planning Agent (MAPA), which performs intelligent deconfliction of the onboard resources that activity execution requires. The resource execution timeline is composed to maximize the “goodness” of all competing activities using a configurable multi-factor figure of merit (FOM). APS’s modular architecture and well-defined interfaces facilitate rapid development and deployment of novel or enhanced capabilities. The level of autonomy is customizable and can be tuned over the course of the mission to allow the satellite more autonomy as it gains trust. These features allow APS to be easily deployed for complex satellite missions with multiple competing mission objectives. APS’s constellation-level collaborative autonomy seamlessly extends its asset-level autonomy. Multiple APS-enabled satellites equipped with inter-satellite links or access to a space network can coordinate without ground station communications, e.g., a constellation of imaging satellites can perform load balancing among themselves to ensure coverage and limit redundancy. Such autonomous collaboration is especially important in scenarios where evolving conditions change mission parameters, e.g., if one satellite collects imagery from a region, and processing of that imagery identifies signatures warranting follow-up tasking, a different satellite overflying the location in the near future can perform the collection. APS has been developed and extended for multi-domain, multi-asset mission applications through multiple programs sponsored by AFRL, DARPA, NASA, and ONR.