On-Board Autonomous Planning System
The Benefits of Planning On Board
Orbit Logic’s onboard Autonomous Planning System (APS) enables less reliance on ground stations, shorter mission timelines, and low-latency responsive tasking. APS is flexible and customizable onboard software that supports both asset-level autonomy and constellation- or swarm-level collaborative autonomy.
Unlike centralized systems, APS is not constrained by the limitations of link availability, latency, and bandwidth, ensuring that satellite directives can be executed promptly and efficiently. APS's autonomous decision-making capabilities enable satellites to function independently, reducing reliance on ground-based control and minimizing the risk of delays in communication. This is particularly valuable in operational scenarios such as missions conducted far from Earth, where communication constraints are pronounced, and rapid decision-making is crucial.
APS manages the planning, scheduling, and execution for complex missions with inter-related tasks, such as the steps in a Tasking, Collection, Processing, Exploitation, and Dissemination (TCPED) pipeline. APS can use inter-asset communication links to collaborate and spread tasking across the swarm or constellation, even when those links have low availability or capacity.
Configurable for any Mission
APS has a decentralized software architecture. An instance of APS can run onboard each asset in the heterogeneous swarm. APS models “resources” on board to predict how various activities would impact the actual resources of the spacecraft. On a given asset, APS employs one or more Specialized Autonomous Planning Agents (SAPAs), software modules that plan onboard activities for specialized missions or needs, depending on the host asset’s capabilities and constraints. Each SAPA is dedicated to a general mission- or systemlevel need 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 asset-wide deconfliction of the onboard resources that activity execution requires. The predicted results are used by APS to plan activities in a way that best utilizes system resources.
The MAPA/SAPA onboard architecture offers the flexibility to plan for different kinds of opportunities, keeps the system modular and efficient enough to be used in constrained computing environments, and makes the system extensible to almost any planning domain.
APS’s vehicle interface utilizes a “plug-in” architecture. All vehicle interface components are designed to communicate directly with the vehicle system in order to relay known vehicle state to APS, carry out APS planned activities, and provide a maintenance port for updating APS binaries and configuration.
Beyond satellite deployments, APS has been applied for autonomous patrol/detect track using unmanned underwater vehicles with the US Navy and heterogeneous Lunar and Martian exploration using rovers, rotorcraft, satellites, and astronauts-in-the-loop with NASA.
Enabling assets with the ability to perform autonomous decision-making allows them to respond much more quickly to capture opportunities that might otherwise be missed. The modular APS architecture modular architecture allows planning systems to be assembled from individual planning components and quickly configured (and reconfigured as necessary) to meet initial and dynamic mission goals.
Robust Application Management
The decentralized nature of APS’s architecture means that it is composed of numerous applications and systems The “Mission Executive” component of APS efficiently manages all of these applications Specifically, the Mission Executive software manages flight-critical, application-level software on a platform running APS software. It is responsible for starting, stopping, and monitoring the APS application suite (e.g. MAPA, SAPA, other services). The Mission Executive makes use of various configurations settings to maintain flexibility over a variety of host platforms. This includes mainly settings to control from where on the file system the software applications are executed.
The Mission Executive is composed of four functional components
A Rolling Timeline for Better Plans
APS operates using a rolling timeline, constantly adding or modifying the existing spacecraft command queue as new information is received in the form of dynamic and frequently ad-hoc events. APS’s agents support both nominal and ad-hoc planning needs, including computing optimized plans for multiple user-specified objectives, as well as the determination of dynamic courses of action in response to triggering events. APS can work completely independently or in conjunction with ground station commanding and/or other APS-equipped autonomous agents.
Flight ProvenAPS is written in C++ using POSIX compliant standards to ensure compatibility on a broad range of operating systems including Windows, Linux, and VxWorks. APS’s performance, flexibility and configurability make it a cost-effective addition to space missions - and one that can be achieved with low risk and minimal impact to schedule. APS raised its TRL level to 7 in June 2021 when deployed on-orbit aboard the YAM- 3 satellite. APS has been deployed to and tested on lightweight flight hardware, including the Unibap e2100, Raspberry Pi, and Beaglebone Black platforms.
APS Data Sheet
On-Board Swarm Control for Autonomy and Responsiveness (OSCAR), (45th Annual AAS Guidance, Navigation and Control Conference, Feb 2023, Breckenridge, CO) Lunar Exploration via Manned-Unmanned Teaming with Autonomous Robotic Swarms, (73rd International Astronautical Congress (IAC), Sept 2022, Paris, France) Autonomous Planning System (APS) for In-Flight Demo Mission, (Small Satellite Conference, Aug 2022, Logan UT) Onboard Autonomous Planning System Paper, (Space Ops Conference, May 2014, Pasadena, CA)