Synopsis
The paper provides a comprehensive review of the role of artificial intelligence (AI) in enabling trusted autonomous operations for future satellite systems and space exploration missions.
Key points:
- AI can enable autonomous capabilities for satellites such as on-board planning, fault detection, smart navigation, rendezvous and docking, and coordination between satellite swarms. This will reduce dependence on ground control and enable extended mission durations.
- AI also brings challenges related to verification, validation, certification, and trust especially for safety-critical applications. Methodologies are needed to provide guarantees on robustness, reliability, and integrity of learning-enabled components.
- The paper discusses various AI techniques being investigated for different satellite subsystems like guidance, navigation and control, health management, payload and mission management, and ground systems. Relevant case studies and prototypes are summarized.
- Architectures and middleware to enable flexible integration of AI components with heritage satellite platforms are outlined. The role of on-orbit learning and reconfigurable systems is also discussed.
- Technology roadmaps for inserting AI into space missions over the next few decades are presented. Recommendations are provided for maturing AI to be qualified for future operational satellite projects.
- Ethical considerations related to autonomous satellites are analyzed including controllability, traceability, transparency, and managing uncertainty. Regulatory and policy matters are also highlighted.
In summary, the rapid growth of AI capabilities presents game-changing opportunities as well as challenges for rethinking satellite system design and operations to enable smarter autonomous functionalities for enhancing performance, resilience, and scientific return from space missions.
