Mission planning for autonomous systems integrating socio-technical-ecological risks

The ocean remains one of Earth's most under-sampled environments, creating a gap between societal demands for ocean knowledge and the limits of costly, infrequent, human-dependent ship and in-situ observations. A potential solution is robotic organizations, i.e., heterogeneous teams of sensing platforms capable of real-time, adaptive, and intelligent data collection that operate safely and efficiently with minimal human presence. This requires integrated systems-of-systems capabilities beyond single robotic platforms. Key challenges include robust collaboration and resilience during long-duration missions in remote, harsh settings, and the integration of domain knowledge about ecosystem dynamics, human legacies, and ocean infrastructure into autonomous reasoning and decision-making. This paper characterizes socio-technical-ecological mission risks associated with ocean robotic organizations and proposes a conceptual approach to mission planning for risk-aware autonomy. Central to the approach is supervisory risk control with online risk models distributed across edge and cloud computing domains, enabling safe robotic interaction under limited underwater communications. We outline how risk models inform adaptive sensing and control, and guide cooperative behaviours. These developments are decisive for achieving scalable, intelligent, and resilient ocean robotic operations and organizations while improving observation quality.

Keywords: Mission planning, Autonomous systems, Robotics, Artificial intelligence (AI), Supervisory risk control (SRC), Decision-support, Socio-technical-ecological, Risks.