To achieve this goal, AI-GOVSATCOM will pursue the following specific objectives:

[O.1 – Transforming Static Operations] Replace static, pre-assigned bandwidth allocation with a demand-driven, real-time adaptive system, capable of anticipating traffic surges and dynamically adjusting capacity.

[O.2 – Boosting Efficiency & Reducing OPEX] Ensure optimal utilization of satellite resources through forecast-informed planning, minimizing bandwidth over-provisioning and reducing operational expenditures (OPEX), without compromising service quality.

[O.3 – Proactive Decision-Making] Leverage advanced AI models, including probabilistic forecasting and Kolmogorov-Arnold Networks (KANs), to accurately predict traffic variations and inform resource allocation strategies in both GEO and LEO contexts.

[O.4 – Benchmarking & Technical Specification] Establish a comprehensive benchmark and gap analysis comparing current GOVSATCOM bandwidth practices with predictive alternatives and deliver a finalised technical specification to guide future developments.

[O.5 – Empowering Stakeholders] Provide satellite operators with actionable insights and decision-support tools to improve operational planning and enhance service reliability for governmental and humanitarian missions.

[O.6 – Secure and Ethical Data Handling] Guarantee robust data protection through security risk assessment, anonymization protocols, and the use of synthetic datasets to support privacy-compliant model development and testing.

[O.7 – Pushing AI SatCom Innovation] Promote the adoption of AI-based forecasting in the satellite domain by demonstrating performance improvements and aligning with ESA’s GOVSATCOM objectives and future architectures.

To achieve this goal, AI-GOVSATCOM will pursue the following specific objectives:

WP1000 – State-of-the-art revision, gap analysis and finalized technical specs: focuses on surveying current forecasting methods, identifying gaps, benchmarking potential solutions, and producing a detailed technical specification and verification plan for GOVSATCOM use cases.

WP2000 – Proof-of-concept for the selected technical solutions: explores and compares different forecasting models (including KANs and probabilistic approaches), quantifying their expected performance, feasibility, and compliance, and defining preliminary proof-of-concept prototypes.

WP3000 – Verification of Proof-of-Concepts: performs the technical validation of the most promising forecasting solutions, ensuring their ability to meet GOVSATCOM constraints via updated designs and compliance assessments.

WP4000 – Verification and implementation plan for the demonstrator: translates the validated solutions into a deployable, containerized solution, producing a complete implementation and verification plan for the demonstrator, including software modularity, testing, and execution strategy.

WP5000 – Implementation and Verification of compliance: involves the practical deployment of all selected solutions and demonstrator modules, running tests, analyzing results, and producing updated performance assessments, prediction consistency evaluations, and user documentation.

WP6000 – Commercial Exploitation Assessment: conducts a final technology evaluation, defines the development roadmap, and outlines steps for the commercial adoption of the forecasting system, including TRL analysis and long-term applicability.