AI-POWERED AUTONOMOUS SATELLITES FOR THREAT DETECTION AND MITIGATION (ARTES 4.0 SPL 4S 3D.038) - EXPRO+

Objective: The objective is to develop and validate an AI-powered autonomous satellite system capable of real-time threat detectionand mitigation targeting LEO constellations. This will reducereliance on ground-based monitoring and improve resilience in satellite operations.TargetedImprovements: - 50% improvement in real-time threat detection latency / delay over ground-based solutions.-60% reduction in response time using AI-initiated autonomous countermeasures.- Development of a simulator for cyber, RF and physical attacks, enabling validation of AI-based countermeasure/mitigation solutionsintroducing a novel capability in ESA MS.Description: Current satellite security solutions depend on ground-based infrastructure for threat detection and mitigation, introducing latency and operational risks. Traditional security mechanismsare static, rely on predefined protocols, and lack adaptability, making them ineffective against sophisticated, evolving threat scenarios.This activity will develop an AI-powered Autonomous Satellite Systemfor real-time cyber threat detection and mitigation. By leveraging AI-driven intrusion detection and automatedresponse mechanisms,the system will enable LEO satellites to autonomously identify, analyse, and counter cyber threats without reliance on ground intervention.Edge computing will facilitate real-time adaptive security policies, ensuring rapid response to attacks while significantlyimproving system resilience.To evaluate the concept, the activity will develop a reference cybersecurity architecture, incorporating an AI-powered intrusion detection engine, automated countermeasure execution,and simulation frameworks for cyberattack modellingand mitigation testing. The system will be deployed and validated using a FlatSat testbed, an attack simulator to model, test, andrefine simulated cyber-attack scenarios, and a simulator for the Inter-satellite link to validate real-time threat mitigation in context of satellite constellation. The simulator environmentshall be capable to analyse and validate AI-driven mitigation strategies across multiple simulated orbital assets in a constellation. Furthermore, the attack simulator model shall analyseand benchmark threat mitigation techniques for additional RF and/or physical threat scenarios.The expected outputs include a functional FlatSat testbed, an AI-driven intrusion detection engine for autonomous response to cyber and signal interference threats, and a simulationenvironment to model, test, and refine specific attack scenarios.