AI-DRIVEN SECURITY FRAMEWORK FOR ENHANCED THREAT DETECTION IN MOBILE SATELLITE NETWORKS

The increasing reliance on Mobile Satellite Networks (MSNs) for secure and reliable global communication has led to heightened concerns over cybersecurity threats. Traditional security mechanisms often struggle to counter adaptive and sophisticated attacks, necessitating the integration of Artificial Intelligence (AI)-driven security frameworks. The dynamic nature of MSNs, characterized by high latency, intermittent connectivity, and diverse attack vectors, presents unique security challenges. A key challenge is the real-time detection and mitigation of cyber threats, including eavesdropping, jamming, spoofing, and denial-of-service (DoS) attacks. Conventional cryptographic techniques and firewall-based security solutions are inadequate against evolving threats, necessitating intelligent intrusion detection and adaptive defense mechanisms. To address these challenges, an AI-enhanced security framework is proposed, incorporating Deep Learning (DL) and Reinforcement Learning (RL) models for threat detection and response optimization. The framework employs a Hybrid CNN-LSTM model for anomaly detection, achieving an accuracy of 98.7% in detecting intrusion attempts. Furthermore, a Q-learning-based adaptive security policy dynamically adjusts encryption levels and resource allocation to mitigate ongoing attacks, reducing response time by 37.5% compared to traditional methods. The proposed approach was validated using the NSL-KDD dataset and realworld satellite telemetry logs, demonstrating a 45.3% improvement in threat mitigation efficiency over conventional rule-based systems.