APF-CE: AI-Powered Policy Fusion for Secure Data Sharing of Consumer Electronics in STNs

The ongoing advancement of satellite communication technology is accelerating the deployment of consumer electronics in space TT&C networks (STNs), where secure data sharing is essential for supporting AI-powered intelligent services. However, divergent perceptions among network entities in STNs often lead to authorization deviations and policy conflicts, which compromise data dissemination efficiency and service quality while introducing security risks. In this paper, we propose APF-CE, a deep reinforcement learning-based policy fusion scheme. The framework incorporates a feedback-driven compensation policy constructed by TT&C base stations, which uses filtered user feedback to establish a data credibility-and utility-aware sharing mechanism. To further enhance adaptability, a hybrid reward mechanism is designed, modeling policy adjustment impacts through similarity differences and environmental feedback. In addition, the multi-policy fusion based on DRL resolves the attribute conflict between compensation and sharing policies, while the attribute policies can obscure user identities to achieve privacy protection. Performance analysis shows that our scheme improves data sharing efficiency and feedback fault tolerance by 9.25% and 6.73%.