A Reassessment on Friendly Jamming Efficiency

With the rapid and continuous growth of various types of wireless devices in IoT, securing the communications among heterogeneous devices becomes an emerging issue. A physical layer security scheme, called "friendly jamming", has drawn great attention recently owing to its ability to protect the confidentiality of the communication as well as to enable message authentication and access control for those already employed, unencrypted, weakly encrypted, or resource constrained devices. We notice that in a large number of cases in which friendly jamming are preferable, the transmitting signals to be protected have varying spectrum utilization at symbol level. In this paper, we rebuild secrecy capacity models and re-evaluate the jamming efficiency by taking this micro time scale non-stationary characteristic into consideration. Our reassessments reveal that jamming efficiency is greatly overestimated in the existing literature. The second part of our work further proposes a waveform design on jamming signal as a means to enhance the jamming efficiency. The basic idea is to consider both time and frequency domain structure of the transmitting signal when designing the jamming signal, making both time and frequency bandwidth largely match to each other. We discuss the implementation details for jamming common QAM and PSK modulated signals. Both simulations and proof-of-concept experiments validate the theoretical correctness of our reassessment and practical effectiveness of our method.