Smart Antenna Allocation for Wireless Surveillance

Shiyu Chen,Zhi Chen,Lingxiang Li,Shaoqian Li
DOI: https://doi.org/10.1109/icnc47757.2020.9049654
2020-01-01
Abstract:This paper investigates the smart transmit/receive antenna allocation problem for a wireless surveillance system, in which a smart full-duplex (FD) legitimate node tries to improve its monitoring rate over a suspicious communication link via adaptively allocating part of its antennas to send jamming signals. We consider a practical surveillance scenario where the wireless channel is time-varying and the channel state information (CSI) is unavailable for the monitor. The monitor’s goal is to maximize the long-term monitoring rate by allocating its number of transmit/receive antennas in an adaptive way. However, it is challenging to obtain an optimal policy in the considered time-varying system. To that end, we formulate it as a Markov decision Process (MDP) and apply Q-Learning (QL) to learn the optimal transmit/receive antenna allocation through interactions with the environment. To further accelerate the learning process, a novelty adaptive method is also proposed, referred to as ALR-Q-Learning (ALRQL), which generates real-time learning rates according to the difference on number of symbols that the monitor decodes successfully between two adjacent time slots. Simulation results validate the efficacy of our proposed methods in terms of the robustness as well as adaptability to the changing surveillance environment.
What problem does this paper attempt to address?