Deep Neural Network - Driven Outage Prediction in Multiuser Two-Way Relay Systems

In this paper, we explore the use of Deep Neural Networks (DNN) for outage evaluation in a cellular multiuser two-way relay system. In this system, a multi-antenna base station communicates with one of several single-antenna mobile stations through a single-antenna relay terminal, utilizing analog network coding. We incorporate beamforming at the BS along with MS selection to optimize the end-to-end signal-to-noise ratio. The exact outage expression from [1] is used in our model. Given the high complexity of the system, directly deriving a closed-form outage formula is challenging. To overcome this, we propose a DNN framework for real-time outage prediction. This approach not only validates our theoretical model but also reduces offline execution time and computational complexity.