Discrete Signaling and Treating Interference as Noise for the Gaussian Interference Channel
編碼及消息理論實驗室
黃昱智教授
The two-user Gaussian interference channel is a fundamental building block in network information theory. The signal model is given by:
編碼及消息理論實驗室
黃昱智教授
The capacity region can be achieved to within 1 bit by Han-Kobayashi scheme, which employs 1) Gaussian signaling and 2) SIC receiver for decoding common message.
We propose a novel coding scheme that involves only discrete signaling and single-user treating interference as noise receiver. We show that this scheme achieves the capacity region to within a constant gap:
