Information-Theoretic Analysis and Performance Evaluation of Optimal Demappers for Multi-Layer Broadcast Systems

Multi-layer broadcast systems distribute services across time and frequency domain by means of power-division multiplexing. Successive interference cancellation is required, in general, in order to extract the content of all services. For a low-complexity implementation, the receiver can obtain the strongest (top-layer) signal assuming underlying signals to behave like thermal noise. The thermal noise assumption may not be valid under certain conditions and a more accurate characterization of the interference could bring improved performance. This paper analyzes the validity of the noise-like assumption considering the power ratio between signals and the required carrier-to-noise ratio for error-free reception. The main contribution of the paper is the proposal of a demapping algorithm that exploits the knowledge of the constellation of underlying signals. Generalized mutual information, performance evaluation, and complexity analysis are provided with the additive white Gaussian noise-like assumptions and with the proposed alternative in order to assess the potential performance improvements that can be achieved.