On the performance of amplifier-aware dense networks: Finite block-length analysis

Abstract

In this paper, we investigate the performance of dense Poisson-point-process-based cellular networks using finite length codewords. Taking the properties of the power amplifiers (PAs) into account, we derive the outage probability, the peruser throughput and the area spectral efficiency in different conditions. Our analysis is based on some recent results on the achievable rates of finite-length codes and we investigate the effect of the codeword length/PAs properties on the system performance. Our numerical and analytical results indicate that the inefficiency of the PAs affects the performance of dense networks substantially. Also, for a given number of information nats per codeword, there is an optimal finite codeword length maximizing the throughput.

Finite Blocklength Communications in Smart Grids for Dynamic Spectrum Access and Locally Licensed Scenarios

“…In [31], [32], the authors investigate retransmission methods and interference networks in this context. Nevertheless, network level analysis is still missing in the literature, except for [33], where the authors utilize Poisson Point Process (PPP) to characterize the network deployment and interference, considering finite block codes in a cellular network context.…”

“…Valuable works have been done in [31], [32] where authors use the finite blocklength notion to analyze the throughput of different wireless networks. In [33], a model similar to ours is investigated where they also use PPP to characterize the cellular networks and evaluate the outage and throughput of the network in which a base station is connected to its nearest neighbor. However, this is not the case in our model.…”

“…However, using stochastic geometry and Poisson point process to model large wireless networks have proven to be a useful tool in solving the challenges faced by the classical methods [43]. Hence, in this paper, we assume a dense network where the position of the interferers is modeled as a Poisson point process [33]. Formally, we are dealing with a Poisson field of interferers [44] where the distribution of nodes that are causing interference follows a 2-dimensional Poisson point processΦ with density λ ą 0 (average number of nodes per m 2 ) [43] over an infinite plane [14] [45].…”