CSMA-based Packet Broadcast in Radio Channels with Hidden Stations

TL;DR

This paper develops a mathematical model for CSMA broadcast networks with hidden stations, deriving closed-form performance solutions and analyzing how station density and access probability affect network throughput and synchronization.

Contribution

It introduces a novel MAC layer modeling methodology based on Markov processes and provides the first analytical steady-state performance solutions for infinite 1-D CSMA networks with hidden stations.

Findings

System enters synchronized transmissions at high access probabilities.

Maximum goodput increases with station density but is sensitive to access probability.

Analytical results validated by simulations in vehicular network scenarios.

Abstract

Carrier Sense Multiple Access (CSMA) MAC protocols are known to suffer from the hidden station (HS) problem. The complete mathematical analysis of CSMA networks with HSs is still an open problem, even for broadcast communication with a simple linear network topology. In this paper we address this challenge by introducing a MAC layer modeling methodology based on time- and space-domain Markov processes. Using this methodology we derive the closed-form solution for the steady-state performance in infinite one-dimensional (1-D) CSMA networks with HSs. The analytical results are validated by simulation and establish that: 1) under the assumption of fixed frame duration, if the conditional channel access probability at each station exceeds a certain threshold, the CSMA system enters the status of "synchronized transmissions", where a large number of adjacent stations transmit in overlap and interfere each other resulting in null system goodput. 2) The maximum system goodput of CSMA broadcast communication increases with increased station density but becomes increasingly sensitive to the conditional channel access probability. In [25] we validate the analytical results gained in this paper by simulation of a multi-lane highway scenario, and provide quantitative guidance for congestion control algorithms in vehicular networks.