On the Optimal Rate of Ad Hoc ALOHA Networks

The interest in asynchronous ALOHA protocols has recently been revived by the diffusion of low-power, long-range communication technologies proposed for Machine-to-Machine scenarios. This work focuses on ad hoc sensor networks where nodes communicate in pairs generating mutual interference, and tackles the fundamental yet still open question of how to set the transmission rate so as to maximise throughput. Elaborating on a previous stochastic geometry model, we consider a SINR decoding threshold that depends on the transmission rate, and analyse the interplay between interference and noise power to find the optimal transmission rate. Results reveal important design insights, showing that: i) the optimal rate relates to other system parameters in a rather complex manner but can be expressed by a single equation and is solvable via numeric evaluation; ii) in the interference-limited region, the optimal rate depends only on the bandwidth and the path-loss exponent; iii) this optimal rate yields close-to-optimal throughput also in the noise-limited region for standard values of the path loss coefficient; iv) using the optimal rate can significantly improve network throughput compared to the use of near-optimal rates.