Energy consumption performance of a class of access protocols for mobile data networks

Abstract

— When user terminals powered by a finite bat- tery source are used for wireless communications, energy con- straints are likely to influence the choice of media access pro- tocols. We use the average number of correctly transmitted packets for a given amount of allocated energy as an appropri- ate metric. In particular, we study different versions of a wire- less access protocol operating over a mobile radio channel us- ing a finite energy source with a flat power profile. The mobile radio channel iteslf is characterized by a correlated Rayleigh fading process, the correlation in the fading process being de- pendent on the speed of the user terminal. We show that the access protocol with an Error Detect feature is energy efficient for pedestrian user speeds, whereas for vehicular speeds a Re- transmission protocol is more efficient. energy efficiency under different channel conditions. We use a stochastic model for jointly tracking the evolution of the protoco ls and the available charge. By considering a discrete-time process which tracks the protocol evolution by means of a state machine, it is possible to define a set of metrics associated with the state transitions. By appropriately defining the metrics and by studying the corresponding reward earned throughout the evolution of the process, we evaluate the energy efficiency of the protocols. Al- though the energy consumption issue addressed in this paper and in (1)-(4) originates from different perspectives (access mecha- nism versus power/error control), they lead to consistent conclu- sions about energy management strategies. In particular, persis- tence is not always a reasonable choice, and adaptive strategies which try to avoid transmission during bad channel periods yield much better energy efficiency.