On the Energy Consumption of a Hybrid Time-Division/Network-Diversity Multiple Access Protocol with Harvesting Capabilities
Ref: CISTER-TR-221002 Publication Date: 2022
On the Energy Consumption of a Hybrid Time-Division/Network-Diversity Multiple Access Protocol with Harvesting Capabilities
Ref: CISTER-TR-221002 Publication Date: 2022Abstract:
This paper deals with a class of random access protocols where collisions are no longer regarded as a complete loss of resources. On the contrary, system-induced collisions based on coordinated multi-user retransmissions can be used as diversity to resolve the original conflict. This scheme is called network diversity multiple access (NDMA). NDMA has been shown to achieve one of the highest throughput values in the literature of random access. This paper shows that in NDMA the average number of transmissions plus retransmissions (including diversity-based and backlog) required for a packet to be correctly received by the destination (i.e., the average energy consumption) is considerably lower than in ALOHA. At infinite signal-to-noise ratio (SNR) and at full traffic, this metric is shown to be asymptotically bounded by the size of the (sub)network, whereas in ALOHA it is known to be unbounded regardless of SNR conditions. In addition, this paper assumes that the requested retransmissions (or induced collisions) will also be used as a source of potential energy to be harvested by idle terminals. Therefore, the objective is to optimize both the network performance (throughput and energy consumption) as well as the average energy that can be harvested per terminal. This goal is achieved by means of multi-objective optimization tools. The results suggest that NDMA not only allows for large values of throughput and reduced energy consumption in comparison with conventional ALOHA, but also higher levels of available energy to be harvested by idle terminals. To further improve efficiency, a combination of NDMA with time-division multiple access (TDMA) is here proposed. This fusion improves throughput and energy consumption at the expense of the delay induced by deterministic scheduling.
Ramiro Robles
Mauricio LaraRecord Date: 24, Oct, 2022