Age of information: pontualidade da informação em aplicações de monitoramento remoto

Abstract

This work proposes using the Age of Information (AoI) as a metric to evaluate how fresh is the decoded information at the destination when compared to the monitored process at the information source. This metric became relevant in the Internet of Things (IoT) context, as, in contrast to traditional metrics such as throughput and delay, it quantifies the importance of an update regarding the freshness and timeliness of its content. Timely updating a destination incurs in using the available resources in the network in the best manner to guarantee a reasonable performance. Throughout the work, two application scenarios are presented with numerical and analytical analysis, emphasizing optimization parameters in a remote monitoring network to obtain a better average AoI. The first scenario intends to evaluate the trade-off between average AoI and energy consumption, considering a source that updates a destination adopting retransmissions. In a complementary manner, the proposition of using receive diversity in addition to the diversity gain obtained by the retransmissions presented a better trade-off between the energy used in the transmissions and the average AoI at the destination. The next topic is a more complex scenario, where the average AoI is evaluated when multiple devices update a destination to a common wireless channel. Therefore, the message decoding depends not only on the large and small channel fading aspects but on the simultaneous transmission of other devices, which causes a collision, and all the transmitted messages are lost. Seeking to recover some information even when a collision happens, the Successive Interference Cancellation (SIC) technique is proposed. With this in mind, numerical and analytical results are presented to evaluate the advantages of such an extension. Additionally, closed expressions for the channel access probability that minimizes the network average AoI are presented. It is possible to infer that all the presented extensions throughout the work incur advantages considering the average AoI as a performance metric. Furthermore, the increased processing incurred by the extensions is only on the destination side, which generally has no processing constraints, thus not increasing power consumption on the message source, which often has power/processing restrictions.