English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

We investigate an energy-harvesting wireless sensor transmittinglatency-sensitive data over a fading channel. The sensor injects captured datapackets into its transmission queue and relies on ambient energy harvested fromthe environment to transmit them. We aim to find the optimal scheduling policythat decides whether or not to transmit the queue's head-of-line packet at eachtransmission opportunity such that the expected packet queuing delay isminimized given the available harvested energy. No prior knowledge of thestochastic processes that govern the channel, captured data, or harvestedenergy dynamics are assumed, thereby necessitating the use of online learningto optimize the scheduling policy. We formulate this scheduling problem as aMarkov decision process (MDP) and analyze the structural properties of itsoptimal value function. In particular, we show that it is non-decreasing andhas increasing differences in the queue backlog and that it is non-increasingand has increasing differences in the battery state. We exploit this structureto formulate a novel accelerated reinforcement learning (RL) algorithm to solvethe scheduling problem online at a much faster learning rate, while limitingthe induced computational complexity. Our experiments demonstrate that theproposed algorithm closely approximates the performance of an optimal offlinesolution that requires a priori knowledge of the channel, captured data, andharvested energy dynamics. Simultaneously, by leveraging the value function'sstructure, our approach achieves competitive performance relative to astate-of-the-art RL algorithm, at potentially orders of magnitude lowercomplexity. Finally, considerable performance gains are demonstrated over thewell-known and widely used Q-learning algorithm.

Accelerated Structure-Aware Reinforcement Learning for Delay-Sensitive Energy Harvesting Wireless Sensors