eduzhai > Applied Sciences > Engineering >

Runtime Adaptation in Wireless Sensor Nodes Using Structured Learning

  • Save

... pages left unread,continue reading

Document pages: 46 pages

Abstract: Markov Decision Processes (MDPs) provide important capabilities forfacilitating the dynamic adaptation and self-optimization of cyber physicalsystems at runtime. In recent years, this has primarily taken the form ofReinforcement Learning (RL) techniques that eliminate some MDP components forthe purpose of reducing computational requirements. In this work, we show thatrecent advancements in Compact MDP Models (CMMs) provide sufficient cause toquestion this trend when designing wireless sensor network nodes. In this work,a novel CMM-based approach to designing self-aware wireless sensor nodes ispresented and compared to Q-Learning, a popular RL technique. We show that acertain class of CPS nodes is not well served by RL methods, and contrast RLversus CMM methods in this context. Through both simulation and a prototypeimplementation, we demonstrate that CMM methods can provide significantlybetter runtime adaptation performance relative to Q-Learning, with comparableresource requirements.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×