Options
Towards Delay-Minimal Scheduling through Reinforcement Learning
Publikationstyp
Conference Paper
Date Issued
2020-02
Sprache
English
Author(s)
Institut
TORE-URI
Citation
GI/ITG KuVS Fachgespräche Machine Learning and Networking (2020)
Contribution to Conference
The rise of wireless sensor networks (WSNs) inindustrial applications imposes novel demands on existing wire-less protocols. Thedeterministic and synchronous multi-channelextension(DSME) is a recent amendment to the IEEE 802.15.4standard, which aims for highly reliable, deterministic trafficin these industrial environments. It offers TDMA-based channelaccess, where slots are allocated in a distributed manner. In thiswork, we propose a novel scheduling algorithm for DSME whichminimizes the delay in time-critical applications by employingreinforcement learning (RL) on deep neural networks (DNN).
Subjects
MLE@TUHH