Designing Intelligent Energy Efficient Scheduling Algorithm To Support Massive IoT Communication In LoRa Networks

<p>We are about to enter a new world with sixth sense ability – “Network as a sensor -6G”. The driving force behind digital sensing abilities is IoT. Due to their capacity to work in high frequency, 6G devices have voracious energy demand. Hence there is a growing need to work on green solutions to support the underlying 6G network by making it more energy efficient. Low cost, low energy, and long-range communication capability make LoRa the most adopted and promising network for IoT devices. Since LoRaWAN uses ALOHA for multi-access of channels, collision management is an important task. Moreover, in massive IoT, due to the increased number of devices and their Adhoc transmissions, collision becomes and concern. Furthermore, in long-range communication, such as in forests, agriculture, and remote locations, the IoT devices need to be powered using a battery and cannot be attached to an energy grid. LoRaWAN originally has a star network wherein IoT devices communicated to a single gateway. Massive IoT causes increased traffic at a single gateway. To address Massive IoT issues of collision and gateway load handling, we have designed a reinforcement learning-based scheduling algorithm, a Deep Deterministic policy gradient algorithm with channel activity detection (CAD) to optimize the energy efficiency of LoRaWAN in cross-layer architecture in massive IoT with star topology. We also design a CAD-based simulator for evaluating any algorithms with channel sensing. We compare energy efficiency, packet delivery ratio, latency, and signal strength with existing state of art algorithms and prove that our proposed solution is efficient for massive IoT LoRaWAN with star topology.</p>

Name thesis:degree_name

Master of Science in Computer Science (MSCS)

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Computer Science

Year dc:date.available

2023

Author dc:creator

• Mhatre, Jui

Contributors dc:contributor

• Dr. Ahyoung Lee
• Dr. Yong Pei

• LoRa
• Deep Deterministic Policy Gradient Reinforcement Learning
• Massive IoT
• Energy Efficiency
• Latency
• Packet Delivery Ratio
• Simulator
• Digital Communications and Networking
• Power and Energy

Repository record dc:identifier

https://digitalcommons.kennesaw.edu/cs_etd/59

OAI identifier oai:identifier

oai:digitalcommons.kennesaw.edu:cs_etd-1064