Bringing Intelligence to Wireless Sensor Nodes: Improving Energy Efficiency and Communication Reliability in Sensor Nodes

Wireless Sensor Networks (WSN) are becoming more dense to enable many new smarter use cases in industry, healthcare and agriculture. WiFi-based wireless sensor networks are becoming very attractive because of high-bandwidth, large coverage and low-powered sensors being cost-effective. Even though WiFi offers low power consumption, resources are still limited in wireless sensor networks and the identification of how to efficiently use the energy of a wireless sensor node has been an open research topic for years. In this paper a lightweight distributed reinforcement learning framework for wireless sensor networks is presented. This framework allows sensor nodes to control their transmit power in such a way that they still communicate reliably with minimum energy consumption which increases the network life-span.