| Summary: Wireless Sensor Networks (WSNs) are used in a wide variety of application areas such as environmental monitoring, home automation, agricultural production, health monitoring, military surveillance, industrial automation, smart grids and smart cities, and their connectivity to the Internet can be ensured by protocol stacks developed by the IETF (Internet Engineering Task Force). The IETF 6TiSCH protocol stack is defined as an extension of the IPv6 Internet protocol to support 
low-power wireless sensor networks with efficient, secure and scalable characteristics for industrial applications. The 6TiSCH protocol has a time-slotted,channel-hopping mechanism with dynamic resource requirements. However, the frequent path changes of wireless nodes require reorganization of resources. This leads to energy consumption and high communication cost due to the additional computational overhead. 6TiSCH networks use the IPv6 Routing Protocol for Low Power and Lossy Networks (RPL) at the routing layer. RPL makes routing decisions based on objective functions. In this paper, we present a new solution to the problem of frequent path changes of nodes in the network by optimizing the RPL objective function inspired by the Q-Learning algorithm. The proposed method improves real-time application problems encountered in routing processes such as high 
end-to-end packet transmission delay and high packet losses in buffer. With the proposed algorithm, a more stable network structure is obtained by utilizing the previous states and critical metrics in the case of limited power, memory and computational capacity of the WSNs, improving the performance and reliability of the network.
Key Words: IoT, Wireless Sensor Networks, Machine Learning, Q-Learning, IETF 6TiSCH Networks, RPL Objective Functions. |
