Adaptive Frequency Hopping in Industrial Wireless Sensor Networks: a Decision-theoretic Framework

This paper proposes an adaptive frequency hopping (AFH) approach that allows Industrial Wireless Sensor Networks (IWSNs) to cognitively switch working channels for high transmission reliability. Assuming the communication spectrum state follows a Markov Process (MP), we build a theoretical AFH framework based on the theory of Markov Decision Process (MDP). With this decision-theoretic framework, we can achieve an AFH strategy that maximizes the expected cumulative transmission reliability over a finite horizon. Judging the high computational complexity of the proposed MDP model, we further propose a myopic AFH with reduced complexity by assuming that each channel evolves independently. Without additional computation burdens or control messages exchange between sensors, the proposed AFH strategies are centrally computed by the network manager. Simulations finally demonstrate the efficiency of the proposed AFH strategies.