Abstract
Communications in Wireless Sensor Networks (WSNs) are affected by dynamic environments, variable signal fluctuations and interference. Thus, prompt actions are necessary to achieve dependable communications and meet quality of service requirements. To this end, the reactive algorithms used in literature and standards, both centralized and distributed ones, are too slow and prone to cascading failures, instability and sub-optimality. We explore the predictive power of machine learning to better exploit the local information available in the WSN nodes and make sense of global trends. We aim at predicting the configuration values that lead to network stability. In this work, we adopt the Q-learning algorithm to train WSNs to proactively start adapting in face of changing network conditions, acting on the available transmission power levels. Our aim is to prove that smart nodes lead to better network performance with the aid of simple machine learning.