Channel Selection Using Machine Learning

The channel plays an important role in any wireless communication system. If there exists only one channel between the transmitter and the receiver, if the link fails, then the communication cannot be established. The reliability of communication can be improved by introducing multiple communication channels. Not only the number of channels but also the type of channels being used has an impact on the system. Although there have been few works done in this direction, works related to long distances have not been given importance. Further, manual channel switching is the recommended choice, but manual switching is greatly impacted by the people involved in the mechanism and may not be accurate all the time. Keeping these in view, this paper proposes a channel selection mechanism based on Wi-Fi and LoRa (Long Range) technologies. The advantage is that this mechanism takes into account both radio technologies to choose the best channel for the given conditions. Further, machine learning-based techniques are introduced to learn the best channel to use based on historical data, which helps in achieving automatic channel selection. This will be particularly useful in dynamic environments, where the channel conditions can change frequently. To validate the proposed concept, various experiments are carried out and from the experimental results, it is observed that the KNN algorithm achieves good performance.