Queries stream optimization in wireless sensor network with machine learning

The use of a wireless sensor network by multiple Internet of Things (IoT) devices results in a high volume of queries with similar attributes. To conserve limited resources such as communication and power, it's crucial to minimize redundant queries. In this paper, we propose an effective strategy that decomposes queries into recognizable patterns, reducing the number of requests sent to sensors and reducing communication messages. One of the most relevant applications today is predicting temperature, pressure, and wind speed from data collected from IoT devices. By utilizing a Long-Short-Term-Memory (LSTM) model, we can decrease the number of messages sent to sensors and identify repetitive patterns to eliminate redundant requests. This approach significantly lowers energy consumption at the sensor level, thus extending their lifespan.