"Machine learning algorithms for Internet of Things – IoT network communication optimization"

Taking advantage of Machine Learning (ML) capabilities into the Internet of Things (IoT) technology, it is estimated that future IoT systems can be significantly improved. The main goal of this PhD thesis is studying, developing and applying ML algorithms into IoT systems, in order to archive Machine to Machine (M2M) communication optimization. It is mandatory to ensure high Quality of Service for all the connected devices, taking into consideration the hardware limitations that IoT devices introduce, as well as the network complexity which is increasing as the number of devices rises. IoT devices usually have low specifications in terms of CPU power, memory and storage. Moreover, due to the great number of devices in an IoT infrastructure, the device heterogeneity is significantly growing, introducing a possible impact in the system performance, which should be studied extensively. Finally, it is essential to study the correlation between computationally intense ML techniques and IoT systems energy consumption.