Automating the Interactions among IoT Devices using Neural Networks
Javier Rojo, Daniel Flores-Martín, José Manuel García-Alonso, Juan M. Murillo, Javier Berrocal · 2020
The number of Internet of Things (IoT) devices is growing at an unstoppable pace. In most cases, the proper functioning of these devices requires human intervention. Due to this growth, people will have to configure more and more devices leading to the investment of a considerable effort and time. Nowadays, some works try to automate the user actions with IoT devices by using machine learning algorithms based on the relationships among devices or the previous human behaviour. However, these proposals do not make use of contextual information obtained from the devices themselves or they employ complex sets of prediction models. This paper proposes a neural network-based solution to predict the devices behaviour by using previous interactions and contextual information as input variables. Thanks to this, device interactions can be predicted and automated, even without having previous records, by knowing behaviours of devices of the same type in similar environment.