Lid Opening Detection in Manholes using RNN
Uros Andrijasevic, Jelena Kocić, Vladan Nešić · 2020
In this paper, the machine learning algorithm for the lid opening detection in telecommunication manholes is presented. Telecommunication manholes network is equipped with smart Internet of Things (IoT) devices that use multiple sensors to monitor the system and to detect various states of a manhole, like a lid opening, detection of the light, detection of the object or person presence, etc. The machine learning algorithm that detects state lid open or lid closed from the data that the IoT manhole monitoring system generates is developed with a recurrent neural network (RNN), more precisely long short-term memory (LSTM). The proposed solution lids to reduced number field interventions, which consequently result in lower operational cost of the manhole monitoring system and increase the system effectiveness.