elPlatform solution is a safety diagnosis AI solution for smart cities and smart factories consisting of a data management module for managing data collected by the IoT device, a device management module for remotely managing the IoT device, a big data pre-processing module for processing omitted (or missing) data, and an AI diagnosis module for predicting data volatility.
The big data pre-processing module is in charge of removing redundant data and restoring missing data.
AI prediction accuracy can be enhanced through big data pre-processing.
The LSTM-based AI algorithm provides a prediction algorithm that predicts the volatility of data obtained from the IoT device and a decision algorithm that determines whether or not the data is abnormal(ex.: fault).
The volatility prediction algorithm includes a built-in prediction algorithm that uses two algorithms
: an LSTM-based algorithm and a decision tree algorithm.
The data storage management module features the following functions: a function that receives data from the IoT device connected through diverse wire/wireless communications, a function that stores the received data in InfluxDB(open source), a data visualization function that visualizes the received data, and a big data pre-processing function that processes missing data.
The IoT device management module manages the remote IoT devices based on the open source Chirpstack network server.
It remotely sets up and registers the remote IoT devices, issues security keys, and remotely manages the IoT devices using the downlink event.