Optimizing LSTM-based temperature prediction algorithm for embedded system deployment

This paper presents a flexible Industrial Internet of Things (IIoT) infrastructure model, highlighting the integration of the Long Short-Term Memory (LSTM) algorithm for predictive analysis. A key component of this model is the Concentrator, a fog-computing local hub that provides a sandbox environment for third-party developments. Within this framework, clients can collect data using mechanisms provided by Original Equipment Manufacturers (OEMs), such as AROL Closure Systems, while independently developing proprietary algorithms. This approach eliminates the need for direct interaction with OEMs. The paper explores the use of the LSTM algorithm to develop a predictor for analyzing machine temperature behavior, allowing for the anticipation of potential faults.