With the rapid urbanization of cities, elevators have become essential for ensuring smooth vertical mobility in high-rise buildings. However, frequent mechanical or electrical faults can lead to accidents, delays, and safety concerns. This research introduces an IoT-based intelligent alert framework designed to monitor, predict, and report potential elevator malfunctions in real time. The system employs a four-layer architecture comprising sensor, edge, cloud, and application tiers. Non-invasive sensors capture vibration, sound, temperature, and motor-current data from the elevator units. Using a Principal Component Analysis (PCA)–Long Short-Term Memory (LSTM) model, the framework reduces high-dimensional data and predicts anomalies before failure occurs. Edge devices handle quick, low-latency inference, while cloud servers perform deeper analysis, visualize performance trends, and trigger alerts for maintenance teams. The proposed solution enhances reliability, minimizes downtime, and supports safer urban mobility. Future improvements may include vision-based diagnostics and federated learning for model updates across multiple buildings.

IoT, Edge Computing, LSTM, Predictive Maintenance, Smart Elevators, Urban Safety