Smart Elevator With Reconfigurable Sensor Monitoring

Elevator systems require robust operation monitoring to ensure safety, performance, and effective fault detection, particularly in environments where sensor configurations may evolve over time. Traditional approaches often rely heavily on expert knowledge and rule-based systems, limiting scalability and responsiveness. With the advent of IoT technologies, real-time data collection and intelligent fault analysis have become both feasible and effective. To address these limitations, we present ElevatorTalk-M, an IoT-based elevator operation monitoring system that introduces a novel finite state machine (FSM)-based framework capable of dynamically generating operational models for arbitrary sensor configurations. Building on the earlier ElevatorTalk platform, ElevatorTalk-M enables sensors to be added or removed transparently while maintaining accurate and consistent operational modeling, a capability that is not available in existing smart elevator solutions and is particularly beneficial for legacy elevators. Experimental results show that all door obstruction events are successfully detected with an average detection time of 0.1 seconds. We further provide a formal correctness proof of the proposed fault detection algorithm. By integrating an auxiliary vibration sensor, the system can significantly enhance detection reliability, reducing the detection error from 0.18% to 0.0054%. These results demonstrate that ElevatorTalk-M offers a scalable, interpretable, and resilient solution that reduces reliance on domain experts while improving real-time elevator operation monitoring.