Abstract
Condition monitoring techniques are widely used in various industries to increase efficiency and reduce maintenance costs, but it has yet to be extensively researched in elevator system. In this paper, we propose a model-based method with high accuracy and strong robustness for real-time speed monitoring in elevator system using low-cost inertial measurement unit (IMU) as the data acquisition device. The algorithm uses attitude correction, Kalman filter within overlapping sliding window, and zero-velocity update to eliminate the interference, including the low precision and random placement angle of IMU, different motion characteristics of the elevator, and slight vibration caused by the external environment. The experimental results indicate that the performance of the algorithm on the test sets can be comparable to that of classical supervised learning models and outperform the direct integration method. Subsequent experiments have verified that the proposed method can maintain high accuracy for a long time in real scenarios.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call