Application of LSTM-Based GNSS Data Imputation for Structural Health Monitoring: A Case Study of a World’s Top 10 Longest Long-Span Floating Bridge in Norway

Abstract

This paper investigates an LSTM-based method to impute missing GNSS data from the Bergsøysund Bridge in Norway. Simulating 10% randomly data loss, the approach demonstrated high quality in restoring time-series displacement measurements, supporting improved structural health monitoring. Future work will address longer time gaps for the data loss, and integration with additional sensors.