Forecasting geodetic measurements using finite impulse response artificial neural networks

Abstract

In this paper, a method for evaluating and forecasting deformation movements present in buildings during tunneling works is described. The data used for processing is gathered from the project ’Prokop’ that has involved tunneling works under residential buildings, all mapped using a geodetic control network, or elevation network. Measurement results from this project are being used by the Finite Impulse Response (FIR) Neural Network as time series to predict future movements/deformations.