Resumen de An Approach to Forecasting and Filtering Noise in Dynamic Systems Using LSTM Architectures
Juan Pedro Llerena, Jesús García Herrero, José Manuel Molina López
Some of the limitations of state-space models are given by the difficulty of modelling certain systems, the filters convergence time or the impossibility of modelling dependencies in the long term. Having agile and alternative methodologies that allow the modelling of complex problems but still provide solutions to the classic challenges of estimation or filtering, such as the position estimation of a mobile with noisy measurements of the same variable, are of high interest. In this work, we address the problem of position estimation of 1-D dynamic systems from a deep learning paradigm, using Long-Short Term Memory (LSTM) architectures designed to solve problems with long term temporal dependencies, in combination with other recurrent networks. A deep neuronal architecture inspired by the Encoder-Decoder language systems is implemented, remarking its limits and finding a solution capable of making position estimations of a moving object. The results are finally compared with the optimal values from the Kalman filter, obtaining comparable results in error terms.
