Using a wavelet-decompositon of a random field, the aim is to efficiently estimate (latent) spatial models.
There is considerable research on improving navigation accuracy of unmanned (aerial) vehicles. This direction of research focuses on improving stochastic sensor calibration to integrate navigation filters and produce more accurate navigation performance.
In this work we put forward a computationally efficient statistical procedure to estimate a wide range of time series models in a robust manner thereby reducing the influence of data contamination on estimation and inference.