Multiscale estimation of terrain complexity using ALSM point data on variable resolution grids

Abstract

Multiscale Kalman smoothers (MKS) have been previously employed for data fusion applications and estimation of topography. However, the standard MKS algorithm embedded with a single stochastic model gives suboptimal performance when estimating non-stationary topographic variations, particularly when there are sudden changes in the terrain. In this work, multiple MKS models are regulated by a mixture-of-experts (MOE) network to adaptively fuse the estimates. Though MOE has been widely applied to onedimensional time series data, its extension to multiscale estimation is new.