PROBE-GK: Predictive Robust Estimation using Generalized Kernels

TL;DR

This paper introduces PROBE-GK, a nonparametric Bayesian method for modeling sensor uncertainty, leading to more accurate and robust state estimation in dynamic environments for computer vision and robotics.

Contribution

It develops a predictive robust estimator using generalized kernels and Bayesian inference, learned from data without fixed noise assumptions.

Findings

Significant localization accuracy improvements over fixed noise models

Effective in synthetic, KITTI, and real-world experiments

Enhanced robustness in dynamic, uncertain environments

Abstract

Many algorithms in computer vision and robotics make strong assumptions about uncertainty, and rely on the validity of these assumptions to produce accurate and consistent state estimates. In practice, dynamic environments may degrade sensor performance in predictable ways that cannot be captured with static uncertainty parameters. In this paper, we employ fast nonparametric Bayesian inference techniques to more accurately model sensor uncertainty. By setting a prior on observation uncertainty, we derive a predictive robust estimator, and show how our model can be learned from sample images, both with and without knowledge of the motion used to generate the data. We validate our approach through Monte Carlo simulations, and report significant improvements in localization accuracy relative to a fixed noise model in several settings, including on synthetic data, the KITTI dataset, and our own experimental platform.