Bayesian Pose Graph Optimization via Bingham Distributions and Tempered Geodesic MCMC
This addresses the challenge of non-convex optimization and uncertainty estimation in robotics and computer vision, offering a novel method for improved initialization in pose graph problems.
The paper tackles the problem of initializing pose graph optimization in scenarios like SLAM and SFM by introducing the Tempered Geodesic MCMC (TG-MCMC) algorithm, which provides reliable initial poses and uncertainty estimates, with evaluation on synthetic and real benchmark datasets showing robustness to missing data and noise.
We introduce Tempered Geodesic Markov Chain Monte Carlo (TG-MCMC) algorithm for initializing pose graph optimization problems, arising in various scenarios such as SFM (structure from motion) or SLAM (simultaneous localization and mapping). TG-MCMC is first of its kind as it unites asymptotically global non-convex optimization on the spherical manifold of quaternions with posterior sampling, in order to provide both reliable initial poses and uncertainty estimates that are informative about the quality of individual solutions. We devise rigorous theoretical convergence guarantees for our method and extensively evaluate it on synthetic and real benchmark datasets. Besides its elegance in formulation and theory, we show that our method is robust to missing data, noise and the estimated uncertainties capture intuitive properties of the data.