Variational State-Space Models for Localisation and Dense 3D Mapping in 6 DoF

TL;DR

This paper introduces a variational state-space model that combines learning and domain knowledge for accurate 6-DoF localisation and dense 3D mapping, optimized end-to-end for UAV navigation.

Contribution

It presents a novel deep Bayesian framework integrating multi-view geometry and dynamics for improved visual SLAM performance.

Findings

Achieves near state-of-the-art accuracy on UAV flight data.

Demonstrates effective generative prediction and planning capabilities.

Abstract

We solve the problem of 6-DoF localisation and 3D dense reconstruction in spatial environments as approximate Bayesian inference in a deep state-space model. Our approach leverages both learning and domain knowledge from multiple-view geometry and rigid-body dynamics. This results in an expressive predictive model of the world, often missing in current state-of-the-art visual SLAM solutions. The combination of variational inference, neural networks and a differentiable raycaster ensures that our model is amenable to end-to-end gradient-based optimisation. We evaluate our approach on realistic unmanned aerial vehicle flight data, nearing the performance of state-of-the-art visual-inertial odometry systems. We demonstrate the applicability of the model to generative prediction and planning.