Simultaneous Localization and Mapping: Through the Lens of Nonlinear Optimization

TL;DR

This paper introduces a unified optimization-based framework for SLAM that encompasses filtering and batch methods, providing insights into their performance differences and enabling flexible algorithm design.

Contribution

It presents a generalized framework that unifies filtering and optimization approaches in SLAM, with reformulations and empirical comparisons demonstrating its versatility.

Findings

Filtering corresponds to specific design choices in the framework.

The reformulated MSCKF performs well on challenging datasets.

Empirical performance interpolates between state-of-the-art methods.

Abstract

Simultaneous Localization and Mapping (SLAM) algorithms perform visual-inertial estimation via filtering or batch optimization methods. Empirical evidence suggests that filtering algorithms are computationally faster, while optimization methods are more accurate. This work presents an optimization-based framework that unifies these approaches, and allows users to flexibly implement different design choices, e.g., the number and types of variables maintained in the algorithm at each time. We prove that filtering methods correspond to specific design choices in our generalized framework. We then reformulate the Multi-State Constrained Kalman Filter (MSCKF), implement the reformulation on challenging image sequence datasets in simulation, and contrast its performance with that of sliding window based filters. Using these results, we explain the relative performance characteristics of these two classes of algorithms in the context of our algorithm. Finally, we illustrate that under different design choices, the empirical performance of our algorithm interpolates between those of state-of-the-art approaches.