TBV Radar SLAM -- trust but verify loop candidates

TL;DR

TBV Radar SLAM introduces a novel radar SLAM approach that verifies loop closure candidates through multiple techniques, significantly reducing errors and demonstrating robustness and generalization across environments.

Contribution

It presents a new introspective loop verification method for radar SLAM that improves accuracy and robustness without environment-specific tuning.

Findings

65% lower error than previous state-of-the-art

High correct-loop-retrieval rate achieved

Generalizes across different environments

Abstract

Robust SLAM in large-scale environments requires fault resilience and awareness at multiple stages, from sensing and odometry estimation to loop closure. In this work, we present TBV (Trust But Verify) Radar SLAM, a method for radar SLAM that introspectively verifies loop closure candidates. TBV Radar SLAM achieves a high correct-loop-retrieval rate by combining multiple place-recognition techniques: tightly coupled place similarity and odometry uncertainty search, creating loop descriptors from origin-shifted scans, and delaying loop selection until after verification. Robustness to false constraints is achieved by carefully verifying and selecting the most likely ones from multiple loop constraints. Importantly, the verification and selection are carried out after registration when additional sources of loop evidence can easily be computed. We integrate our loop retrieval and verification method with a fault-resilient odometry pipeline within a pose graph framework. By evaluating on public benchmarks we found that TBV Radar SLAM achieves 65% lower error than the previous state of the art. We also show that it's generalizing across environments without needing to change any parameters.