Claxton, OwenMalone, ConnorCarson, HelenFord, Jason J.Bolton, GabeShames, ImanMilford, Michael2025-05-232025-05-23http://www.scopus.com/inward/record.url?scp=85207328281&partnerID=8YFLogxKhttps://hdl.handle.net/1885/733751291Visual Place Recognition (VPR) systems often have imperfect performance, affecting the 'integrity' of position estimates and subsequent robot navigation decisions. Previously, SVM classifiers have been used to monitor VPR integrity. This research introduces a novel Multi-Layer Perceptron (MLP) integrity monitor which demonstrates improved performance and generalizability, removing per-environment training and reducing manual tuning requirements.We test our proposed system in extensive realworld experiments, presenting two real-time integrity-based VPR verification methods: a single-query rejection method for robot navigation to a goal zone (Experiment 1); and a history-of-queries method that takes a best, verified, match from its recent trajectory and uses an odometer to extrapolate a current position estimate (Experiment 2). Noteworthy results for Experiment 1 include a decrease in aggregatemean along-track goal error from≈9.8 mto ≈3.1 m, and an increase in the aggregate rate of successful mission completion from ≈41% to ≈55%. Experiment 2 showed a decrease in aggregate mean along-track localization error from ≈2.0 m to ≈0.5 m, and an increase in the aggregate localization precision from≈97% to≈99%.Overall, our results demonstrate the practical usefulness of a VPR integrity monitor in real-world robotics to improve VPR localization and consequent navigation performance.The work of Michael Milford was supported in part by the ARC Laureate Fellowship under Grant FL210100156, in part by the QUT Centre for Robotics, in part by the Centre for Advanced Defence Research in Robotics and Autonomous Systems, and in part by Australian Government under Grant AUSMURIB000001 through ONR MURI under Grant N00014-19-1-2571. This research is partially supported by an ARC Laureate Fellowship FL210100156 to M.Milford, the QUT Centre for Robotics, the Centre for Advanced Defence Research in Robotics and Autonomous Systems, and received funding from the Australian Government via grant AUSMURIB000001 associated with ONR MURI grant N00014-19-1-2571. The work of C.Malone and H.Carson was supported in part by an Australian Postgraduate Award.8enPublisher Copyright: © 2024 IEEE.acceptability and trustLocalizationvision-based navigation202410.1109/LRA.2024.348304585207328281