aiqbench analysis
SODA-CitrON addresses a critical gap in robotics and autonomous systems by offering an unsupervised, online clustering method specifically optimized for static object tracking amidst heterogeneous sensor data. The approach outperforms state-of-the-art techniques like JPDA and DBSTREAM in key metrics such as F1 score and position RMSE while maintaining loglinear computational complexity and full output explainability. This advancement is particularly relevant for environmental mapping applications where motion models are ineffective and clutter management is essential.
SODA-CitrON: Static Object Data Association by Clustering Multi-Modal Sensor Detections Online
arXiv:2602.22243v3 Announce Type: replace Abstract: The online fusion and tracking of static objects from heterogeneous sensor detections is a fundamental problem in robotics, autonomous systems, and environmental mapping. Although classical data association approaches such as JPDA are well suited for dynamic targets, they are less effective for static objects observed intermittently and with heterogeneous uncertainties, where motion models provide minimal discriminative power with respect to clutter. In this paper, we propose a novel method for static object data association by clustering multi-modal sensor detections online (SODA-CitrON), while simultaneously estimating positions and maintaining persistent tracks for an unknown number of objects. The proposed unsupervised machine learning approach operates in a fully online manner and handles temporally uncorrelated and multi-sensor measurements. Additionally, it has a worst-case loglinear complexity in the number of sensor detections while providing full output explainability. We evaluate the proposed approach in different Monte Carlo simulation scenarios and compare it against state-of-the-art methods, including POM-based filtering, DBSTREAM clustering, and JPDA. The results demonstrate that SODA-CitrON consistently outperforms the compared methods in terms of F1 score, position RMSE, MOTP, and MOTA in the static object mapping scenarios studied.
Abstract:The online fusion and tracking of static objects from heterogeneous sensor detections is a fundamental problem in robotics, autonomous systems, and environmental mapping. Although classical data association approaches such as JPDA are well suited for dynamic targets, they are less effective for static objects observed intermittently and with heterogeneous uncertainties, where motion models provide minimal discriminative power with respect to clutter. In this paper, we propose a novel method for static object data association by clustering multi-modal sensor detections online (SODA-CitrON), while simultaneously estimating positions and maintaining persistent tracks for an unknown number of objects. The proposed unsupervised machine learning approach operates in a fully online manner and handles temporally uncorrelated and multi-sensor measurements. Additionally, it has a worst-case loglinear complexity in the number of sensor detections while providing full output explainability. We evaluate the proposed approach in different Monte Carlo simulation scenarios and compare it against state-of-the-art methods, including POM-based filtering, DBSTREAM clustering, and JPDA. The results demonstrate that SODA-CitrON consistently outperforms the compared methods in terms of F1 score, position RMSE, MOTP, and MOTA in the static object mapping scenarios studied.
8 pages, 5 figures; \c{opyright} 2026 IEEE. Accepted for the 2026 International Conference on Information Fusion (FUSION 2026)
Robotics (cs.RO)
arXiv:2602.22243 [cs.RO]
(or arXiv:2602.22243v3 [cs.RO] for this version)
Full text under the publisher's open licence · 211 words