inttransИнтеллектуальный транспортIntelligent transport3033-6007АО «НИИАС»10.24412/3033-6007-2026-339-37-52inttrans-115Research ArticleИСКУССТВЕННЫЙ ИНТЕЛЛЕКТ И МАШИННОЕ ОБУЧЕНИЕПостроение компактных графов сцены на основе топологической карты для автономной навигации мобильного роботаBuilding Compact Scene Graphs Based on a Topological Map for Autonomous Navigation of a Mobile RobotМуравьевК. Ф.MuravyevK. F.

к.т.н., научный сотрудник

PhD, Research Fellow

kmuravev@frccsc.ruРоманенкоВ. И.RomanenkoV. I.

студент факультета компьютерных наук

Student of the Faculty of Computer Science

viktorsamr@yandex.ruФедеральный исследовательский центр «Информатика и управление» Российской академии наук (ФИЦ ИУ РАН)РоссияFederal Research Center «Computer Science and Control» of the Russian Academy of Sciences (FRC CSC RAS)Russian FederationНациональный исследовательский университет «Высшая школа экономики» (НИУ ВШЭ)РоссияNational Research University Higher School of Economics (NRU HSE)Russian Federation202627092026103(39)3752Copyright © Муравьев К.Ф., Романенко В.И., 20262026Муравьев К.Ф., Романенко В.И.Muravyev K.F., Romanenko V.I.Данная работа распространяется под лицензией Creative Commons Attribution 4.0.This work is licensed under a Creative Commons Attribution 4.0 License.https://www.intelligent-transport.ru/jour/article/view/115

Автономная навигация мобильного робота в человеко-ориентированных средах требует наличия карты, которая является не только геометрической моделью среды для планирования маршрутов, но также содержит информацию об объектах среды (двери, мебель, офисная техника и т.д.). Такое представление карты обеспечивают графы сцены, в которых вершинами являются помещения, локации и объекты, а ребра кодируют пространственные связности или отношения между объектами. Большинство современных методов построения графов сцены обладают высокой вычислительной сложностью, и построенные ими графы являются избыточными для обеспечения автономной навигации робота. В данной работе предлагается метод построения компактных графов сцены (Compact Scene Graph, CSG), в основе которого лежит вычислительно эффективный метод топологического картирования PRISM-TopoMap и привязка семантических объектов к локациям топологической карты. Построенный граф сцены позволяет строить маршруты до объектов по локациям топологической карты и осуществлять локализацию с высокой точностью. Предложенный метод был экспериментально исследован в симуляционной среде Habitat. Результаты экспериментов показали, что предложенный CSG потребляет значительно меньше памяти, чем традиционные метрические карты и графы сцен, и обеспечивает надежную локализацию по топологической карте за счет привязки к семантическим объектам.

Autonomous navigation of a mobile robot in human-centered environments requires a map that contains not only a geometric model of the environment for path planning, but also information about environmental objects (doors, furniture, office equipment, etc.). Scene graphs provide such a map representation, where nodes correspond to rooms, locations, and objects, and edges encode spatial connectivity or relationships between objects. Most modern scene graph construction methods have high computational complexity, and the graphs they produce are redundant for the purposes of autonomous robot navigation. This paper proposes a method for constructing Compact Scene Graphs (CSG), which is based on the computationally efficient topological mapping method PRISM-TopoMap and the association of semantic objects with locations on the topological map. The resulting scene graph enables route planning to objects via topological map locations and achieves high-precision localization. The proposed method was experimentally evaluated in the Habitat simulation environment. The experimental results demonstrate that the proposed CSG consumes significantly less memory than traditional metric maps and scene graphs, while providing reliable localization on the topological map through association with semantic objects.

граф сценытопологическая картасемантическая сегментацияавтономная навигациясимуляционная средаscene graphtopological mapsemantic segmentationautonomous navigationsimulated environmentReferencesLabb´e, M. RTAB-Map as an open-source lidar and visual simultaneous localization and mapping library for large-scale and long-term online operation / M. Labb´e, F. Michaud // Journal of Field Robotics. – 2019. – Vol. 36, No. 2. – P. 416–446. – DOI 10.1002/rob.21831.Labbé, M., & Michaud, F. (2019). RTAB-Map as an open-source lidar and visual simultaneous localization and mapping library for large-scale and long-term online operation. Journal of Field Robotics, 36(2), 416–446. https://doi.org/10.1002/rob.21831GLIM: 3D range-inertial localization and mapping with GPU-accelerated scan matching factors / K. Koide, M. Yokozuka, S. Oishi, A. Banno // Robotics and Autonomous Systems. – 2024. – Vol. 179. – P. 104750. – DOI 10.1016/j.robot.2024.104750.Koide, K., Yokozuka, M., Oishi, S., & Banno, A. (2024). GLIM: 3D range-inertial localization and mapping with GPU-accelerated scan matching factors. Robotics and Autonomous Systems, , Article 104750. https://doi.org/10.1016/j.robot.2024.104750Muravyev, K. Evaluation of RGB-D SLAM in large indoor environments / K. Muravyev, K. Yakovlev // International Conference on Interactive Collaborative Robotics. – 2022. – P. 93–104. – DOI 10.1007/978-3-031-23609-9_9.Muravyev, K., & Yakovlev, K. (2022). Evaluation of RGB-D SLAM in large indoor environments. In International Conference on Interactive Collaborative Robotics (pp. 93–104). https://doi.org/10.1007/978-3-031-23609-9_9Muravyev, K. Evaluation of topological mapping methods in indoor environments / K. Muravyev, K. S. Yakovlev // IEEE Access. – 2023. – Vol. 11. – P. 132683–132698. – DOI 10.1109/ACCESS.2023.3335818.Muravyev, K., & Yakovlev, K. S. (2023). Evaluation of topological mapping methods in indoor environments. IEEE Access, 11, 132683–132698. https://doi.org/10.1109/ACCESS.2023.PRISM-TopoMap: Online topological mapping with place recognition and scan matching / K. Muravyev, A. Melekhin, D. Yudin, K. Yakovlev // IEEE Robotics and Automation Letters. – 2025. – Vol. 10, No. 4. – P. 3126–3133. – DOI 10.1109/LRA.2025.3541454.Muravyev, K., Melekhin, A., Yudin, D., & Yakovlev, K. (2025). PRISM-TopoMap: Online topological mapping with place recognition and scan matching. IEEE Robotics and Automation Letters, 10(4), 3126–3133. https://doi.org/10.1109/LRA.2025.3541454Topological semantic graph memory for image-goal navigation / N. Kim, O. Kwon, H. Yoo [et al.] // Proceedings of the 6th Conference on Robot Learning (CoRL), PMLR. – 2023. – Vol. 205. – P. 393–402.Kim, N., Kwon, O., Yoo, H., Choi, Y., Park, J., & Oh, S. (2023). Topological semantic graph memory for image-goal navigation. In Proceedings of the 6th Conference on Robot Learning (CoRL) (Vol. 205, pp. 393–402). PMLR.Clio: Real-time task-driven open-set 3D scene graphs / D. Maggio, Y. Chang, N. Hughes [et al.] // IEEE Robotics and Automation Letters. – 2024. – Vol. 9, No. 10. – P. 8921–8928. – DOI 10.1109/LRA.2024.3451395.Maggio, D., Chang, Y., Hughes, N., Trang, M., Griffith, D., Dougherty, C., Cristofalo, E., Schmid, L., & Carlone, L. (2024). Clio: Real-time task-driven open-set 3D scene graphs. IEEE Robotics and Automation Letters, 9(10), 8921–8928. https://doi.org/10.1109/LRA.2024.3451395Hughes, N. Hydra: A real-time spatial perception system for 3D scene graph construction and optimization / N. Hughes, Y. Chang, L. Carlone // arXiv. – 2022. – arXiv:2201.13360. – DOI 10.48550/arXiv.2201.13360.Hughes, N., Chang, Y., & Carlone, L. (2022). Hydra: A real-time spatial perception system for 3D scene graph construction and optimization (arXiv:2201.13360). arXiv. https://doi.org/10.48550/arXiv.2201.13360Khronos: Middleware for simplified time management in CPS / S. Peros, S. Delbruel, S. Michiels [et al.] // Proceedings of the 13th ACM International Conference on Distributed and Event-based Systems. – 2019. – P. 127–138. – DOI 10.1145/3328905.3329507.Peros, S., Delbruel, S., Michiels, S., Joosen, W., & Hughes, D. (2019). Khronos: Middleware for simplified time management in CPS. In Proceedings of the 13th ACM International Conference on Distributed and Event-based Systems (pp. 127–138). https://doi.org/10.1145/3328905.Habitat: A platform for embodied AI research / M. Savva, A. Kadian, O. Maksymets [et al.] // Proceedings of the IEEE/CVF International Conference on Computer Vision. – 2019. – P. 9338–9346. – DOI 10.1109/ICCV.2019.00943.Savva, M., Kadian, A., Maksymets, O., Zhao, Y., Wijmans, E., Jain, B., Straub, J., Liu, J., Koltun, V., Malik, J., Batra, D., & Mottaghi, R. (2019). Habitat: A platform for embodied AI research. In Proceedings of the IEEE/CVF International Conference on Computer Vision (pp.–9346). https://doi.org/10.1109/ICCV.2019.00943A modular robotic system for autonomous exploration and semantic updating in large-scale indoor environments / S. H. Allu, I. Kadosh, T. Summers, Y. Xiang // arXiv. – 2024. – arXiv:2409.15493. – DOI 10.48550/arXiv.2409.15493.Allu, S. H., Kadosh, I., Summers, T., & Xiang, Y. (2024). A modular robotic system for autonomous exploration and semantic updating in large-scale indoor environments (arXiv:2409.15493). arXiv. https://doi.org/10.48550/arXiv.2409.15493Carpenter, G. A. Adaptive resonance theory : Tech. Rep. / G. A. Carpenter, S. Grossberg. – Boston, MA, USA : Boston University, Center for Adaptive Systems and Department of Cognitive and Neural Systems, 1993.Carpenter, G. A., & Grossberg, S. (1993). Adaptive resonance theory (Technical Report). Boston University, Center for Adaptive Systems and Department of Cognitive and Neural Systems.TopoNav: Topological navigation for efficient exploration in sparse reward environments / J. Hossain, A. Z. M. Faridee, N. Roy [et al.] // 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). – 2024. – P. 693–700. – DOI 10.1109/IROS58592.2024.10802380.Hossain, J., Faridee, A. Z. M., Roy, N., Freeman, J., Gregory, T., & Trout, T. (2024). TopoNav: Topological navigation for efficient exploration in sparse reward environments. In IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (pp. 693–700). https://doi.org/10.1109/IROS58592.2024.10802380SemanticTopoLoop: Semantic loop closure with 3D topological graph based on quadric-level object map / Z. Cao, Q. Zhang, J. Guang [et al.] // IEEE Robotics and Automation Letters. – 2024. – Vol. 9, No. 5. – P. 4257–4264. – DOI 10.1109/LRA.2024.3374169.Cao, Z., Zhang, Q., Guang, J., Wu, S., Hu, Z., & Liu, J. (2024). SemanticTopoLoop: Semantic loop closure with 3D topological graph based on quadric-level object map. IEEE Robotics and Automation Letters, 9(5), 4257–4264. https://doi.org/10.1109/LRA.2024.3374169Learning 3D semantic scene graphs from 3D indoor reconstructions / J. Wald, H. Dhamo, N. Navab, F. Tombari // Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. – 2020. – P. 3961–3970. – DOI 10.1109/CVPR42600.2020.00402.Wald, J., Dhamo, H., Navab, N., & Tombari, F. (2020). Learning 3D semantic scene graphs from D indoor reconstructions. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 3961–3970). https://doi.org/10.1109/CVPR42600.2020.00402Exploiting edge-oriented reasoning for 3D point-based scene graph analysis / C. Zhang, J. Yu, Y. Song, W. Cai // Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. – 2021. – P. 9705–9715. – DOI 10.1109/CVPR46437.2021.00958.Zhang, C., Yu, J., Song, Y., & Cai, W. (2021). Exploiting edge-oriented reasoning for 3D pointbased scene graph analysis. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 9705–9715). https://doi.org/10.1109/CVPR46437.2021.00958SGFormer++: Semantic graph transformer for incremental 3D scene graph generation / M. Qi, C. Lv, Z. Fu [et al.] // arXiv. – 2026. – arXiv:2606.15328. – DOI 10.48550/arXiv.2606.15328.Qi, M., Lv, C., Fu, Z., Zhang, X., & Ma, H. (2026). SGFormer++: Semantic graph transformer for incremental 3D scene graph generation (arXiv:2606.15328). arXiv. https://doi.org/10.48550/arXiv.2606.15328Learning transferable visual models from natural language supervision / A. Radford, J. W. Kim, C. Hallacy [et al.] // Proceedings of the 38th International Conference on Machine Learning, PMLR. – 2021. – Vol. 139. – P. 8748–8763.Radford, A., Kim, J. W., Hallacy, C., Ramesh, A., Goh, G., Agarwal, S., Sastry, G., Askell, A., Mishkin, P., Clark, J., Krueger, G., & Sutskever, I. (2021). Learning transferable visual models from natural language supervision. In Proceedings of the 38th International Conference on Machine Learning (Vol. 139, pp. 8748–8763). PMLR.SGRec3D: Self-supervised 3D scene graph learning via object-level scene reconstruction / S. Koch, N. Vaskevicius, M. Colosi [et al.] // Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision. – 2024. – P. 3392–3402. – DOI 10.1109/WACV57701.2024.00337.Koch, S., Vaskevicius, N., Colosi, M., Hermosilla, P., & Ropinski, T. (2024). SGRec3D: Self-supervised 3D scene graph learning via object-level scene reconstruction. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (pp. 3392–3402). https://doi.org/10.1109/WACV57701.2024.00337vS-Graphs: Integrating visual SLAM and situational graphs through multi-level scene understanding / A. Tourani, S. Ejaz, H. Bavle [et al.] // arXiv. – 2025. – arXiv:2503.01783. – DOI 10.48550/arXiv.2503.01783.Tourani, A., Ejaz, S., Bavle, H., Morilla-Cabello, D., Sanchez-Lopez, J. L., & Voos, H. (2025). vS-Graphs: Integrating visual SLAM and situational graphs through multi-level scene understanding (arXiv:2503.01783). arXiv. https://doi.org/10.48550/arXiv.2503.01783S-Graphs+: Real-time localization and mapping leveraging hierarchical representations / H. Bavle, J. L. S´anchez-L´opez, M. Shaheer [et al.] // IEEE Robotics and Automation Letters. – 2023. – Vol. 8, No. 8. – P. 4927–4934. – DOI 10.1109/LRA.2023.3290512.Bavle, H., Sánchez-López, J. L., Shaheer, M., Civera, J., & Voos, H. (2023). S-Graphs+: Real-time localization and mapping leveraging hierarchical representations. IEEE Robotics and Automation Letters, 8(8), 4927–4934. https://doi.org/10.1109/LRA.2023.3290512ConceptGraphs: Open-vocabulary 3D scene graphs for perception and planning / Q. Gu, A. Kuwajerwala, S. Morin [et al.] // 2024 IEEE International Conference on Robotics and Automation (ICRA). – 2024. – P. 5021–5028. – DOI 10.1109/ICRA57147.2024.10610243.Gu, Q., Kuwajerwala, A., Morin, S., Jatavallabhula, K. M., Sen, B., Agarwal, A., Rivera, C., Paul, W., Ellis, K., Chellappa, R., Gan, C., de Melo, C. M., Tenenbaum, J. B., Torralba, A., Shkurti, F., & Paull, L. (2024). ConceptGraphs: Open-vocabulary 3D scene graphs for perception and planning. In 2024 IEEE International Conference on Robotics and Automation (ICRA) (pp. 5021–5028). https://doi.org/10.1109/ICRA57147.2024.10610243Grounding DINO: Marrying DINO with grounded pre-training for open-set object detection / S. Liu, Z. Zeng, T. Ren [et al.] // Computer Vision – ECCV 2024, Springer. – 2024. – Vol. 47. – P. 38–55. – DOI 10.1007/978-3-031-72970-6_3.Liu, S., Zeng, Z., Ren, T., Li, F., Zhang, H., Yang, J., Jiang, Q., Li, C., Yang, J., Su, H., Zhu, J., & Zhang, L. (2024). Grounding DINO: Marrying DINO with grounded pre-training for open-set object detection. In Computer Vision – ECCV 2024 (Vol. 47, pp. 38–55). Springer. https://doi.org/10.1007/978-3-031-72970-6_3

The authors declare that there are no conflicts of interest present.