{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,9]],"date-time":"2026-06-09T16:24:00Z","timestamp":1781022240543,"version":"3.54.1"},"reference-count":52,"publisher":"MDPI AG","issue":"2","license":[{"start":{"date-parts":[[2026,2,14]],"date-time":"2026-02-14T00:00:00Z","timestamp":1771027200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IJGI"],"abstract":"<jats:p>General-purpose large language models excel at open-domain question answering, but in railway operation and maintenance (O&amp;M) scenarios they still suffer from hallucinated knowledge and poor domain adaptation. In practice, railway O&amp;M knowledge mainly arises from two heterogeneous sources: spatio-temporal data such as train trajectories, which are organized along the spatial layout of railway lines, and domain documents such as operating rules, which exhibit varying degrees of structural regularity. Traditional retrieval-augmented generation (RAG) systems usually flatten these multi-source data into a single unstructured text space and perform global retrieval in one embedding space, which easily introduces noisy context and makes it difficult to precisely target knowledge for specific lines, sections, or equipment states. To overcome these limitations, we propose CARAG, a context-aware RAG framework tailored to railway O&amp;M data. CARAG treats domain documents and spatial data as a unified knowledge substrate and builds a spatial knowledge graph with concept and instance levels. On top of this knowledge graph, a GraphReAct-based multi-turn interaction mechanism guides the LLM to reason and act over the concept knowledge graph, dynamically navigating to spatially and semantically relevant candidate regions, within which vector retrieval and instance-level graph retrieval are performed. Experiments show that CARAG significantly outperforms baseline RAG methods on RAGAS metrics, confirming the effectiveness of structure-guided multi-step reasoning for question answering over multi-source heterogeneous railway O&amp;M data.<\/jats:p>","DOI":"10.3390\/ijgi15020078","type":"journal-article","created":{"date-parts":[[2026,2,16]],"date-time":"2026-02-16T14:55:47Z","timestamp":1771253747000},"page":"78","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":2,"title":["CARAG: Context-Aware Retrieval-Augmented Generation for Railway Operation and Maintenance Question Answering over Spatial Knowledge Graph"],"prefix":"10.3390","volume":"15","author":[{"ORCID":"https:\/\/orcid.org\/0009-0008-8503-7289","authenticated-orcid":false,"given":"Wenkui","family":"Zheng","sequence":"first","affiliation":[{"name":"School of Software, Henan University, Kaifeng 475004, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0000-8958-824X","authenticated-orcid":false,"given":"Mengzheng","family":"Yang","sequence":"additional","affiliation":[{"name":"School of Software, Henan University, Kaifeng 475004, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yanfei","family":"Ren","sequence":"additional","affiliation":[{"name":"School of Software, Henan University, Kaifeng 475004, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Haoyu","family":"Wang","sequence":"additional","affiliation":[{"name":"School of Software, Henan University, Kaifeng 475004, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Chun","family":"Zeng","sequence":"additional","affiliation":[{"name":"School of Software, Henan University, Kaifeng 475004, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8803-2055","authenticated-orcid":false,"given":"Yong","family":"Zhang","sequence":"additional","affiliation":[{"name":"BNRist, DCST, RIIT, Tsinghua University, Beijing 100084, China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2026,2,14]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"47","DOI":"10.1016\/j.jrst.2025.08.001","article-title":"A critical review of operations research on the operation and maintenance of railway systems","volume":"1","author":"Huang","year":"2025","journal-title":"J. Railw. Sci. Technol."},{"key":"ref_2","doi-asserted-by":"crossref","unstructured":"Yin, L., and Wang, Y. (2023). Spatiotemporal Evolution Analysis of the Chinese Railway Network Structure Based on Self-Organizing Maps. ISPRS Int. J. Geo-Inf., 12.","DOI":"10.3390\/ijgi12040161"},{"key":"ref_3","doi-asserted-by":"crossref","unstructured":"Ait-Lamallam, S., Yaagoubi, R., Sebari, I., and Doukari, O. (2021). Extending the IFC Standard to Enable Road Operation and Maintenance Management through OpenBIM. ISPRS Int. J. Geo-Inf., 10.","DOI":"10.3390\/ijgi10080496"},{"key":"ref_4","doi-asserted-by":"crossref","unstructured":"Yang, J., Jang, H., and Yu, K. (2024). Geographic Knowledge Base Question Answering over OpenStreetMap. ISPRS Int. J. Geo-Inf., 13.","DOI":"10.3390\/ijgi13010010"},{"key":"ref_5","doi-asserted-by":"crossref","unstructured":"Li, W., Wang, S., Chen, X., Tian, Y., Gu, Z., Lopez-Carr, A., Schroeder, A., Currier, K., Schildhauer, M., and Zhu, R. (2023). GeoGraphVis: A Knowledge Graph and Geovisualization Empowered Cyberinfrastructure to Support Disaster Response and Humanitarian Aid. ISPRS Int. J. Geo-Inf., 12.","DOI":"10.3390\/ijgi12030112"},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"5282","DOI":"10.11591\/ijece.v13i5.pp5282-5295","article-title":"An approach of ontology and knowledge base for railway maintenance","volume":"13","author":"Ragala","year":"2023","journal-title":"Int. J. Electr. Comput. Eng."},{"key":"ref_7","first-page":"601","article-title":"An intelligent railway operation and maintenance management approach based on BIM and semantic web","volume":"45","author":"He","year":"2024","journal-title":"J. Graph."},{"key":"ref_8","doi-asserted-by":"crossref","unstructured":"Li, W., Song, M., and Tian, Y. (2019). An Ontology-Driven Cyberinfrastructure for Intelligent Spatiotemporal Question Answering and Open Knowledge Discovery. ISPRS Int. J. Geo-Inf., 8.","DOI":"10.3390\/ijgi8110496"},{"key":"ref_9","doi-asserted-by":"crossref","unstructured":"Jia, L., Wang, L., and Qin, Y. (2022). Decision Support System and Demonstration Application of High-Speed Railway Train Operation Under Emergency Conditions. High-Speed Railway Operation Under Emergent Conditions: Theoretical Methodology and Applications, Springer.","DOI":"10.1007\/978-3-662-63033-4"},{"key":"ref_10","doi-asserted-by":"crossref","first-page":"118050","DOI":"10.1016\/j.eswa.2022.118050","article-title":"Data-driven decision support for rail traffic control: A predictive approach","volume":"207","author":"Luo","year":"2022","journal-title":"Expert Syst. Appl."},{"key":"ref_11","unstructured":"Achiam, J., Adler, S., Agarwal, S., Ahmad, L., Akkaya, I., Aleman, F.L., Almeida, D., Altenschmidt, J., Altman, S., and Anadkat, S. (2023). GPT-4 Technical Report. arXiv."},{"key":"ref_12","unstructured":"Guo, D., Yang, D., Zhang, H., Song, J., Zhang, R., Xu, R., Zhu, Q., Ma, S., Wang, P., and Bi, X. (2025). Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning. arXiv."},{"key":"ref_13","unstructured":"Yang, A., Li, A., Yang, B., Zhang, B., Hui, B., Zheng, B., Yu, B., Gao, C., Huang, C., and Lv, C. (2025). Qwen3 technical report. arXiv."},{"key":"ref_14","first-page":"1","article-title":"A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questions","volume":"43","author":"Huang","year":"2025","journal-title":"ACM Trans. Inf. Syst."},{"key":"ref_15","doi-asserted-by":"crossref","unstructured":"Xia, Y., Kim, J., Chen, Y., Ye, H., Kundu, S., Hao, C.C., and Talati, N. (2024). Understanding the Performance and Estimating the Cost of LLM Fine-Tuning. 2024 IEEE International Symposium on Workload Characterization (IISWC), IEEE.","DOI":"10.1109\/IISWC63097.2024.00027"},{"key":"ref_16","first-page":"9459","article-title":"Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks","volume":"Volume 33","author":"Larochelle","year":"2020","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref_17","unstructured":"Christodoulopoulos, C., Chakraborty, T., Rose, C., and Peng, V. (2025). MultiDocFusion: Hierarchical and Multimodal Chunking Pipeline for Enhanced RAG on Long Industrial Documents. 2025 Conference on Empirical Methods in Natural Language Processing, Association for Computational Linguistics."},{"key":"ref_18","first-page":"23424","article-title":"LightPROF: A Lightweight Reasoning Framework for Large Language Model on Knowledge Graph","volume":"39","author":"Ao","year":"2025","journal-title":"Proc. AAAI Conf. Artif. Intell."},{"key":"ref_19","unstructured":"Edge, D., Trinh, H., Cheng, N., Bradley, J., Chao, A., Mody, A., Truitt, S., Metropolitansky, D., Ness, R.O., and Larson, J. (2024). From local to global: A graph rag approach to query-focused summarization. arXiv."},{"key":"ref_20","unstructured":"Che, W., Nabende, J., Shutova, E., and Pilehvar, M.T. (2025). Hierarchical Document Refinement for Long-context Retrieval-augmented Generation. 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), Association for Computational Linguistics."},{"key":"ref_21","unstructured":"Christodoulopoulos, C., Chakraborty, T., Rose, C., and Peng, V. (2025). KERAG: Knowledge-Enhanced Retrieval-Augmented Generation for Advanced Question Answering. Proceedings of the Findings of the Association for Computational Linguistics: EMNLP 2025, Association for Computational Linguistics."},{"key":"ref_22","doi-asserted-by":"crossref","unstructured":"Pirolli, P., and Card, S. (1995, January 7\u201311). Information foraging in information access environments. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, Denver, CO, USA.","DOI":"10.1145\/223904.223911"},{"key":"ref_23","unstructured":"Yao, S., Zhao, J., Yu, D., Du, N., Shafran, I., Narasimhan, K.R., and Cao, Y. (2023, January 1\u20135). ReAct: Synergizing Reasoning and Acting in Language Models. Proceedings of the Eleventh International Conference on Learning Representations, Kigali, Rwanda."},{"key":"ref_24","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1145\/3657631","article-title":"Interactive Question Answering Systems: Literature Review","volume":"56","author":"Biancofiore","year":"2024","journal-title":"ACM Comput. Surv."},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"717","DOI":"10.13052\/jwe1540-9589.1785","article-title":"A Review of Question Answering Systems","volume":"17","author":"Ojokoh","year":"2018","journal-title":"J. Web Eng."},{"key":"ref_26","doi-asserted-by":"crossref","first-page":"605","DOI":"10.1016\/j.jmsy.2020.04.007","article-title":"Transferable two-stream convolutional neural network for human action recognition","volume":"56","author":"Xiong","year":"2020","journal-title":"J. Manuf. Syst."},{"key":"ref_27","doi-asserted-by":"crossref","first-page":"261","DOI":"10.1016\/j.jmsy.2020.07.011","article-title":"Digital twin improved via visual question answering for vision-language interactive mode in human\u2013machine collaboration","volume":"58","author":"Wang","year":"2021","journal-title":"J. Manuf. Syst."},{"key":"ref_28","doi-asserted-by":"crossref","first-page":"122999","DOI":"10.1016\/j.eswa.2023.122999","article-title":"Modeling and application of a customized knowledge graph for railway alignment optimization","volume":"244","author":"Pu","year":"2024","journal-title":"Expert Syst. Appl."},{"key":"ref_29","doi-asserted-by":"crossref","unstructured":"Tang, X., Huynh, V.N., Xia, H., and Bai, Q. (2025). A Domain Knowledge-Based Railway Equipment Fault Diagnosis Framework. Knowledge and Systems Sciences, Springer.","DOI":"10.1007\/978-981-96-0178-3"},{"key":"ref_30","doi-asserted-by":"crossref","first-page":"494","DOI":"10.1109\/TNNLS.2021.3070843","article-title":"A Survey on Knowledge Graphs: Representation, Acquisition, and Applications","volume":"33","author":"Ji","year":"2022","journal-title":"IEEE Trans. Neural Netw. Learn. Syst."},{"key":"ref_31","first-page":"489","article-title":"Knowledge graph refinement: A survey of approaches and evaluation methods","volume":"8","author":"Paulheim","year":"2016","journal-title":"Semant. Web"},{"key":"ref_32","doi-asserted-by":"crossref","unstructured":"Suchanek, F.M., Kasneci, G., and Weikum, G. (2007). Yago: A core of semantic knowledge. 16th International Conference on World Wide Web, Association for Computing Machinery.","DOI":"10.1145\/1242572.1242667"},{"key":"ref_33","unstructured":"Moore, J.D., Teufel, S., Allan, J., and Furui, S. (2008). The Tradeoffs Between Open and Traditional Relation Extraction. ACL-08: HLT, Association for Computational Linguistics."},{"key":"ref_34","first-page":"9523","article-title":"Span Model for Open Information Extraction on Accurate Corpus","volume":"34","author":"Zhan","year":"2020","journal-title":"Proc. AAAI Conf. Artif. Intell."},{"key":"ref_35","unstructured":"Ding, L., Zhou, S., Xiao, J., and Han, J. (2024). Automated construction of theme-specific knowledge graphs. arXiv."},{"key":"ref_36","unstructured":"Cao, Y., Feng, Y., and Xiong, D. (2024). AutoRE: Document-Level Relation Extraction with Large Language Models. 62nd Annual Meeting of the Association for Computational Linguistics (Volume 3: System Demonstrations), Association for Computational Linguistics."},{"key":"ref_37","unstructured":"Carta, S., Giuliani, A., Piano, L., Podda, A.S., Pompianu, L., and Tiddia, S.G. (2023). Iterative zero-shot llm prompting for knowledge graph construction. arXiv."},{"key":"ref_38","unstructured":"Ku, L.W., Martins, A., and Srikumar, V. (2024). SAC-KG: Exploiting Large Language Models as Skilled Automatic Constructors for Domain Knowledge Graph. 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), Association for Computational Linguistics."},{"key":"ref_39","unstructured":"Christodoulopoulos, C., Chakraborty, T., Rose, C., and Peng, V. (2025). LightRAG: Simple and Fast Retrieval-Augmented Generation. Findings of the Association for Computational Linguistics: EMNLP 2025, Association for Computational Linguistics."},{"key":"ref_40","doi-asserted-by":"crossref","unstructured":"Sarmah, B., Mehta, D., Hall, B., Rao, R., Patel, S., and Pasquali, S. (2024). HybridRAG: Integrating Knowledge Graphs and Vector Retrieval Augmented Generation for Efficient Information Extraction. 5th ACM International Conference on AI in Finance, Association for Computing Machinery.","DOI":"10.1145\/3677052.3698671"},{"key":"ref_41","first-page":"38154","article-title":"HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in Hugging Face","volume":"Volume 36","author":"Oh","year":"2023","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref_42","first-page":"24824","article-title":"Chain-of-Thought Prompting Elicits Reasoning in Large Language Models","volume":"Volume 35","author":"Koyejo","year":"2022","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref_43","unstructured":"Wang, X., Wei, J., Schuurmans, D., Le, Q.V., Chi, E.H., Narang, S., Chowdhery, A., and Zhou, D. (2023, January 1\u20135). Self-Consistency Improves Chain of Thought Reasoning in Language Models. Proceedings of the Eleventh International Conference on Learning Representations, Kigali, Rwanda."},{"key":"ref_44","first-page":"11733","article-title":"Improving Factuality and Reasoning in Language Models through Multiagent Debate","volume":"Volume 235","author":"Salakhutdinov","year":"2024","journal-title":"41st International Conference on Machine Learning"},{"key":"ref_45","unstructured":"Muresan, S., Nakov, P., and Villavicencio, A. (2022). Fantastically Ordered Prompts and Where to Find Them: Overcoming Few-Shot Prompt Order Sensitivity. 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), Association for Computational Linguistics."},{"key":"ref_46","unstructured":"G\u00fcnther, M., Mohr, I., Williams, D.J., Wang, B., and Xiao, H. (2024). Late chunking: Contextual chunk embeddings using long-context embedding models. arXiv."},{"key":"ref_47","unstructured":"Zhao, J., Ji, Z., Feng, Y., Qi, P., Niu, S., Tang, B., Xiong, F., and Li, Z. (2024). Meta-Chunking: Learning Text Segmentation and Semantic Completion via Logical Perception. arXiv."},{"key":"ref_48","unstructured":"Aletras, N., and De Clercq, O. (2024). RAGAs: Automated Evaluation of Retrieval Augmented Generation. 18th Conference of the European Chapter of the Association for Computational Linguistics: System Demonstrations, Association for Computational Linguistics."},{"key":"ref_49","first-page":"8634","article-title":"Reflexion: Language agents with verbal reinforcement learning","volume":"Volume 36","author":"Oh","year":"2023","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref_50","unstructured":"Gao, Y., Xiong, Y., Gao, X., Jia, K., Pan, J., Bi, Y., Dai, Y., Sun, J., Wang, H., and Wang, H. (2023). Retrieval-augmented generation for large language models: A survey. arXiv."},{"key":"ref_51","unstructured":"Al-Onaizan, Y., Bansal, M., and Chen, Y.N. (2024). LumberChunker: Long-Form Narrative Document Segmentation. Findings of the Association for Computational Linguistics: EMNLP 2024, Association for Computational Linguistics."},{"key":"ref_52","unstructured":"Sarthi, P., Abdullah, S., Tuli, A., Khanna, S., Goldie, A., and Manning, C.D. (2024, January 7\u201311). RAPTOR: Recursive Abstractive Processing for Tree-Organized Retrieval. Proceedings of the Twelfth International Conference on Learning Representations, Vienna, Austria."}],"container-title":["ISPRS International Journal of Geo-Information"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2220-9964\/15\/2\/78\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,2,16]],"date-time":"2026-02-16T15:10:37Z","timestamp":1771254637000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2220-9964\/15\/2\/78"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,2,14]]},"references-count":52,"journal-issue":{"issue":"2","published-online":{"date-parts":[[2026,2]]}},"alternative-id":["ijgi15020078"],"URL":"https:\/\/doi.org\/10.3390\/ijgi15020078","relation":{},"ISSN":["2220-9964"],"issn-type":[{"value":"2220-9964","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,2,14]]}}}