Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Dessislava Petrova-Antonova has a total of 169 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    A Scalable, Interoperable Reference Architecture for AI-Enabled Urban Digital Twins: From Geospatial Data Requirements to Operational Deployment

    by Dessislava Petrova-Antonova, Senthil Rajendran, Emil Hristov

    Urban Digital Twins (UDTs) increasingly combine heterogeneous geospatial data, real-time streams, analytics, and artificial intelligence (AI) to support urban governance, yet their reliability depends on how data quality, semantics, provenance, and uncertainty are handled architecturally. This study proposes a prescriptive, technology-agnostic reference architecture for AI-enabled UDT platforms covering GIS, BIM, IoT streams, imagery, and regulatory documents. The architecture combines presentation, logic, persistence, and cross-cutting support layers with reusable patterns for semantic mediation, data-type-specific persistence, spatiotemporal processing, scalable 3D visualisation, and explicit data-quality and provenance control. A four-stage methodology links stakeholder requirements to architectural patterns and validates their operational feasibility using the AQARI national real estate digital twin in Bahrain. The implementation demonstrates integration of spatial, transactional, sensor, 3D, and regulatory data and supports building management and zoning-compliance use cases. Operational tests show sub-second parcel-query response under typical conditions and scalable real-time ingestion. The results indicate that trustworthy AI-enabled UDTs require data-quality controls and interoperability mechanisms to be treated as architectural concerns rather than downstream application tasks. The proposed architecture provides reusable guidance for scalable spatial intelligence and digital-twin deployment, while quantitative uncertainty propagation and independent validation of AI models remain priorities for future work.

  • Open Access

    WRF–City–LES Modelling of the Urban Thermal Environment in Sofia

    by Ryota Karube, Lidia Lazarova Vitanova, Evgeny Shirinyan, Takuto Sato, Dessislava Petrova-Antonova, Hiroyuki Kusaka

    Urban areas are increasingly exposed to thermal stress owing to the combined effects of global warming and the Urban Heat Island (UHI). High-resolution numerical modelling is therefore needed to assess local thermal environments and support heat-mitigation planning. This study applies a high-resolution WRF–City-LES downscaling approach to simulate the urban thermal environment of Sofia, Bulgaria. The mesoscale Weather Research and Forecasting (WRF) model provides regional- to urban-scale meteorological forcing, whereas the microscale City-LES model resolves city-block-scale thermal and wind environments in the Lozenets district by explicitly representing buildings, trees, land-surface characteristics, and building-related anthropogenic heat (AH). Compared with the mesoscale WRF simulation, City-LES more accurately represents the near-surface air temperature at the observation sites, reducing the mean absolute error (MAE) from 1.4 °C to 0.3 °C and the root mean square error (RMSE) from 1.4 °C to 0.3 °C. The high-resolution micro-simulation also captures fine-scale spatial variability in near-surface air temperature and outdoor thermal comfort that WRF did not resolve. The spatial distributions of Mean Radiant Temperature (MRT) and Universal Thermal Climate Index (UTCI) reveal city-block-scale differences related to buildings, tree canopies, and radiative conditions. The AH sensitivity experiment showed that building-related AH had a limited effect on domain-averaged near-surface air temperature under the present implementation. These results demonstrate that the WRF–City-LES downscaling approach provides a useful basis for urban climate assessment and future urban digital twin applications, including scenario-based evaluation of heat-mitigation strategies.

  • Open Access

    User-Centric Pedestrian Navigation: Transfer Study from Sofia to Tokyo

    by Ivan Mitev, George Taylor, Kaloyan Karamitov, Lidia Lazarova Vitanova, Dessislava Petrova-Antonova, Anna Yokokubo, Yusuke Sasaki, Takeo Hamada, Yoshio Ishiguro, Noboru Koshizuka, et al.

    Pedestrian route-finding applications are commonly used for urban navigation. These applications typically focus on finding the shortest or fastest route and rarely consider pedestrian safety or comfort. Such non-utilitarian factors can be accounted for by incorporating a walkability index in the routing algorithm. This study transfers an existing user-centric pedestrian routing approach from Sofia to the Tokyo wards of Shinjuku and Bunkyo. The approach considers the heterogenous needs of different user groups by using an adjustable weighting scheme. Data from OpenStreetMap and local open datasets on streetscape features such as sidewalk width and incline are standardised and walkability scores are calculated using the developed framework. Results indicate that the Sofia routing approach is cleanly transferable to Tokyo. Wasserstein’s, Cliff’s delta, and Kolmogorov-Smirnov’s statistical tests show that calculated walkability scores remain consistent with the intentions of the framework. Calculated routes in Tokyo successfully increase average walkability compared to the shortest path, without excessively increasing length. Walkability scores are unevenly distributed in the study area, with street segments in pedestrianised residential areas with many parks receiving high scores, while major roads that lack dedicated pedestrian infrastructure receive lower scores. The different user profiles lead to spatially distinct walkability patterns. Scores for tourists are high around densely concentrated commercial facilities, while scores for wheelchair users are high along arterial roads with wide sidewalks and low incline. By successfully adapting an existing pedestrian routing method, this study demonstrates that using user-specific walkability scores for route-finding can be a viable approach in different contexts.

  • Open Access

    Agentic Urban Planning Assistant: A Parametric Approach to Spatial and Regulatory Evaluation in Urban Planning

    by Emil Hristov, Angel Spasov, Denis Dimitrov, Dessislava Petrova-Antonova

    Urban planning decision support often requires a joint interpretation of legal rules, parcel geometry, zoning parameters, street context, and proposal-specific constraints. Generic retrieval-augmented generation is insufficient for this setting since many answers depend on structured spatial evidence and deterministic regulatory calculations rather than on retrieved prose alone. This paper presents an agentic urban planning assistant designed for parcel-level first-pass analysis. It combines deterministic request classification, route-specific retrieval, parcel lookup, exact-index zoning-parameter resolution, parcel-side normalization, setback-rule assignment, gated planning calculations, and final large-language-model (LLM) answer synthesis. The implementation uses the LLM mainly for the final answer generation stage, while the evidence and reasoning foundation is constructed through structured data preparation and deterministic workflow logic. An evaluation framework is developed that aligns benchmark design and metric selection with the actual assistant’s route structure. It combines deterministic checks, complementary LLM-as-a-judge metrics, benchmark generation from reviewed seed cases, and execution-level observability. A reviewed 40-case architecture-aware benchmark for the Sofia/Bulgarian case study is prepared. On this benchmark, the deterministic pass rate reached 1.00, while the complementary judge-based scoring yielded a faithfulness value of 0.931 under a revised structured-evidence evaluation setup. The remaining weaknesses are concentrated in proposal-feasibility synthesis and parcel-zoning descriptive completeness.

  • Open Access

    Urban Heat-Aware Walkability Index Based on LLMs Analysis: A Case Study of Sofia

    by Denis Dimitrov, Lidia Lazarova Vitanova, Dessislava Petrova-Antonova

    A novel Artificial Intelligence (AI)-driven urban heat-aware walkability index (UHAWI) is proposed to assess pedestrian wellbeing and heat exposure. The approach integrates a conventional walkability index with the Wet Bulb Globe Temperature (WBGT) to assess how thermal stress affects the suitability of urban streets and areas for walking. AI methods are used to analyse spatial patterns and identify thermally comfortable and low-risk pedestrian zones. A Large Language Model (LLM) is used to analyse the calculated index and suggest an intervention. The pipeline is applied to street segments in Sofia, Bulgaria, across 5 consecutive July dates, producing UHAWI scores ranging from 23.1 to 66.5 (mean 43.9), which are then normalised for visualisation, indicating that summer thermal stress binds pedestrian conditions city-wide. Street segments with critical heat-walkability deficits are identified and targeted interventions are suggested, each grouped by empirically documented cooling and walkability effects, with street tree planting yielding the largest mean improvement (+13.3 points). Rubric-based evaluation of the LLM recommendations by a human (overall mean 3.25/5) and an independent LLM-as-judge (2.75/5) found strong spatial coherence and actionability but limited numerical faithfulness, identifying output grounding as the principal limitation of small locally deployed models. The proposed approach offers municipalities a reproducible, privacy-preserving decision support tool for prioritising climate adaptation investment at the street level, and the compute-then-reason architecture generalises to other urban analytics domains where AI assistance must remain locally executable and fully auditable.

  • Open Access

    Future Building Energy Atlas of Sofia City

    by Lidia Lazarova Vitanova, Darin Pavlov, Tereza K. Trendafilova, Ryota Karube, Koki Nakamura, Agata Zabuska, Quang-Van Doan, Dessislava Petrova-Antonova, Sylvia Ilieva

    This study develops a Future Building Energy Atlas (FBEA) for Sofia, Bulgaria, integrating land-use change modelling and spatial energy-demand assessment within a GIS framework. Future land-use patterns toward 2074 are simulated using a hybrid Artificial Neural Network–Cellular Automata (CA-ANN) model, while building energy consumption is estimated using cadastral data, gross floor area calculations, and energy coefficients derived from more than 2,500 certified buildings across ten functional categories under the medium-energy consumption. Future energy demand is projected based on historical trends observed between 2000 and 2020, during which building energy consumption increased steadily. The results indicate continued urban expansion primarily at the expense of cropland and forest areas. Cropland is projected to decline from 36.4 km² to 26.5 km², while forest areas are projected to decrease from 16.7 km² to 14.4 km² by 2074. At the same time, building energy demand increases significantly. The daily mean maximum energy consumption is projected to increase from 35.2 W/m² in 2024 to approximately 115 W/m² by 2074. These results highlight the strong relationship between urban growth, land-use transformation, and increasing energy demand in the building sector.

  • Open Access

    AI-Empowered Urban Digital Twins: Integrating Geospatial and Legal Data for Spatial Planning and Governance

    by Mila Koeva, Dessislava Petrova-Antonova, Sisi Zlatanova, Rob Peters, Claudio Persello, Ville Lehtola, Pirouz Nourian, Johannes Flacke, Raoul Grouls, Alfred Stein

    Urban planning increasingly requires accessible ways to connect regulatory frameworks with clear spatial visualisation and machine-assisted guidance. However, the integration of legal and geospatial information into platforms that support decision-making, communication, and participation remains limited. In particular, local authorities lack an operational way to translate legal and regulatory provisions into spatially explicit permit systems, and existing Urban Digital Twin architectures rarely formalise this legal-geospatial link. This paper uses Urban Digital Twins (UDTs) and Artificial Intelligence (AI), with a focus on generative AI, for smart, participatory, and responsible spatial planning. It proposes a conceptual framework for an AI-empowered Urban Digital Twin that integrates legal and 2D/3D geospatial data in both static and dynamic forms. Its aim is to support decision-making, information provision, and participation in spatial planning processes. The framework comprises five layers: data acquisition and harmonisation, data modelling, analytics and simulation, visualisation and interaction, and governance and decision support. Each of them are supported by AI models, and is illustrated through four case studies developed with the Province of Utrecht. The framework is supported by theoretical and methodological reflections and by experimental case studies. We conclude that integrating legal and geospatial data through AI-empowered UDTs strengthens spatial planning practices, improves regulatory interpretation, supports governance processes, and enables more informed and transparent decisions, provided that current limitations in AI reliability, explainability, and human oversight are explicitly addressed.

  • Open Access

    LLM-AQRA: Leveraging LLMs for Efficient Urban Air Quality Research

    by Shanshan Jiang, Hanan Waheed, Dessislava Petrova-Antonova, Dumitru Roman

    Urban air quality research increasingly relies on integrating heterogeneous data from sources such as monitoring stations, satellite imagery, community sensors, built-environment datasets, and meteorological forecasts. However, researchers often face challenges in discovering, understanding, and integrating these datasets due to inconsistent metadata, mismatched spatial and temporal resolutions, diverse formats, and varying semantics. Moreover, effective use of air quality analytical models typically requires specialized expertise to interpret model requirements and identify suitable input datasets, leading to labor-intensive, inefficient workflows. This paper introduces LLM-AQRA, a prototype Large Language Model (LLM)-driven Air Quality Research Assistant designed to streamline these processes by automatically generating standard-compliant metadata, extracting model requirements, assessing dataset-model compatibility, and supporting map-based exploration of heterogeneous datasets. Initial user feedback indicates that LLM-AQRA is perceived as useful for dataset understanding and has the potential to reduce manual effort in analytical workflows. Future work will focus on retrieval augmented generation, ontology-guided constraints, expanded domain support, and broader user evaluations to improve reliability and applicability for real-world environmental analysis.

  • Open Access

    A Workflow for Interactive Visualization of Urban CFD Data in Unreal Engine and via Web Streaming

    by Orfeas Eleftheriou, Vasilis Naserentin, Mariya Pantusheva, Dessislava Petrova-Antonova, Anders Logg

    To analyze current and predict future conditions, smart-city planning relies on physics-based simulations to support sustainable design, climate adaptation, and citizen engagement. Yet, the analysis results often remain locked inside specialist tools and static plots that are difficult for non-experts to access and interpret. As part of a broader framework developed for scientific results visualization in an urban context, this paper presents a semi-automated workflow for transforming urban Computational Fluid Dynamics (CFD) outputs into interactive 3D visualizations using Unreal Engine and a web-based streaming application. The feasibility of the workflow is demonstrated through a wind-flow case study of an area in central Sofia, Bulgaria. The proposed approach covers the complete process, from problem formulation and data preparation through simulation execution, results processing, and integration into Unreal Engine. The visualization component is implemented through a modular actor-based architecture that supports both point-based and trajectory-based flow representations, which can be flexibly combined and interactively modified within the scene. Additionally, the visualization system is deployed through Pixel Streaming, providing a practical and viable way to transform complex CFD data into exploratory tools that are accessible to a wide range of urban stakeholders.

  • Open Access

    Privacy-Preserving Smart-City Data Sharing for Urban Environmental Analytics: A k -Anonymity Case Study

    by Abubakari Alidu, Flavio De Paoli, Dessislava Petrova-Antonova, Emil Hristov, Evgeny Shirinyan

    Urban analytics increasingly depends on combining sensitive municipal administrative records with heterogeneous sensing data, yet such integration is often constrained by privacy, governance, and data-sharing rules. This paper asks whether a municipality can report a monthly district-level ambient particulate-matter exposure proxy for an enrolled kindergarten cohort without sharing raw child-level records. The case study links kindergarten enrolment data in Sofia with citizen-sensed PM 2.5 and PM 10 measurements through a three-pipeline workflow: municipality-side privacy processing, air-quality curation, and consumer-side analytics using only a released cohort and public district centroids. Air-quality data alone can describe ambient pollution, but the cohort release is required to define which children, age ranges, kindergartens, and districts are represented in the reporting population. For January 2025, the release contains 3,440 children and satisfies k-anonymity at k = 12 with respect to district, age range, and kindergarten, with no violating equivalence classes. Across k ∈ {5, 8, 12, 16, 20}, retention decreases from 97.5% to 87.3%, and represented districts decrease from 24 to 22. On the overlapping district set, district mean PM 2.5 summaries remain highly stable relative to the k = 5 release, with Spearman ρ ≥ 0.988 for k ≥ 8 and MAE below 0.3 μg/m 3 . In a geocoding diagnostic, 45.9% of sampled administrative addresses were resolved under the evaluated protocol, motivating centroid linkage as a coverage-versus-precision design choice. The findings are bounded to one city, one reporting month, and a district-scale reporting task.