GeoSmart cities: An event-driven approach for processing big geodata in smart cities

This thesis examines the transformation of smart cities from passive data providers to active information and service providers, leveraging sensor data to address complex urban challenges and improve productivity, efficiency, and quality of life. By exploiting the spatio–temporal properties of sensor data, the research develops mechanisms for real–time decision-making applications in smart cities.
The study identifies two primary challenges: conceptual challenges and technical challenges. Conceptual challenges relate to aligning the development of information services with the broader goals and ideals of smart cities, including sustainability, citizen well-being, and stakeholder collaboration. Technical challenges involve integrating heterogeneous sensors, ensuring efficient data management, and enabling the real-time processing of large volumes of sensor data while maintaining scalability and reliability.
To address these challenges, the research proposes a two-pronged solution. First, it introduces a framework for the co-design of event-driven applications, enabling collaboration among developers, stakeholders, and city authorities to conceptualize applications that reuse sensor data for actionable insights. Event-driven applications focus on detecting geographic events—specific occurrences with spatial and temporal properties—to provide situational awareness and support real-time decision-making. The framework includes definitions, workflows, and a co-design matrix to systematically document the conceptualization process.
The framework's applicability is demonstrated through three use cases addressing common urban concerns:
1. Air Pollution Monitoring: Utilizing sensor data to alert citizens of potential exposure risks and enabling informed decisions to protect public health.
2. Ambient-Aware Running: Combining air quality and wearable sensor data to promote outdoor activities while minimizing exposure to pollutants and unfavorable weather conditions.
3. Responsive Commute Services: Enhancing public transportation efficiency and commuter experiences by integrating data from fixed and mobile sensors.
Second, the thesis introduces a Reference Architecture for Smart City Applications (RASCA) to address the technical challenges of implementing event-driven applications. RASCA combines microservices and event-driven architectures to process sensor data streams efficiently, enabling the detection of geographic events in real time. The architecture incorporates complex event processing (CEP) and spatio–temporal analytics, providing capabilities to integrate heterogeneous sensor networks and manage high-throughput data streams. Prototypes based on RASCA, such as GEDSys, were tested using real-world sensor data from the SmartSantander and Algemeen Luchtmeetnet sensor networks. The prototypes demonstrated the feasibility, scalability, and performance of the architecture, highlighting its potential for deployment in decentralized computing environments. Challenges such as potential failure points were identified, emphasizing the need for reliability improvements in future implementations.
A critical innovation in this thesis is the development of a domain-specific language (GeDL) for defining geographic events in event-driven applications. GeDL provides a formal and expressive mechanism to specify spatial, temporal, and attributive relationships, addressing the limitations of existing event processing engines, which lack spatial matching capabilities. A prototype implementation of GeDL was integrated into GEDSys and evaluated using real-world sensor data, demonstrating its effectiveness in declaring geographic events for various applications. While GeDL shows promise as a concise and powerful language, further refinements are recommended to enhance usability, including adding constructs for common temporal patterns and importing event definitions.
In summary, this thesis offers a comprehensive approach to unlocking the potential of sensor data in smart cities. It addresses critical challenges in data integration, real-time processing, and geographic event detection, providing a framework (GeoSmart City Framework), an architecture (RASCA), and a language (GeDL) for developing scalable and high-performance event-driven applications. These contributions advance the fields of smart city technology and geographic information science by enabling the effective use of sensor data for real-time urban management, promoting collaboration among stakeholders, and fostering sustainable and intelligent urban development.