TLDR: This paper introduces a multidimensional, data-driven, and AI-assisted framework for assessing urban comfort, defining it beyond a single metric to include environmental and social aspects. It outlines a four-step process: comfort surveys, model development, scaling predictions, and regional validation, significantly enhanced by AI technologies like CNNs and DNNs for data processing and prediction. A case study in Singapore demonstrates its application for mapping urban thermal comfort using street view imagery, providing valuable insights for urban planning while also highlighting areas for future research, such as expanding to a fully multidimensional comfort model.
Ensuring cities are liveable and comfortable for their residents is a core goal of urban planning. While many studies have looked at factors like greenery, thermal conditions, and walkability, a clear and comprehensive definition of urban comfort, along with a robust evaluation framework, has remained elusive. A recent research paper, Urban Comfort Assessment in the Era of Digital Planning: A Multidimensional, Data-driven, and AI-assisted Framework, addresses this gap by proposing a new framework that emphasizes multidimensional analysis, data support, and AI assistance.
Defining Urban Comfort
The paper clarifies that urban comfort is not a simple, single measure but a complex, multidimensional concept. It’s defined as the degree of satisfaction a person has with their environment, specifically within an urban setting. Unlike general comfort, which can be individual, urban comfort is a collective experience shaped by how environmental attributes interact with human perception. This means it cannot be reduced to a single numerical index, as it varies greatly from person to person based on their preferences and experiences.
The framework categorizes urban comfort into two main dimensions:
-
Environmental Comfort: This refers to people’s satisfaction with the physical aspects of the built environment, including thermal comfort (how hot or cold it feels), visual comfort (what people see), and acoustic comfort (what people hear).
-
Social Comfort: This relates to how well the urban environment supports human interactions and behavioral needs, focusing on the functionality of spaces for activities like walking (walkability), ease of access (accessibility), and feelings of safety.
The researchers propose that urban comfort should be understood as a vector, capturing these multiple dimensions rather than aggregating them into a single value. This approach acknowledges that different comfort dimensions interact and contribute uniquely to overall perception depending on the urban context and individual preferences.
A Four-Step Assessment Framework
To effectively evaluate this multidimensional urban comfort, the paper introduces a structured, data-driven assessment framework with four key steps:
-
Conducting Comfort Surveys: This involves gathering data on both geospatial characteristics of an urban location and human comfort perceptions.
-
Developing Comfort Models: Using the collected data, models are built to understand the complex, non-linear relationships between the urban environment and human comfort.
-
Scaling Up Urban Comfort Predictions: Once models are developed, they can be applied to broader urban areas using extensive geospatial data to predict comfort levels across an entire city.
-
Performing Regional Validation: The comfort models are then validated using additional survey data from distinct locations to confirm their reliability and accuracy.
The Role of AI in Comfort Modelling
Artificial intelligence plays a crucial role in enhancing this data-driven framework, offering significant advantages over traditional methods. AI models excel in handling high-dimensional data, modeling non-linear relationships, and providing scalability and generalizability across different urban areas. For instance:
-
Convolutional Neural Networks (CNN): These are highly effective for extracting features from image data, such as satellite and street view imagery, which are vital for representing urban environments.
-
Deep Neural Networks (DNN): These models are robust for predicting complex, non-linear problems like human comfort and can be applied to untrained data, allowing comfort models developed for one area to be extended to others.
-
Long Short-Term Memory Networks (LSTM): These are ideal for analyzing time-series data, making them valuable for understanding dynamic urban observations over time.
Integrated AI technologies can significantly improve comfort modeling and mapping, helping urban planners identify areas needing improvement, highlight critical environmental elements, and inform decisions through continuous geospatial analysis.
Case Study: Thermal Comfort in Singapore
To demonstrate the framework, the researchers applied it to assess urban thermal comfort in Singapore, a high-density tropical environment. Instead of traditional in-field surveys, they used an online approach leveraging street view images and AI-based CNN feature extraction for the initial comfort survey phase. This method is more efficient, reducing the need for extensive human and financial resources, though in-field surveys were still conducted for model validation.
Two types of models were developed: a Deep Neural Network (DNN) for predicting thermal comfort levels across 92,233 street view images in Singapore, and a linear regression-based inference model to identify factors significantly influencing thermal comfort. The outcome was a detailed map of Singapore’s urban thermal comfort, pinpointing areas with higher and lower comfort levels, which can guide urban renewal and planning efforts.
Also Read:
- Enhancing Urban Mobility Simulations with AI: The Preference Chain Approach
- Artificial Intelligence Revolutionizes Global Weather Forecasting, Offering Rapid and Accessible Climate Insights
Contributions and Future Directions
This research provides a valuable multidimensional framework for understanding urban comfort, outlines a four-step assessment process, and highlights the supportive role of AI in digital urban planning. The Singapore case study successfully demonstrates the application of this framework for thermal comfort analysis. However, the study acknowledges limitations, such as not fully exploring the diverse auxiliary roles of various AI technologies and focusing only on thermal comfort rather than a fully multidimensional comfort model. Future research will aim to address these limitations by incorporating more AI techniques and developing a comprehensive multidimensional comfort model.
