Advancing generative AI for enhanced analytics in urban and environmental monitoring

Kazemi, Amir

Permalink

https://hdl.handle.net/2142/127495 Copy

Description

Title

Advancing generative AI for enhanced analytics in urban and environmental monitoring

Author(s)

Kazemi, Amir

Issue Date

2024-12-05

Director of Research (if dissertation) or Advisor (if thesis)

• Tessum, Christopher
• Kindratenko, Volodymyr

Doctoral Committee Chair(s)

Tessum, Christopher

Committee Member(s)

• Salapaka, Srinivasa M
• Guest, Jeremy S

Department of Study

Civil & Environmental Eng

Discipline

Civil Engineering

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• Smart Cities
• Intelligent Transportation Systems
• Urban Sensing
• Machine Learning
• Unsupervised Learning
• Supervised Learning
• Transfer Learning
• Optimal Transport
• Computer Vision
• Operator Learning

Abstract

Generative AI, notably through optimal transport and denoising diffusion models (DDM), is enhancing the synthesis and analysis of data in urban and environmental monitoring systems. This technology is crucial for creating high-fidelity simulations and models that predict dynamics, supporting the development of resilient infrastructures and optimized processes in the face of operational and environmental challenges. Optimal transport theory is pivotal in learning traffic data distributions and enhancing urban surveillance capabilities. The novel Iterative Generative Adversarial Networks for Imputation (IGANI) architecture introduces a new data-completion approach using an iterative GAN leveraged by optimal transport. Additionally, optimal transport theory is applied to rigorously evaluate one-shot generative models for data augmentation, focusing on radiofrequency (RF) signal patterns in proliferating Unmanned Aerial Vehicles (UAV). Such one-shot generative distribution matching enhances RF-based UAV identification in data-limited environments, enabling more reliable urban surveillance. DDMs have also emerged as powerful generative tools in computer vision and scientific machine learning. Prompt-guided outpainting DDMs are explored to generate contextually rich images for traffic monitoring and expand datasets with automatic annotations. AI-generated Dataset of Outpainted Vehicles for Eye-level Classification and Localization (AIDOVECL) reduces the need for manual labeling and improves performance metrics for vehicle detection. Moreover, a novel approach has been proposed to capture the dynamics of physical processes using DDMs. This new DDM training and sampling algorithm demonstrates its capability to simulate phenomena such as heat distribution, presenting the potential for predictive analytics in environmental planning and management. Overall, this dissertation illustrates how generative AI, mainly through optimal transport and DDMs, may address data scarcity and improve the fidelity and utility of synthetic data in urban and environmental monitoring. The implementations presented highlight potential in data-driven decision-making within infrastructure systems for real-world scenarios. This work contributes insights and methodologies that could inspire further research and practical applications in infrastructure systems and the environment.

Graduation Semester

2024-12

Type of Resource

Thesis

Handle URL

https://hdl.handle.net/2142/127495

Copyright and License Information

Copyright 2024 Amir Kazemi

Owning Collections