AI-driven Urban Growth Prediction: A Cellular Automata Approach with Geospatial and Machine Learning Integration

Abstract

Urban expansion is a critical challenge in rapidly growing cities, necessitating accurate prediction models for sustainable planning. This study employs a cellular automaton (CA) based machine learning model approach to predict urban growth in Jaipur city using Landsat 8 OLI data from 2011-2021. The analysis incorporates multiple spatial drivers including spatial factors for efficiently projecting the urban growth. The study revealed a 28.4% increase in built-up areas from 2011-2021 predominantly replacing green and barren lands. The transition schematics confirms that urbanization follows major road corridors while green spaces and water bodies accounting to its growth. Using CA technique, projected 2031 land use land cover (LULC) indicates continued outward expansion particularly in Jaipur’s western and northern peripheries. The classification accuracy of LULC maps for 2011 and 2021 is 85.6% and 88.3% respectively with a kappa coefficient of 0.81. The CA-based prediction achieved a model accuracy of 0.89 validating its reliability. The findings highlight the urgent necessity for strategic zoning regulations, green infrastructure and transport-oriented development to ensure balance and systematic urban growth. This research serves as a valuable decision-support for urban planners, offering insights into future land-use changes and providing a foundation for sustainable urban policy formulation.