The evolution of human civilizations is closely tied to the advancement of infrastructure, a relationship that has become increasingly evident with the rapid urbanization of recent years. As we stand on the brink of an era of ‘smart cities,’ driven by technologies such as artificial intelligence (AI), the Internet of Things, and big data analytics, the promise of sustainable urban development has never been greater. However, the progress toward these smart cities is being challenged by the growing impacts of climate change.

Natural disasters, particularly earthquakes, pose significant threats to buildings and infrastructure. A prime example of this is soil liquefaction—a hazardous phenomenon where saturated soil loses its strength and rigidity under stress, often due to earthquake-induced shaking or sudden loading. When this occurs, the soil behaves like a liquid, compromising its ability to support structures and creating substantial risks for urban environments.

To address this critical challenge, researchers at the Shibaura Institute of Technology in Japan have developed an advanced AI predictive model designed to generate comprehensive soil liquefaction risk maps. Led by Professor Shinya Inazumi, along with team members Arisa Katsuumi and Yuxin Cong, the study integrates state-of-the-art machine learning techniques with essential geotechnical and geographical data. This innovative model has already been applied to optimize urban planning and infrastructure development in Yokohama, Japan—a city that faces unique soil liquefaction challenges due to its extensive reclaimed lands and frequent seismic activity.

“We were motivated to pursue this research after recognizing the urgent need to improve urban resilience to earthquakes, particularly in rapidly urbanizing areas prone to seismic activity,” said Prof. Inazumi. “There are critical weaknesses in existing geotechnical risk assessments and urban planning strategies. Traditional methods for predicting soil liquefaction are often limited by the scale of data integration and the speed of analysis, leading to gaps in emergency preparedness and risk management. By leveraging advanced technologies such as AI and machine learning, we aimed to develop a more dynamic and accurate predictive model.”

The AI-driven model utilized a combination of advanced machine learning techniques, including artificial neural networks and gradient-boosting decision trees, significantly enhancing the precision of soil liquefaction risk predictions. The researchers achieved exceptional accuracy in forecasting soil classifications and N-values, both crucial for assessing soil liquefaction risk. The model’s effectiveness was validated through rigorous testing against extensive geotechnical survey data.

Prof. Inazumi further explained, “The real-world application of our research is the development of hazard maps, which help urban planners and engineers visualize and identify areas at high risk for soil liquefaction, enabling informed decisions about infrastructure development. Beyond bolstering emergency response planning, this AI-driven approach can also facilitate community engagement and education by providing clear and accessible information about at-risk areas.”

This research highlights significant advancements in geotechnical engineering through the integration of AI technology in predicting soil liquefaction risks. By strengthening initiatives aimed at enhancing urban resilience and sustainability, this innovative approach offers a vital tool for the smart cities of the future.

By Impact Lab