Enhancing construction safety through Data Science and Machine Learning
Jun 01, 2025
14:58:38
Dr. Donie Jardeleza, a researcher and faculty member at BUV, is advancing structural engineering with innovative predictive modelling techniques. With expertise in Data Science, Machine Learning, and IoT, his research improves the accuracy of concrete compressive strength predictions – key to building safer and more durable infrastructure.
By combining advanced data analytics with civil engineering, Dr Jardeleza develops smarter methods for assessing construction materials. His latest work integrates K-Medoids clustering and Quantile Regression Forests (QRF) – two machine learning techniques that enhance prediction accuracy while minimising data inconsistencies.
In his work Predictive Modeling of Compressive Strength composition values for structural studies using K-medoids clustering and Quantile Regression Forests, he introduces an innovative approach that integrates K-medoids clustering with Quantile Regression Forests (QRF) to minimize noise, enhance accuracy, and provide robust prediction intervals. Unlike conventional clustering techniques, K-medoids reduces the impact of outliers, offering more stable categorization of concrete mixtures. Simultaneously, QRF ensures precise forecasting of compressive strength values within a 90% confidence interval. This methodology has paved the way for more reliable and efficient predictive modeling in structural engineering.
In addition, his work on Ubiquitous Robot for Structural Studies: Comparative Analysis on Concrete Crack Images using Feature Detection and Matching advances structural assessment technologies by developing a Level 4 autonomous robotic system designed for inspecting concrete structures. This research addresses the growing demand for automated solutions in long-term structural monitoring. By leveraging feature detection and matching algorithms, the robot analyzes concrete crack images, identifying key points between real-time data and pre-trained datasets. Through extensive experimentation, he demonstrated the system’s ability to process images under various transformations, including scaling, rotation, and skewing. This innovation has set a new benchmark for unmanned structural inspections, reducing reliance on manual labor while significantly improving efficiency and accuracy in evaluating structural integrity.
Dr Jardeleza’s work exemplifies BUV’s commitment to cutting-edge research and real-world impact, demonstrating how data-driven approaches can revolutionise traditional engineering practices. Through his contributions, he is not only advancing academic knowledge but also shaping the future of sustainable and safe construction.