<![CDATA[Urban flooding remains a disastrous challenge for rapidly expanding cities in developing nations. Despite the fact deep learning models and IoT sensing are individually established in hydrology, their seamless integration into a unified, cost-effective Cyber-Physical System (CPS) specifically architected for data-scarce and infrastructure challenged environments remains a critical research gap. This research contributes a novel, end-to-end framework that bridges this divide by harmonizing three distinct pillars: a low-cost, energy-autonomous IoT sensor network, a hybrid CNN-LSTM predictive model, and a dynamic geospatial visualization dashboard. Unlike conventional systems designed for data-rich environments, our framework is contextually adapted for the unique topographical and socio-technical realities of Nigerian urban centers. Validated through a six-month deployment in the high-density Ajeromi-Ifelodun region of Lagos, the system achieved a Nash-Sutcliffe Efficiency (NSE) of 0.89 and a critical 4.5-hour forecast lead time.]]>
Copyrights © 2026
