TalaNivo : Une infrastructure IoT résiliente pour la surveillance des crues et l'alerte multimodale en milieu urbain : Cas de Kinshasa.

Abstract

Urban flooding is a major and recurring hazard in the city-province of Kinshasa, where the 2024 historical flood (6.26 m) caused over 43 deaths and displaced hundreds of households within a single week. The primary limitation of existing early warning systems (EWS) lies in their dependency on unstable internet infrastructure and their inadequacy given the digital divide prevalent in low-resource contexts. This paper presents TalaNivo, an IoT system featuring a hybrid architecture combining: (i) local ultrasonic measurement (accuracy ±0.3 cm over 0–4 m), (ii) real-time Cloud transmission via NodeMCU ESP8266, (iii) resilient SMS alerting via GSM/SIM800L module (latency < 10 s), and (iv) an acoustic buzzer providing universal coverage without technological prerequisites. A signal processing pipeline based on an exponential moving average filter reduces measurement error rates by 92%. Experimental results demonstrate a 30–45 minute reduction in alert delay compared to manual observation methods. TalaNivo thus constitutes a reproducible, low-cost technological foundation for African smart cities facing similar infrastructure constraints.