IoT and machine learning framework for smart emergency response systems in Nigerian cities: A case study of Adamawa state, Nigeria

Existing Emergency Response Systems (ERS) grapple with the immense challenge of managing crises in rapidly growing cities across developing nations, largely because of inadequate infrastructure. This paper demonstrates a proposed Internet of Things (IoT) integrated with Machine Learning (ML) backed smart emergency response system for urban cities in Nigeria using the state of Adamawa as a reference point. Parameterized in the attached guide, the scope of this paper does not include claims to physical hardware deployment. Instead, using simulated data points and feedback from stakeholders, we analyze our provided framework. This research used a compiled dataset of 1200 emergency records. These were able to be cross referenced with localized metrics analyzed in the results and conclusion with the software we simulated. Analyses of Random Forest (RF), Support Vector Machine (SVM), and K- Means were used to determine how well we can predict classifications of incidents to aid in better resource management. Parameters for the simulation assume locally deployed sensors and established comms parameters which would be most optimal for low bandwidth regional restrictions. There was significant evidence to suggest improvements in response time from initial report. Our smart city prototype provides a starting point for future production smart cities in sub-Saharan region.