Disaster Management through Integrative AI
Swarna Kamal Paul, Parama Bhaumik · 2022
Artificial Intelligence (AI) methods when integrated with IoT devices can play key role in prevention, response and recovery phase of disaster management. IoT devices can collect real time data for various events which can be useful for prediction and monitoring of disastrous events using AI methods. However, deploying an application for a specific scenario is highly contextual and development requires a large engineering effort with low reusability. In many scenarios, such approach becomes infeasible due to long development life cycle. Thus, an integrative AI platform has been designed to integrate multiple AI components, IoT devices and data sources in the form of microservices using an easy-to-use dataflow graph-based programming model. The programming model is pre-built with essential primitive functions to implement any arbitrary programming logic. Treating different components as microservices enhances reusability and applications can be designed by integrating components in loosely coupled fashion. Two use cases have been discussed under disaster prevention category and showcased prototype solutions in the proposed integrative AI platform.