Evacuation of coastal communities, particularly in rural areas, can be a challenge due to the topography, dispersed residential and population patterns, and limited number of roadways that lead further inland. This Smart and Connected Community Planning project will assess the viability of leveraging human knowledge and social media data with an artificial intelligence (AI) system to create a human-AI teaming (HAT) paradigm addressing flood evacuation decision making in isolated coastal rural communities. In the process of creating a HAT paradigm, we will integrate transportation network data with river information as well as volunteers' observations and social media data to leverage the strengths of local members of a Community Emergency Response Team (CERT) with AI. This Planning project advances the field by conducting a feasibility study of this HAT decision-making tool, which will be tested through real-world flood evacuation examples in Charleston, Berkeley, and Dorchester Counties located in the Lowcountry region of South Carolina (SC). Additionally, we will determine the barriers and motivations for understanding the usability of the researched HAT decision-making tool using qualitative (interview) protocols. This project supports education and diversity by providing research experiences to diverse students, as well as focusing on vulnerable, rural communities. Additionally, this planning project supports NSF's mission to promote the progress of science and to advance the nation's health, prosperity, and welfare by seeking to (i) enhance flood evacuation by automatic data-driven decision making and (ii) identify potential barriers to the adoption of technology in rural volunteer communities.

This research will co-create and implement a pilot solution for flood evacuation decision making by including systems thinking, human–machine engagement, human development training, and AI-driven decision making. The focus is on the advancement of flood evacuation techniques by transferring the traditional determination techniques (the expert evaluation approach) toward new and coherent HAT computing. By leveraging HAT protocols and applying them to a flood evacuation decision-making tool, our project has the potential to be transformative. Tackling these issues will enable us to harness the full potential of AI as a partner in emergency management and response. This planning project brings together researchers from water resources engineering, social sciences and communication, transportation engineering, disaster science, computer science, and numerous volunteers and stakeholders to co-create solutions, build/strengthen collaboration with key stakeholders and CERT organizations and identify potential barriers to technology use in rural communities challenged by a higher incidence of flood hazards and substandard infrastructure.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

Vidya Samadi
Dr. Samadi is trained as a water resource engineer and works to advance the field of hydroinformatics and cyber-physical modeling systems. Much of her current research is focused on machine learning applications in flood/stormwater computing and decisions making, impacts of flooding on critical infrastructure, smart cities and infrastructure, big data analytics, and Geographic Information systems (GIS). Dr. Samadi's team has developed many packages and modeling systems including Flood Analytics Information System (FAIS), FAIRDNN, Flood Image Classifier, and Watershed Toolkit, as well as physical-based models.
Performance Period: 10/01/2021 - 09/30/2024
Institution: Clemson University
Award Number: 2125283