AI system for real time flood detection
Addressing Flood-Related Roadway Incidents
Current Challenges in Flood-Reporting Tools
Scarcity of Real-Time Tools
Flood-related roadway incidents rank among the top causes of fatalities nationwide, yet the scarcity of real-time flood-reporting tools hampers accurate road condition assessments.
Limitations of Existing Tools
Current tools, including traffic cameras, water sensors, and social media feeds, can detect flooding, yet they aren't optimized for road-specific conditions and typically operate independently. A coordinated sensors network could improve flood level monitoring, but the expense of large-scale deployment is prohibitive.
Introduction of OpenSafe Fusion
Overview of the System
Rice University engineers have introduced a potential remedy: OpenSafe Fusion, an automated data fusion system. Officially termed Open Source Situational Awareness Framework for Mobility using Data Fusion, OpenSafe Fusion harnesses existing reporting tools and public data to swiftly detect road conditions during the increasingly frequent urban flooding events.
Development and Collaboration
Jamie Padgett, the Stanley C. Moore Professor and chair of Rice's Department of Civil and Environmental Engineering, collaborated with postdoctoral researcher Pranavesh Panakkal to analyze data from nine sources in Houston, leading to the development of their automated data system framework.
Study and Publication
Research Details
The study, "More Eyes on the Road: Sensing Flooded Roads by Fusing Real-Time Observations from Public Data Sources," is featured in the journal Reliability Engineering & System Safety.
Theoretical Insights
"While there are few sources that directly monitor flooded roads, urban centers have an abundance of systems that observe flooding and road conditions, either directly or indirectly," said Padgett.
The researchers, Padgett and Panakkal, theorized that leveraging an automated system to unify these real-time insights could revolutionize flood awareness, eliminating the need for extensive new sensor deployments.
Framework Application and Testing
Integration of Data Sources
According to Padgett, this research presents a method for communities to effectively monitor and respond to urban stressors, including flooding, by leveraging current data sources.
Our work extends from our ongoing collaboration with the SSPEED Center at Rice, Which has been at the forefront of developing sophisticated flood alert systems, and concentrates on the impact of floods on transportaion infrastructure, aiming to enhance flood models with complementary data, especially concerning roadways and mobility.
The framework integrates data from traffic alerts, cameras, and traffic speeds, utilizing machine learning and data fusion ot forecast road flooding conditions.
Real-World Testing
The importance of these data sources became clear during Hurricane Harvey in 2017, when many in Houston, including emergency responders, had to manually analyze data to estimate road conditions due to the absence of reliable real-time information.
Model Performance
In assessing the OpenSafe Fusion system, researchers employed historical flood data from Harvey, reconstructing the scenario with data from roughly 62,000 Houston roads,
"The model was able to track approximately 37,000 road links, equating to 60% of the network we examined--a significant progress," Panakkal commented.
Future Directions and Considerations
Potential Data Sources
Potential data sources for the framework include water-level sensors, citizen portals, crowdsourcing platforms, social media, flood modeling, and the "human-in-the-loop" approach.
Importance of Human Element
The human element, Panakkal points out, is especially important in OpenSafe Fusion, as it allows for the responsible deployment of AI.
According to Panakkal, the goal is to avoid a system that operates entirely without human intervention.
Safeguards and Responsible AI
"The model could potentially make incorrect predictions, posing risks to community members who rely on these forecasts to make travel decision. Thus, we implemented safegurads centered on responsible AI usage. The need for responsible AI in such applications reamins a crucial area for ongoing research, and we plan to explore it further as we continue testing our methods."
Impact of Community Access
The research also evaluated how flooding affects community access to essential services, including hospitals and dialysis centers, during natural disasters.
According to Panakkal, this allows community members and emergency responders to identify flooded roads and determine the safest paths to travel.
Conclusion and Future Plans
Exploration and Validation
The researchers, Padgett explains, plan to focus on rigorous testing, validation, and exploration to discover how communities with diverse scales and resources could apply this framework to enhance flood road condition sensing.
Addressing Climate Change
According to Padgett, the escalating frequency and intensity of flood events driven by climate change necessitate the development of more effective solutions to handle their repercussions on infrastructure.
Labels: AI System, AI technology, flood detection