Skip to main content

Designing an Open-Source Database and Benchmark Framework for Incident Response in Smart Cities

Primary Investigators:
Abhishek Dubey
Ayan Mukhopadhyay

Brief Description of Project:
Emergency response management (ERM) is a challenge faced by communities across the globe. First responders need to respond to a variety of incidents such as fires, traffic accidents, and medical emergencies. They must respond quickly to incidents to minimize the risk to human life. Consequently, considerable attention in the last several decades has been devoted to studying emergency incidents and response. Data-driven models help reduce both human and financial loss as well as improve design codes, traffic regulations, and safety measures. Such models are increasingly being adopted by government agencies. Nevertheless, emergency incidents still cause thousands of deaths and injuries and result in losses worth billions of dollars directly or indirectly each year (Hattis, 2015). This is in part due to the fact that emergency incidents (like accidents, for example) are on the rise with rapid urbanization and increasing traffic volume.

While data-driven modeling has shown to improve emergency response to traffic accidents, the availability of open-source data and standardized frameworks for evaluating algorithmic approaches to the response remains a challenge. This project will collate data pertaining to traffic incidents, traffic, weather, roadway geometry, among others, from open-source data repositories to create the first standardized open-source data source for learning forecasting and resource allocation models for improving emergency response. Then, the data will be used to learn machine-learning models by using state-of-the-art approaches such as random forests, neural networks, and others to create benchmarks for the research community.

Desired Qualifications:
Fluency in Python, data management, scikit-learn, keras, tensorflow.
Some background knowledge of machine learning.
Ideally, we expect the student to have taken undergraduate-level machine learning and big data courses 

Nature of Supervision:
The students will directly work with the team of students and investigators and will be given weekly tasks and will participate in discussions.
A Brief Research Plan (period is for 10 weeks):
The first two weeks will be reserved for literature review and getting familiar with the data and problem space.
The next five weeks will be reserved for technical implementations.
The next two weeks will be reserved for writing the paper and building the presentation.
The final week will be reserved for presenting the research work.
Number of Open Slots: 1
Contact Information:
Name: Abhishek Dubey
Department: Computer Science