Souma Chowdhury receives NSF CAREER Award

Project aims to design teams of robots that can work together in real-world environments

Map showing location of several drones.

Inside UB's SMART Motion Capture Lab, researchers create a simulated environment to demonstrate how autonomous air and ground robots can work together. Photo by Douglas Levere.

by Jane Stoyle Welch

Published May 6, 2021

An assistant professor in the University at Buffalo Department of Mechanical and Aerospace Engineering (MAE) is being recognized for his impactful research with a National Science Foundation (NSF) CAREER award.

“Among all the popular science buzz on future robots taking away jobs or going rogue, and disproportionally abundant dystopian themes on TV associated with it, it’s easy to forget about the unique ways in which autonomous robots could impact our lives when we are at our most vulnerable.”
Souma Chowdhury, assistant professor
Department of Mechanical and Aerospace Engineering

Souma Chowdhury

Souma Chowdhury recently received NSF’s most prestigious award to early career faculty. The award will support his research to design teams of robots, also referred to as swarms, that can operate collectively in a predictable way and adapt to a variety of environmental conditions.

The research has implications in a wide variety of fields, including disaster response, environment monitoring, military operations and space exploration.

“As someone who is academically trained in engineering design and swarm algorithms, I have strongly felt that there is an untapped opportunity to bring the machinery of design automation to deliver predictable performance for robots that operate as a team,” says Chowdhury.

Most existing approaches to designing the behavior of robot swarms are unable to predict how individual robots will respond when they are operating in a group and cannot adapt to the differing environments that they face in the real world.

“This project will support theoretical research to concurrently tackle these limitations by leveraging machine learning tools combined with rigorous engineering design approaches to systematically study how knowledge representation and the physical design of individual robots influence their collective behavior,” says Chowdhury.

Education and outreach will be achieved through experiential learning programs such as computer gaming that will teach students how to design robots, and use flying drones to monitor tree health at the University at Buffalo.

“We also plan to engage with local fire departments to explore the technical and socio-economic barriers to transitioning indoor hazard localization capabilities of swarm robots to practice,” says Chowdhury.

"Souma’s research will advance the design and implementation of dependable swarm systems using imitation learning of individual agent behavior," says Francine Battaglia, professor and chair of the Department of Mechanical and Aerospace Engineering. "It will undoubtedly build a firm foundation for his future work in addressing the fundamental questions related to the effectiveness of innovating dependable swarm systems."  

“His research aligns with the School of Engineering and Applied Sciences’ and University-wide initiatives focused on robotics, autonomy, cyber-physical systems and the environment, especially Sustainable Manufacturing and Advanced Robotic Technologies (SMART) and Research and Education in Energy, Environment, and Water (RENEW),” added Battaglia.

Chowdhury’s research is at the intersection of engineering optimization, evolutionary computing and physics-infused machine learning, with applications to drones, mobile robots, large multi-robot and multi-agent systems and renewable energy. He is affiliated with UB’s Artificial Intelligence Institute, and the SMART Community of Excellence.

“Among all the popular science buzz on future robots taking away jobs or going rogue, and disproportionally abundant dystopian themes on TV associated with it, it’s easy to forget about the unique ways in which autonomous robots could impact our lives when we are at our most vulnerable,” says Chowdhury.

The five-year award, entitled “Automated Design of Decentralized Robust and Explainable Swarm Systems (ADDRESS),” is funded with $500,000 from the NSF’s Division of Civil, Mechanical and Manufacturing Innovation (CMMI).

In addition to his CAREER award, Chowdhury is also the PI on a DARPA award that explores machine learning approaches and the concept of learning from humans playing computer games, to develop tactical strategies for operating large swarms of mobile robots in adversarial environments. He has authored two book chapters, 40 peer-reviewed journal articles, over 90 full-length conference publications and has one design patent.

He is a member of the American Society of Mechanical Engineers (ASME) and the Institute of Electrical and Electronics Engineers (IEEE), a senior member of the American Institute of Aeronautics and Astronautics (AIAA), and a selected member of the AIAA Multidisciplinary Design Optimization (MDO) Technical Committee. He serves as a reviewer for over 20 international journals in the areas of engineering design, machine learning, robotics and renewable energy.

Notable honors and awards include the Renewable Energy Top Paper Award from Elsevier in 2015, and UB’s School of Engineering and Applied Sciences’ Early Career Researcher of the Year award in 2019.

Chowdhury joined the University at Buffalo in 2016. Prior to this, he was an assistant research professor at Mississippi State University and a research assistant professor at Syracuse University. He earned his PhD in Mechanical Engineering from Rensselaer Polytechnic Institute in 2012.