This week’s UTC-IASE faculty spotlight is on Dr. Ashwin Dani, who is an assistant professor in the Department of Electrical and Computer Engineering. He received his B.S in Mechanical Engineering from the University of Pune in India, and then went on to receive a Ph.D in Mechanical and Aerospace Engineering from the University of Florida. He began his career in academia as a post-doctoral research associate at the University of Illinois at Urbana-Champaign in 2011, and then moved to the University of Connecticut in 2013 and currently works as an assistant professor.

 

Dr. Dani is the PI for the Robotics and Controls Lab, which focuses on solving various estimation and control challenges in engineering domains such as robotics, automation, industrial and biomedical applications. His research falls into areas such as (1) Model building from data using machine learning, (2) Human-Robot collaboration and safety issues in manufacturing environments, (2) GPS-denied navigation of unmanned aerial systems and improved autonomy, (4) sensor data fusion, and (5) estimation and control for neuro-prosthesis. He has also done previous work in areas such as estimation and control theory, robotics, autonomous navigation, localization and mapping, and vision-based control.

 

This summer, Dr. Dani attending the American Control Conference in Milwaukee, where a paper titled “Learning Stable Nonlinear Dynamical Systems with External Inputs using Gaussian Mixture Models” was presented by the UTC-IASE Fellow Iman Salehi. Iman Salehi, a UTC-Fellowship graduate student, worked under the counsel of Dr. Dani this summer. Salehi’s project is titled “Physics Informed Machine Learning Project”, which is researching how to develop models for heat exchangers using machine learning that embeds dynamical system properties. The efforts in this project bring a paradigmatic change to how model building using machine learning is looked at to include physical properties of the system. This paper presented a data-driven modeling method with convergence guarantees embedded in machine learning. This paper is closely related to the “Physics-Informed Machine Learning Project”.

 

Professor Dani worked with UTC this summer on a project titled “Physics Informed Machine Learning” and “Wire Harness Assembly using Robots”, with a goal to develop methodologies and architectures for data-driven model learning, containing dynamical system properties coming from physics of the system embedded in machine learning. In terms of research publications, Dr. Dani and his lab worked on extending the conference manuscripts presented over the summer, to journal manuscripts, for projects  “Physics Informed Machine Learning”.