Vaibhav V. Unhelkar

PhD Candidate
Interactive Robotics Group
AeroAstro · CSAIL
MIT

Email
lastname[at]mit.edu
Lab
32-314, Stata Center

GitHub · LinkedIn
ResearchGate · Google Scholar

I am a PhD candidate in the Interactive Robotics Group(IRG) at Massachusetts Institute of Technology advised by Prof. Julie A. Shah. My research focuses on developing decision-making and inference algorithms for robots that interact with humans. For developing these algorithms, I use tools from planning under uncertainty, Bayesian nonparametric inference, machine learning, and insights from human factors.

During my academic training, I have worked on several autonomous systems: satellites, aerial vehicles and mobile robots. As part of my M.S. research at MIT, I worked on developing a mobile robotic assistant for automotive final assembly lines. Prior to joining IRG, I was a researcher with the Dynamics and Controls group at the Department of Aerospace Engineering, IIT Bombay, where my work focused on System Identification of miniature aerial vehicles.

I obtained my undergraduate education at the Indian Institute of Technology (IIT) Bombay through the integrated B.Tech. + M.Tech. programme in Aerospace Engineering. My Master's and Bachelor's theses focused on spacecraft attitude estimation and GPS-INS integration, respectively. At IIT Bombay, I was a member of the Pratham student satellite team. As a summer intern, I have also worked for Turbomeca on the problem of volcanic ash in helicopter engines.

I am passionate about education and have been involved in educational outreach and mentoring activities. Along with research and education, I cherish interesting conversations, the works of Isaac Asimov, and instrumental music.

Updates

Jun '18
I will be attending the RSS Pioneers workshop in Pittsburgh, PA. More details coming soon!
May '18
I will be presenting joint work on integrating prediction and planning for human-robot collaboration at the International Conference on Robotics and Automation, Brisbane, Australia. For more details on the work, see our recent publication in IEEE Robotics and Automation Letters. [Details and Video]
Apr '18
I will be talking about on-going research on algorithms for learning latent states of human-robot interaction at the workshop on Robot Teammates at ICRA 2018. [Details]
Mar '18
Read more about the AAAI Fall Symposium on Human-Agent Groups, which I co-organized in November 2017 in the upcoming issue of AI Magazine. Shout out to the AAAI for their excellent support with organizing the symposium. [Details]
Feb '18
Our joint work with BMW on introducing mobile robots between people in the automotive final assembly will appear in IEEE Robotics and Automation Magazine later this year. [Details]
Jan '18
Our work on integrating prediction and planning for human-robot collaboration will appear in IEEE Robotics and Automation Letters later this year. [Details]
Nov '17
I am co-organizing the AAAI Fall Symposium on Human-Agent Groups at Washington, DC. [Details]
Jun '17
I will be presenting my on-going research on Decision-making for Communication in Human-Robot Teams in the workshop on human-robot interaction at the Robotics: Science and Systems conference. My talk will focus on challenges that robots face in exercising its communication capability while collaborating with humans. [Details]
Mar '17
I will be at the conference on Human-Robot Interaction in Vienna, Australia. At the conference, X. Jessie Yang will be presenting our research on modeling and evaluating Trust in Automation. [Details]
Jun '16
I will be presenting on-going research on Planning with Explicit Modeling of Time in workshops at the Robotics: Science and Systems conference and at the International Joint Conference on Artificial Intelligence. We are designing incremental planning algorithms that explicitly model time as part of the state space, thereby allowing a robot to effectively reason using evolving predictions in dynamic environments.
Feb '16
I will be presenting our paper on Decision-Making for Communication at the AAAI Conference on Artificial Intelligence, Phoenix, AZ. By explicitly modeling the belief of its teammate, our algorithm Contact allows a robot to efficiently utilize its communication capability. [Details]
Jun '15
I completed my M.S. at the Interactive Robotics Group, MIT. My master's thesis is titled "Towards a Mobile Robotic Assistant for Automotive Final Assembly Lines: Control, Sensing and Human Robot Interaction". Special thanks to my collaborators for making the project a success!
May '15
I will be presenting joint research on Human-Robot Co-Navigation at the International Conference on Robotics and Automation, Seattle, WA. Through a study of human walking motion, we confirmed the existence of anticipatory indicators of human walking motion. Further, we used this insight for prediction of human motion using time-series classification. [Details]
Jan '15
Amidst Boston's seemingly unending snow, completed my qualifier exams. Now a PhD Candidate!