Bhaskar Vundurthy, Ph.D.
 |
Postdoctoral Research Associate
|
Education and Experience
Postdoctoral Research Associate, University of Notre Dame, IN, USA, 2021-present.
Software Engineer - Advanced Visualization, The MathWorks Inc., Bengaluru, India, 2019-2021.
Ph.D. in Robotics from Indian Institute of Technology (IIT) Madras, Chennai, India, 2019.
M.Tech. in Control and Instrumentation from Indian Institute of Technology (IIT) Madras, Chennai, India, 2013.
B.E. (Hons.) in Electronics and Instrumentation Engineering from The Birla Institute of Technology and Science (BITS) Pilani, Pilani, India, 2011.
M.Sc. (Hons.) in Chemistry from BITS Pilani, Pilani, India, 2011.
Research Interests
My doctoral dissertation primarily uses tools from Computational Geometry to compute an optimal location for rendezvous of multiple robots in an environment cluttered with obstacles, while constraints are imposed on distances and times of travel. The research also employs tools from Game Theory to study the effect of an intelligent self-sufficient adversary in the rendezvous of multi-agent systems. An important part of the research involves the implementation of the designed algorithms on in-house fabricated bipedal and mobile robots. My time at The MathWorks Inc. led to the development of advanced examples involving robotic applications to (i) deploy deep learning algorithms on ARM devices like Raspberry Pi and (ii) create a scenario for a high fidelity vehicle model to generate and track waypoints in minimum time using an agent-based modeling approach.
My current postdoctoral research aims at computing optimal strategies for an intelligent player amidst n boundedly rational players in a repeated games framework. In particular, the work identifies how an intelligent player can change the expected outcome of the game when every other player employs Fictitious Play. Further, an ongoing collaborative endeavor focuses on designing incentive schemes for agents in a federated learning setup to encourage truthful elicitation of information.
In the past, I have had the opportunity to develop various hardware constrained algorithms and demonstrate their implementation on in-house fabricated bipedal and mobile robots. Going forward, I am looking to explore various learning paradigms in developing algorithms that simultaneously track and evolve, as opposed to the conventional algorithms that adhere to a strict (and more often a single) constraint. In addition, I have been collaborating with my prior advisor to design and fabricate soft robot hands with an aim to understand their kinematics and highlight their efficacy across multiple gripping strategies. In summary, my research interests span (but are not limited to) the following fields:
- Computational Geometry
- Game Theory
- Robotics
- Multi-agent systems
- Deep Learning
- Autonomous Driving
Selected Publications
B. Vundurthy and K. Sridharan, "Protecting an Autonomous Delivery Agent Against a Vision-Guided Adversary: Algorithms and Experimental Results," in IEEE Transactions on Industrial Informatics, vol. 16, no. 9, pp. 5667-5679, Sept. 2020 (Journal Impact Factor: 9.112).
B. Vundurthy and K. Sridharan, "Multiagent Gathering With Collision Avoidance and a Minimax Distance Criterion—Efficient Algorithms and Hardware Realization," in IEEE Transactions on Industrial Informatics, vol. 15, no. 2, pp. 699-709, Feb. 2019 (Journal Impact Factor: 9.112).
Sharad Kumar Singh, Puduru Viswanadha Reddy, and B. Vundurthy, "Study of multiple target defense differential games using receding horizon based switching strategies," in IEEE Transactions on Control Systems Technology, accepted for Publication in August 2021.
B. Vundurthy and K. Sridharan, "Time Optimal Rendezvous for Multi-Agent Systems Amidst Obstacles - Theory and Experiments," IECON 2018 - 44th Annual Conference of the IEEE Industrial Electronics Society, Washington, DC, 2018, pp. 2645-2650.