Symposium honors autonomous and connected systems graduate student research
“Autonomous and complex systems control has been highlighted as one of the key strategic research areas that the College would like to grow and strengthen,” says Dana Weinstein, associate dean for graduate education.
“ICON aims to enhance Purdue’s research and education on autonomy by joint efforts from researchers with different expertise,” says Shaoshuai Mou, co-director of the Center for ICON. “This symposium is to show ICON’s most recent research on autonomy from our outstanding students.”
Shreyas Sundaram, co-director of ICON added, “ICON faculty and graduate students are doing world-class research on enabling the next generation of autonomous and connected systems, and the symposium highlighted the breadth and depth of expertise that is available here at Purdue.”
The 2021 ICON Award Committee consisted of Martin Corless (chair, AAE), Hubo Cai (CE), Jianghai Hu (ECE), Mo Rastgaar (Poly) and Song Zhang (ME). The committee evaluated the research of all nominees and selected three for the 2021 ICON Outstanding Student Research Award based on their presentation, research and advisors’ comments.
“It was difficult to pick awardees,” says Corless. “The outstanding body of research presented by graduate students ranged from the theoretical to the very applied.”
The awardees are Wanxin Jin (AAE), Brian R. Page (ME), and Mythra V. Balakuntala (Poly):
Wanxin Jin
Advisor: Shaoshuai Mou, assistant professor of aeronautics and astronautics
“Control-induced Learning Applications to Human-Robot Autonomy”
“Wanxin’s research has mainly focused on robotics, especially application of inverse optimal control (IOC) into human-robot teaming,” says Shaoshuai Mou, assistant professor of aeronautics and astronautics and Jin’s advisor. Mou is also co-director of ICON. “One of Wanxin’s research contributions is development of the first distributed algorithm for solving the IOC problem. His research has been recognized by top robotic journals, including The International Journal of Robotics Research and IEEE Transactions on Robotics (T-RO), TRO); Automatica, a leading archival journal in the field of systems and control; and top machine learning conferences such as Neural Information Processing Systems (NeurIPS).
Brian R. Page
Advisor: Nina Mahmoudian, associate professor of mechanical engineering
“Persistent Autonomous Maritime Operation with an Underwater Docking Station”
“Brian’s research contributions are important and have lasting impact,” says Nina Mahmoudian, associate professor of mechanical engineering and Page’s advisor. “His most significant contribution is the design, optimization, and experimental validation of a novel autonomous docking system that can support a wide range of underwater vehicles. The contribution has been recognized by the U.S. Navy through patent protection. Long-term, this work extends the operational life of autonomous underwater vehicles from hours to days, even to months due to the ability to transfer data and recharge vehicles on-site in an autonomous and mobile manner.”
Mythra V. Balakuntala
Advisor: Richard M. Voyles, the Daniel C. Lewis Professor of the Purdue Polytechnic
“Coaching by Demonstration for Learning Contact-rich Tasks”
“Mythra has been doing some excellent work on Machine Learning for Coaching Robots by Demonstration in both a simulation environment for controlling medical robots under austere communications constraints, and co-robotic applications in small-batch manufacturing,” says his advisor Richard M. Voyles, the Daniel C. Lewis Professor of the Purdue Polytechnic and head of the Collaborative Robotics Lab. “He is extending the Learning from Demonstration paradigm to include sporadic coaching feedback to fine-tune the motion strategy. This includes the ability to estimate and parameterize the 'style' of a demonstrator/coach to recognize the difference between a surgeon and a medic, for example. As part of this work, he wrote his own proposal to Amazon Research, which was awarded a $100,000 grant."
Additional Student Presenters:
Ziping Liu (ME)
“Automatic Unloading-On-the-Go for Agricultural Combine Harvesters”
Wei Deng (Math)
“Dynamic Importance Sampling and Beyond”
Qingyi Gao (STAT)
“On the Latent Space of Generative Models”
Hemant Gehlot (CE)
“Optimal Sequencing Policies for Recovery of Systems Deteriorating After Disasters”
Sangjun Lee (Poly)
“Attack Resilient Vehicle Platooning”
Mridul Agarwal (IE)
“Multi-Objective Reinforcement Learning With Non-linear Scalarization”
Ciyuan Zhang (ECE)
“Estimation and Distributed Eradication of SIR Epidemics Over Networks”
Videos from the event: https://engineering.purdue.edu/ICON/events/icon-outstanding-student-research-symposium