SMART Lab - Purdue

Excited to share that 8 papers from our lab (www.smart-laboratory.org) will be presented at (IEEE International Conference on Robotics and Automation)! This includes 6 accepted papers from the original ICRA submissions and 2 papers transferred from RA-L for oral presentations. Huge congratulations to all our lab members and collaborators on this unprecedented achievement!

Soyun Choi, a PhD visiting student, joins SMART Lab in Fall 2024. Soyun received her BS in Information and Communication Engineering from Andong National University and MS in Electrical and Computer Engineering from Inha University, respectively. She is currently a PhD student in Immersive Media Engineering at Sungkyunkwan University, under the supervision of Prof. Sungeun Hong (https://www.csehong.com/). Her research interests include Semantic/Instance Segmentation, Multimodal Learning, Prompt Learning, and Amodal Segmentation.

Please join us in welcoming Soyun to the SMART Lab team!! 🥳
Check out Larry further at http://www.smart-laboratory.org/group/Soyun_Choi.html

Excellent News! Shyam Sundar Kannan has successfully defended his PhD! He became the 7th student to earn a PhD degree from the lab, and the 1st student to complete both an MS and a PhD with us. During his PhD, Shyam published 4 journal papers and 12 conference papers (Check out Shyam’s Google Scholar: https://scholar.google.com/citations?user=HGZCqPoAAAAJ&hl=en; he has already over 200 citations!). Currently, 2 papers are under review. Along the way, he has received numerous awards and honors, such as the Purdue Engagement Award in 2022, 2nd place at Autonomous Day 2.0 in 2022, and the Purdue Research Foundation (PRF) Fellowship in 2020.

Since starting his PhD in 2019, Shyam has been actively involved in multiple projects. For example, in his first 2 years, he primarily worked on delivery robots in the context of human-robot interaction, a project funded through his PhD fellowship. This project was particularly challenging as it required a deep understanding of human-robot interaction theory, extensive user studies, and field experiments in real-world environments. Shyam significantly contributed to the success of this project with his excellence in formulating engineering problems, rapid development, and his design and ex*****on of research experiments.

Following that, Shyam participated in the 2021 Inaugural Indy Autonomous Challenge as an Autonomous Systems Engineer on the Black & Gold Autonomous Racing Team. Since then, he has been working on a project involving vision-based localization and mapping for autonomous mobility. In fact, the main concept of this project was originally developed by Shyam, and its success to date is largely due to his creative thinking and hard work. We have been pleased to witness Shyam’s passion, commitment, and quality contributions to his research.

Thank you and congratulations, Shyam!

(A recording of his defense presentation is available at: https://www.youtube.com/watch?v=npj5xpN4ykA)

Ziqin (Larry) Yuan, a Ph.D. student, joins SMART Lab in 2024 Fall. Larry received his BS in Robotics Engineering from Beijing University of Chemical Technology and MS in Mechatronics and Robotics from New York University, respectively. His research interests are in robotics, specifically in human-robot interaction, multi-human multi-robot teaming, and socially-aware navigation.
Please join us in welcoming Larry to the SMART Lab team!! 🥳

Check out Larry further at http://www.smart-laboratory.org/group/Ziqin_Yuan.html

Congratulations, Dr. Min, on his new role as the inaugural director of the Purdue Applied AI Research Center (AARC)! 🥳

Welcome Message: https://www.youtube.com/watch?v=FG3zzmlFnkA
AARC Website: https://polytechnic.purdue.edu/aarc

Welcome to Purdue Applied AI Research Center | Byung-Cheol Min, Director To learn more about the Applied AI Research Center (AARC) at Purdue Polytechnic, visit: https://polytechnic.purdue.edu/aarc

Come and check this interesting talk if you are around!

Don't miss the opportunity to attend this Applied AI Research Center seminar next Friday August 16th at 3pm on Purdue's West Lafayette campus (POTR 234). This talk will address Autonomous Driving Innovations and many other related topics. For more information visit: https://polytechnic.purdue.edu/aarc/news-and-events

Excited to share that 3 of our submitted papers have been accepted for presentation at the 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems ( )! These accepted papers highlight the ongoing research in my lab focusing on and . More specifically, papers 1 and 2 below address formation and task allocation challenges in multi-robot systems, introducing our new approaches using learning from demonstration and LLMs. Paper 3 introduces LLM-based approach for robot learning in room segmentation.

1. “SMARTLLM: Smart Multi-Agent Robot Task Planning using Large Language Models”, authored by Shyam Sundar Kannan, Vishnunandan Venkatesh, and Byung-Cheol Min
arXiv: https://arxiv.org/abs/2309.10062
video: https://www.youtube.com/watch?v=TnyCKwgTm3U
website: https://sites.google.com/view/smart-llm/

2. “Learning from Demonstration Framework for Multi-Robot Systems Using Interaction Keypoints and Soft Actor-Critic Methods”, authored by Vishnunandan Venkatesh and Byung-Cheol Min
arXiv: https://arxiv.org/abs/2404.02324
video: https://www.youtube.com/watch?v=IF6t4A7lZEw

3. “Semantic Layering in Room Segmentation via LLMs”, authored by Taehyeon Kim and Byung-Cheol Min
arXiv: https://arxiv.org/abs/2403.12920
video: https://www.youtube.com/watch?v=gBEG_KySboY
website: https://sites.google.com/view/selros

Proud of our team's hard work and looking forward to sharing our findings at IROS2024, Abu Dhabi, UAE!

Purdue University | Purdue Polytechnic | Purdue CIT

Exciting News! Our paper titled “MOCAS: A Multimodal Dataset for Objective Cognitive Workload Assessment on Simultaneous Tasks” has just been published in IEEE Transactions on Affective Computing ( )!

This paper introduces , a multimodal dataset dedicated to human cognitive workload (CWL) assessment. MOCAS represents a valuable addition to current research in multimodal fusion CWL assessment. To our knowledge, MOCAS is the first open-access dataset to include both physiological and behavioral data, along with CWL and emotion annotations, including personal traits and background information of subjects from realistic closed-circuit television (CCTV) monitoring tasks. This increases its applicability for real-world human-robot/machine interaction scenarios. All data were collected using off-the-shelf, user-friendly sensors; thus, models built from our dataset can be easily applied in everyday applications with efficiency.

Publication (IEEE Xplore): https://ieeexplore.ieee.org/document/10557131
Open-access preprint (arXiv): https://arxiv.org/pdf/2210.03065
Supplementary video: https://youtu.be/BxVVj7R9b70
Dataset: https://zenodo.org/records/7023242

Kudos to all authors, especially the first co-authors, Wonse Jo and Ruiqi Wang!

Our paper entitled “PlaceFormer: Transformer-based Visual Place Recognition using Multi-Scale Patch Selection and Fusion” has just been published in IEEE Robotics and Automation Letters ( )!

This paper introduces , a transformer-based model for visual place recognition that uses patch tokens to create global image descriptors. Unlike traditional methods, it focuses on task-relevant regions and refines image recognition through a self-attention mechanism, re-ranking images based on similarity scores from geometrically verified multi-scale patches. Extensive tests on benchmark datasets demonstrate PlaceFormer's superior accuracy and computational efficiency.

Publication (IEEE Xplore): https://ieeexplore.ieee.org/document/10543144
Open-access pre-print (arXiv): https://arxiv.org/abs/2401.13082

Kudos to Shyam Sundar Kannan! More to come from his stellar PhD work!

Ruiqi Wang has passed his PhD prelim exam. Congratulations, Ruiqi!

Vishnunandan Venkatesh, PhD student, has been selected as the Member of the Year for the SMART Lab (www.smart-laboratory.org) in recognition of his outstanding achievements and significant contributions throughout the 2023-24 academic year. The SMART Lab Member of the Year is chosen through a vote by all lab members, and the winner receives a certificate and cash award. Congratulations Vishnu on this well-deserved recognition!

Vishnunandan Venkatesh received the Employee Recognition Award and the Bravo Award from the University. Congratulations, Vishnu! We are so proud of you! 🥳