Resume
Researcher
ORCID:
https://orcid.org/0000-0002-5141-2406
Piyush
Gupta
Research Scientist at Honda Research Institute
San Jose, United States
Citations
-
Contact Info
Resume
About
I am a Research Scientist at Honda Research Institute, focused on exploring the use of Large Language Models (LLMs) in mobility decision-making. I hold a PhD (2023) and MS (2020) in Electrical Engineering from Michigan State University, and a B.Tech. (2015) in Mechanical Engineering from the Indian Institute of Technology Delhi. My expertise lies in behavior prediction, reinforcement learning, autonomous vehicle planning, and human-robot interaction. Proficient in deep neural networks, ROS, Python, and C++, I am passionate about developing innovative solutions to enhance safety and efficiency in mobility.
Professional SkillsPython
C++
Robotics
Large Language Models (LLMs)
Pytorch
ROS
CARLA
TensorFlow
Reinforcement Learning
Research ExperienceResearch Scientist
Honda Research Institute US via TriCom Quest
February 2024 - Present
San Jose, CA, USA
- Developed a versatile multi-agent mission planning system utilizing Large Language Models (LLMs) for optimized coordination and task management.
- Minimized hallucinations in LLMs by implementing function-calling techniques for solving graph-based mathematical problems.
Motion Planning
Control Theory
Machine Learning
Large Language Models
Autonomous Vehicles
Research Intern
Honda Research Institute US
May 2022 - February 2024
San Jose, CA, USA
- Designed an interaction-aware trajectory planner for autonomous vehicles, adeptly integrating neural network predictions with model predictive control, ensuring robust convergence guarantees.
- Enhanced the interaction-aware planner’s runtime efficiency fivefold by employing knowledge distillation to develop a compact one-shot prediction network using a Social-GAN-based teacher model.
- Devised and evaluated a robust Gaussian lane-keeping behavior prediction model in ROS/C++, integrating constant heading predictions with an interactive Intelligent Driver Model for autonomous driving software stack.
Autonomous Vehicles
Motion Planning
Controls
Machine Learning
Research Assistant
Michigan State University College of Engineering
August 2017 - October 2023
East Lansing, MI, USA
- Enhanced self-driving vehicle resilience to edge cases in CARLA through multi-agent adversarial reinforcement learning and dueling double deep Q networks, achieving a 44% improvement in complex driving scenarios.
- Researched evaluative feedback’s impact on human skill development in sequential decision-making, using maximum entropy inverse reinforcement learning to analyze human reward responses across diverse feedback strategies.
- Formulated an efficient Deterministic Sequencing of Exploration and Exploitation (DSEE) algorithm for model-based reinforcement learning, achieving sub-linear regret growth over time.
- Designed a sophisticated human-in-the-loop system using semi-Markov Decision Processes, optimizing human cognitive resources to enhance human-robot interaction.
- Adopted a game-theoretic strategy (CPR games) to incentivize collaboration in decentralized, diverse teams
Human-in-the-loop
Game Theory
Motion Planning
Machine Learning
Deep Learning
Control Theory
Work ExperienceResearch Scientist
Honda Research Institute US via TriCom Quest
February 2024 - Present
San Jose, CA
- Developed a versatile multi-agent mission planning system utilizing Large Language Models (LLMs) for optimized coordination and task management.
- Minimized hallucinations in LLMs by implementing function-calling techniques for solving graph-based mathematical problems.
Research Intern
Honda Research Institute US
May 2022 - February 2023
San Jose, CA, USA
- Designed an interaction-aware trajectory planner for autonomous vehicles, adeptly integrating neural network predictions with model predictive control, ensuring robust convergence guarantees.
- Enhanced the interaction-aware planner’s runtime efficiency fivefold by employing knowledge distillation to develop a compact one-shot prediction network using a Social-GAN-based teacher model.
- Devised and evaluated a robust Gaussian lane-keeping behavior prediction model in ROS/C++, integrating constant heading predictions with an interactive Intelligent Driver Model for autonomous driving software stack.
R&D Engineer
Honda Automobile R&D Center
July 2015 - August 2017
Haga, Tochigi, Japan
- Designed and validated advanced control systems, including high-pressure fuel pump, super knock, torque, and fuel injection, for the turbocharged V-6 engine in the 2020 Honda Acura TLX and MDX, utilizing C++ and Simulink.
- Developed an electric water pump control system enhancing turbo-cooling efficiency post engine shut-off.
Global Intern
Daikin Industries
May 2014 - July 2014
Shiga, Japan
• Modeled and simulated Daikin’s swing compressor in MSC ADAMS for noise and vibration reduction.
• Measured frequency response of the compressor using Scan and Paint and achieved 𝟐𝟔% 𝐫𝐞𝐝𝐮𝐜𝐭𝐢𝐨𝐧 𝐢𝐧 𝐧𝐨𝐢𝐬𝐞 with optimized accumulator mass and design.
Education
Michigan State University
East Lansing, MI Aug 2017 - Oct 2023
Doctor of Philosophy (PhD)
, Electrical Engineering
Thesis: Optimal & Game Theoretic Feedback Design for Efficient Human Performance in Human-Supervised Autonomy
Michigan State University
East Lansing, Michigan Aug 2017 - Dec 2020
Master of Science (MS)
, Electrical and Computer Engineering
Indian Institute of Technology Delhi (IITD)
New Delhi, India Jun 2011 - May 2015
Bachelors of Technology
, Mechanical Engineering
Undergraduate Thesis: Electrowetting based liquid lens with variable focal length
Patent(Granted): A membrane-less variable focus liquid lens and an imaging device
Judge of The Works of Others
Reviewer
Jan 2024 - Present
IEEE Robotics and Automation Letter (RAL)
Reviewer
Jan 2024 - Present
IEEE Intelligent Vehicles Symposium (IV)
Reviewer
Jan 2023 - Present
IEEE Transactions on Automatic Control (TACON)
Reviewer
Jan 2023 - Present
IEEE International Conference on Robotics and Automation (ICRA)
Reviewer
Jan 2023 - Present
Computational Intelligence and Neuroscience
Reviewer
Jan 2023 - Present
International Conference on Models and Technologies for Intelligent Transportation Systems (MT-ITS)
Reviewer
Jan 2022 - Present
IEEE Conference on Decision and Control (CDC)
Reviewer
Jan 2022 - Present
IEEE Transactions on Control of Network Systems (TCNS)
Reviewer
Jan 2022 - Present
IEEE Control and Systems Letter (LCSS)
Reviewer
Jan 2020 - Present
IEEE American Control Conference (ACC)
Presentations/Talks
Gaussian Lane Keeping: A Robust Prediction Baseline
Edmonton, Alberta, Canada Sep 2024
IEEE International Conference on Intelligent Transportation Systems (ITSC)
Towards Scalable & Efficient Interaction-Aware Planning in Autonomous Vehicles using Knowledge Distillation
Jeju, South Korea Jun 2024
IEEE Intelligent Vehicles Symposium (IV)
Interaction-Aware Trajectory Planning for Autonomous Vehicles with Analytic Integration of Neural Networks into Model Predictive Control
London, UK Jun 2023
IEEE International Conference on Robotics and Automation (ICRA)
Towards Physically Adversarial Intelligent Networks (PAINs) for Safer Self-Driving
San Diego, CA, USA Jun 2023
IEEE American Control Conference (ACC)
Deterministic sequencing of exploration and exploitation for reinforcement learning
Cancun, Mexico Dec 2022
IEEE Conference on Decision and Control (CDC)
On robust and adaptive fidelity selection for human-in-the-loop queues
Virtual Jun 2021
IEEE European Control Conference (ECC)
Optimal fidelity selection for human-in-the-loop queues using semi-Markov decision processes
Philadelphia, PA, USA Jun 2019
IEEE American Control Conference (ACC)
Achieving efficient collaboration in decentralized heterogeneous teams using common-pool resource games
Nice, France Dec 2019
IEEE Conference on Decision and Control (CDC)
Membership
Member
Mar 2024 - Permanent
IEEE Intelligent Transportation Systems Society
Member
Jan 2023 - Permanent
IEEE Robotics and Automation Society
Member
Jan 2023 - Permanent
IEEE Sensors Council
Member
Mar 2019 - Permanent
IEEE Control Systems Society
Member
Mar 2019 - Permanent
IEEE Systems Council
Graduate Student Member
Jan 2019 - Permanent
Insititute of Electrical and Electronics Engineers (IEEE)
Member
Jan 2019 - Permanent
IEEE Young Professionals
Graduate Student Member
Jan 2019 - Dec 2023
Society of Automotive Engineers (SAE)
Member
Mar 2019 - Dec 2021
IEEE Nanotechnology Council
Certifications
Japanese-Language Proficiency Test N2
Issued on Jul 2016
JLPT
Japanese-Language Proficiency Test N3
Issued on Dec 2015
JLPT
Patents
Enhancing Reasoning Capabilities in Vision-Language Models via System 2 Inductive Bias with GFlowNet
Pending | Sep 09 2024 | US | 63/692596 Interaction-aware trajectory planning
Issued | May 23 2024 | US | 18188323
Other Inventors:
Piyush Gupta ,David F Isele ,Sangjae Bae
Gaussian Lane Keeping: A Robust Prediction Baseline
Pending | May 18 2024 | US | 94135379
Other Inventors:
David F Isele ,Piyush Gupta ,Sangjae Bae
Autonomous Vehicle Trajectory Planning using Neural Network Trained Based on Knowledge Distillation
Pending | Dec 06 2023 | US | 18/531510
Other Inventors:
Piyush Gupta ,David F Isele ,Sangjae Bae
A Membrane-less Variable Focus Liquid Lens and an Imaging Device
Issued | Nov 22 2023 | IN | 471854
Other Inventors:
Supreet Singh Bahga ,Sasidhar Kondaraju ,Shivam Agarwal ,Roshan Patra ,Piyush Gupta ,Ankur
Prizes and Awards
Outstanding PHD Graduate Student Award
Michigan State University
|
Mar 2022
Teaching Experience
Robotics
417 | Spring | 2022 | Undergraduate
Realted Researchers
Message
