Sumanth Tangirala
About
Publications
Skills
Experience
Projects
Contact
Resume
Sumanth Tangirala
Sumanth Tangirala
Robotics ResearcherRutgers University
PRACSYS - RutgersRutgersTekionISRODA-IICT
Sumanth Tangirala
Robotics Researcher at PRACSYS Lab, Rutgers University
Passionate about tackling complex, interdisciplinary challenges that have a tangible and positive impact on the real world.Primary Areas of Focus: Robotics, Motion Planning, and Reinforcement Learning.Currently pursuing a thesis under the guidance of Professor Kostas Bekris, exploring robot kinodynamic motion planning algorithms for high-velocity robots with first-order goals.
Ph.D. in Computer Science | Rutgers University (2024 - Now)Master of Science in Computer Science | Rutgers University (2022-2024)Bachelor of Technology in Information and Communication Technology | DA-IICT (2016-2020)
Publications
Preprint:
KRAFT: Sampling-Based Kinodynamic Replanning and Feedback Control over Approximate, Identified Models of Vehicular Systems
Aravind Sivaramakrishnan, Sumanth Tangirala, Dhruv Metha Ramesh, Edgar Granados, Kostas E. Bekris
Conference Papers:
The State of Robot Motion Generation
Kostas E. Bekris, Joe Doerr, Patrick Meng, Sumanth Tangirala
Proceedings of the International Symposium of Robotics Research (ISRR), 2024
Roadmaps with Gaps over Controllers: Achieving Efficiency in Planning under Dynamics
Aravind Sivaramakrishnan, Sumanth Tangirala, Noah R. Carver, Kostas E. Bekris
Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2024
MORALS: Analysis of High-Dimensional Robot Controllers via Topological Tools in a Latent Space
Ewerton R. Vieira*, Aravind Sivaramakrishnan*, Sumanth Tangirala, Edgar Granados, Konstantin Mischaikow, Kostas E. Bekris
Proceedings of the IEEE International Conference on Robotics and Automation (ICRA), 2024
Finalist for the IEEE ICRA Best Paper Award in Automation, 2024
Stacked Autoencoder Based Feature Extraction and Superpixel Generation for Multifrequency PolSAR Image Classification
Tushar Gadhiya, Sumanth Tangirala, Anil K. Roy
Proceedings of the 2019 International Conference on Pattern Recognition and Machine Intelligence (PReMI), 2019
Skills
Machine & Deep LearningI have over five years of extensive experience in ML and DL, applying these in Computer Vision, Robotics, and NLP projects with PyTorch and TensorFlow
Reinforcement LearningI've spent the past year focusing on Reinforcement Learning for robot motion planning, primarily using PPO and SAC-HER at the PRACSYS lab
Robot Operating SystemIn my current research, I use ROS for robot control and navigation, including developing a perception system for PRACSYS Lab's MuSHR platform
Physics Simulation EnginesMy current work extensively involves using MuJoCo and NVIDIA IsaacGym for environments to simulate robots and develop robot controllers
Python, C++With over seven years of expertise in Python and C++, I have applied these programming languages in diverse projects spanning various fields
CUDA, Parallel ProgrammingI have used CUDA C++ to parallelize image segmentation from scratch and have also used it on a high-level to accelerate the training of neural networks.
Experience
  • PRACSYS Lab, Rutgers UniversityI am researching robot motion planning algorithms for high-velocity robots in dynamic conditions using AI/ML and Reinforcement Learning techniques
  • Rutgers UniversityAs a graduate of Masters of Computer Science, Class of 2024, I continue to pursue my passion for AI/ML and Robotics
  • Tekion Corp.Starting with a Winter 2020 software engineering internship and advancing to a full-time role as Software Engineer, I developed significant frontend components for a leading Dealership Management System.
  • Center of WIMS2, University of MichiganIn a collaborative project between WIMS2 and Smart City Lab, DAIICT, I developed end-to-end IoT solutions for an environmental logging sensor network
  • Indian Space Research OrganizationDuring a Summer Internship, I developed a time-series Machine Learning technique for accurate land cover classification using satellite imagery and this tool continues to be used by ISRO in their ongoing projects
  • Smart City Lab, DA-IICTAs a Research Assistant, I worked to advance the lab's goal of developing intelligent systems that use multi-modal data. My work was focused on PolSAR image segmentation and classification using Deep Learning techniques
  • DA-IICTI graduated with a Bachelor of Technology in Information and Communication Technology. I was also a member of the DA-IICT Google Developer Student Club, the Student Council and the Student Placement Cell
Projects
Obstacle Navigation with a Robot Dog
GPU Parallelization of Fuzzy Image Segmentation
Mobile Robot Social Navigation
Just Walk Out Shopping
Land Cover Classification using GeoSpatial Imagery
Contact Me
Email:tangiralasumanth@gmail.com||sumanth.t@rutgers.edu