Experience
Prasanth specializes in tackling challenging AI problems in fulfillment with Robotics. These include
fundamental computer vision (CV) and robot manipulation work for optimal behavior learning. His
research at the THINC Lab primarily focuses on human-robot collaboration, with an emphasis on developing
human-centered AI solutions. His expertise spans a broad spectrum, encompassing areas such as object recognition,
reinforcement learning (RL), and general machine learning (ML) for enhanced decision making in Robotics.
Graduate Research Assistant
THINC Lab, University of Georgia
Supervisor: Dr.Prashant Doshi
- Developed a robust, flexible and automated software system to sort produce on a conveyor line using a collaborative robot (PrecisionSort is on track to be deployed commercially in produce sorting sheds through InversAI)
- Developed ML-based perception pipeline for real-time feedback to the robot’s control stack.
- Developed a robotic system (PrecisionDetect) - with an ML-based classification model to distinguish and segregate fertile eggs from infertile ones on a conveyor line
January 2019 - Present
Robotics Research Intern
Mitsubishi Electric Research Laboratories
Supervisor: Dr.Siddarth Jain and Dr.Diego Romeres
- Developed a state-of-the-art mathematical model and inverse RL algorithm to solve flexible collaboration between humans and robots in collaborative tasks
- Submitted two research papers – one to IROS 2024 and another to IJCAI 2024
- Patent pending
August 2023 - December 2023
Robotics Intern
RIF Robotics
Supervisor: Dr.Kevin DeMarco
- Collaboratively developed a seamless perception-based Sterile Processing pipeline using robotic manipulators
- Generated synthetic datasets to train an object detection model that accurately recognizes similar-looking instruments and bio-burden
May 2022 - August 2022
Robotics Intern
Lunar Outpost
Supervisor: Justin James
- Worked on ROS Navstack for Autonomous Navigation on an ARM controller using 3D LiDAR Mapping for a Mobile Robotics Platform
- Developed Teleoperation capabilities, Multi-sensor integration, RGB cam based 360 deg live streaming, helped develop on board Electronics, and Mechanical Subsystems
June 2018 - August 2018
Graduate Researcher
MirrorLab, Colorado School of Mines
Supervisor: Dr.Tom Williams
- Integrated Robot Architectures DIARC and ROS to realize an Alpine Search and Rescue(S&R) Scenario in ROS Gazebo
- Pipeline development between DIARC and ROS to pass NLP commands to ROS and the feedback to DIARC
- Worked with researchers from University of Bremen, Germany to implement this pipeline on a ROS Gazebo based UAV alpine S&R Platform
Aug 2017 - May 2018