Student in Master of Science in Robotics

Developed simulations for a UGV with a manipulator that autonomously navigates the Waste Isolation Pilot Plant (WIPP)• Combined the 3D models of the UGV and manipulator and integrated the model in simulation environments in Gazebo.• Adapted existing manipulator scripts for use with the UGV, allowing users to simultaneously control the UGV and the manipulator.• Improved the accuracy and tracking of a rope identification script for the use of grasping ropes with the manipulator

A participant of the Opportunity Research Scholars program, a two-semester research program exclusively for students in the Electrical and Computer Engineering Department at Georgia Tech.Students are exposed to many interest areas of the electrical and computer engineering fields through researching in a team of three to four undergraduates. Each team is led by a Ph.D. mentor, who guides the students on their research.Participants of ORS also attend research workshops that inform students about proper research conduct, ethics, and procedures. By the end of the spring semester, each research team must create a presentation board and present it in front of judges. Currently, I am part of a robotics research group that focuses on the Teleoperation of Semi-Autonomous Robots through Uncertain Environments.

The goal of this research is to have the Roomba navigate the Klaus Advanced Computing Building (KACB) using Ultra-Wideband (UWB) localization techniques, specifically the Chan Ho and Harbi localization algorithms. Due to COVID-19, it was difficult to have many students work in the lab. I led the sub-team that developed a simulation using ROS and Gazebo, allowing the experiments to take place virtually. In the simulation, I imported 3D models of the Roomba and all three floors of the KACB. I then developed the controls for the Roomba in the simulation. I also implemented UWB devices and their functionalities, and the localization algorithms in the simulation. Afterward, I utilized a built-in navigation package in ROS that provides local and global planning tools. One of these packages is the Global Planner package. This package uses a cost map and the Dijkstra or A* algorithm to generate a path. By using an image of the KACB building the Global Planner generated a path, which the Roomba would begin to follow.

I integrated the Terrain Elevation-based Robot Path Planning (TERP) into ARL's ROS/Unity-based Autonomy Stack (Phoenix). I rewired and updated drivers of a physical UGV platform (Aion R6) for debugging TERP and calibrating a ZED 2 Camera and its point cloud generation. I also calibrated a DWA motion planner to respect the kino-dynamic constraints of an ATV-based UGV (Aion L4) and improved the performance of TERP by optimizing its path planning calculations.

I participated as a research intern at the University of Central Florida under Dr. Aman Behal. I worked on a program that allows users to deviate from an arm’s preset nominal trajectory using a haptics device. The haptics device’s feedback force would be used to encourage and discourage specific movement of the arm’s end effector. The arm’s end effector would have preset trajectory and the user can push or pull along the direction of the preset trajectory to control the speed at which the end effector follows the trajectory.

I participated in a research internship with Dr. Aman Behal and mentor Yannick Roberts at the University of Central Florida. I assisted with the continuation of Amirhossein Jabalameli’s exploration of edge-based recognition of objects for robotic grasping. Amirhossein Jabalameli designed an algorithm that calculated the optimal method of grabbing an object based on its edges. Yannick Roberts implemented the algorithm in a ROS-based software called GripIt. However, GripIt can only work with two-fingered grippers, so the research was to make the algorithm more robust for grippers with more fingers.