Portfolio

Geometric 3D reconstruction augmented with Deep Learning features

Reflecting on the moment I captured this photo, I remember being struck by the beauty of the landscape and the joyful presence of my dog Simba. But now, when I look back at the picture, the vividness of that day escapes me. The depth and scale of the scenery, the true size of my fluffy companion—these details are lost in the flatness of the 2D image. It’s a common dilemma with traditional photography. This realization spurred my interest in developing methods to transform such 2D images into 3D models, allowing us to relive our memories with a richness that matches the original experience. My work in geometric 3D reconstruction is a step towards this future, where capturing life in all its dimensions could be as simple as taking a picture with our phones..

Catching a Moving Target: My Venture into Dynamic Path Planning

Delving into the realm of dynamic environments in my Planning and Decision-making in Robotics course, I faced the intricate challenge of intercepting a moving target—a task that pushed the boundaries of strategic robotic planning. My solution involved A* search and the implementation of backward Dijkstra based heuristic calculation in C++, focusing on optimizing performance times and memory efficiency..

Comparing Four Sampling-Based Path Planners for an N DOF Robot Arm.

In this project, I confronted the dynamic challenge of sampling based path planning for a robotic arm. This required precise path planning under constraints, for which I developed a sophisticated C++ planner.

Varied Bite Acquistion in Robot Assisted Feeding

A major chunk of my research at the Robotic Caregiving and Human Interaction Lab, undertaken at Carnegie Mellon University, focuses on enhancing the capabilities of robot-assisted feeding systems. Aimed at improving the quality of life for individuals who struggle with self-feeding, this research develops a control primitive enabling a robotic arm to adeptly scoop varying amounts of food.

Developing a Symbolic Planner for Automated Action Sequencing

My crack at implementing a generic symbolic planner involved constructing a planner capable of parsing an environment’s symbolic description and generating an actionable sequence to transition from an initial state to a goal. The environment, exemplified by the Blocks world from class, featured a set of objects and a table with specific initial and goal conditions that my planner had to navigate.

Automatic Speech Recognition with Synthetic Speech

Could Synthetic Speech revolutionize Automatic Speech Recognition?

Detection of Cracks Using Image Processing

Multi-Goal Grasp Planning for Mobile Manipulators in Cluttered Environments

This project presents a grasp planning system for 2D planning of a robot arm and the Stretch mobile manipulator, designed for handling multiple objects on a table. It incorporates Multi Goal RRT* with a legible costmap, motivated by the gap in existing methods where the most probable grasp might require a more costly overall path. The system aims to enhance the robot’s capability in choosing the best grasping paths in cluttered environments and improve safety and intuitiveness for collaborative use.

Comparative Study of SAE 8620 and EN 36C in an Automotive Differential

This research provides a comprehensive comparison of two case hardening alloy steels, SAE 8620 and EN 36C, in the context of their application in automotive differential gearboxes. The study aims to evaluate and contrast these materials based on various performance metrics crucial for automotive applications.