old_projects

Hypergradient Descent Implementation

Repository: GitHub Repository

Co-authors: Mohbat Tharani, Huzaifa Arif, Tobias Park, Momin Abbas, Xinyan Sun

Reimplemented the "Online learning rate adaptation with hypergradient descent" paper, developing advanced optimization algorithms for machine learning. This project involved deep understanding of optimization theory, gradient-based methods, and adaptive learning rate techniques for improved model training efficiency.

Autonomous Pipeline Following Robot

Developed a UAV that autonomously tracked pipelines using camera and laser scanner technology. The UAV was equipped with gas sensors that enabled it to detect leakages, stop at the detection site, and publish precise coordinates of the leakage location. This project combined computer vision, autonomous navigation, and environmental sensing for industrial applications.

3D Mapping using Lenovo Phab 2 Pro

Implemented real-time 3D mapping using ROS (Robot Operating System) to establish communication between different machines - a laptop and Lenovo Phab-2 Pro. Utilized the Phab 2 Pro's advanced sensors to build comprehensive 3D maps in real-time, with visualization capabilities through rviz on the laptop. This project demonstrated expertise in mobile sensing, ROS architecture, and real-time data processing.

Noise Cancelling Headphones

Led a team of four members to develop active noise cancellation technology. Implemented signal processing algorithms to generate inverted signals based on detected ambient noise, which were then interfered with the original noise to produce filtered audio output. This project showcased understanding of signal processing, acoustics, and real-time audio processing.

LC Meter

Designed and built a precision measurement instrument using PIC microcontrollers capable of accurately determining the values of connected inductors and capacitors. This project demonstrated proficiency in embedded systems design, microcontroller programming, and electronic measurement principles.