Project Posts

Potential Field Library

This project offers a potential field based motion planning library for robotic manipulators, written in C++ and compatible with ROS2. Utilizing attractive and repulsive force generation in SE(3), whole-body obstacle avoidance using FCL collision geometry, users only need to submit their robot’s URDF and inverse kinematics solvers to plan smooth, collision-free trajectories. SE(3) Task-Space Planning Whole-Body Velocity Planner Repository Structure Core C++ Library The PotentialField class manages the overall potential field representation, and offers path planning methods among others to interact/query the field. The core library has the following features:

Tuesday, December 2, 2025 | 6 minutes Read

Perseverance

Perseverance is an open‑source mobile robot inspired by NASA’s Mars rover. The project focuses on robust state estimation for a 6‑wheel rocker‑bogie platform using an Extended Kalman Filter (EKF) within a modular ROS 2 package for sensor fusion and, later, SLAM and navigation. I derive the non‑linear kinematic model and Jacobians, implements multiple drive‑base models (differential drive, mecanum, and rocker-bogie), and benchmark performance using IMU and wheel‑odometry data. The primary purpose of this project is to learn more about the full robotics stack for a mobile robot, including EKF SLAM, navigation, and control. I’ve selected the 6-wheel MARS Rover form factor because it’s my favorite robot and I want to try and build a novel motion model and autonomous navigation stack for the rover and open-source it for others to try.

Tuesday, August 12, 2025 | 2 minutes Read

Reinforcement Learning Exploration

The Jackal is a simple differential drive robot with an impressive payload capacity and can be equipped with various sensors. In this project, we simulated the jackal with a 3D LiDaR (Velodyne VLP16) in Mujoco and setup a PPO pipeline to teach the mobile robot to navigate towards a goal and avoid obstacles.

Thursday, June 12, 2025 | 1 minute Read

Teaching an Old Dog New Tricks

The Unitree Go2 Dog is an incredibly advanced robot with an impressive locomotion system. This project utilized the Genesis simulation platform to train policies for the robot using PPO (Proximal Policy Optimization) to accomplish various tasks such as strafing, running, jumping, and jumping over obstacles. This post is still under construction, check out our Final Presentation for more details Reward Function Design The reward function design was where the majority of the work was done to train the robot on different tasks. The Genesis platform offers an end-to-end example of a training script, environment, and evaluation script for teaching the robot to walk. The reward function for walking encouraged the dog to maintain a forward velocity while penalizing deviations from the body height target and from the initial joint points.

Wednesday, March 19, 2025 | 2 minutes Read

Open Source Delta Robot

An open source ROS package for controlling delta robots with position and velocity control. Configurable for adaptive forward and inverse kinematics. Trajectory generation and visualization examples are also offered. Designed for public use and easy integration with new delta robots as long as their configuration parameters are setup correctly. Robot Design and CAD are also available. Robot Kinematics The robot’s forward and inverse kinematics were first implemented in a jupyter notebook to visualize the robot’s configuration space and workspace.

Friday, March 14, 2025 | 12 minutes Read

Mobile Manipulation

This project incorporates several robotics concepts to perform a pick and place task in simulation using a mecanum-wheeled mobile robot with a 5 degree-of-freedom robot arm. Omnidirectional Mobile Base Kinematics (Mecanum Wheels) This robot uses mecanum wheels, which are omnidirectional wheels with 45-degree rollers that allow the robot to move in any direction without changing its orientation. When controlling the robot, we command the wheel velocities, which can be described by the following equation:

Thursday, December 12, 2024 | 7 minutes Read

ToastBot

Using a 7-DoF Franka Emika Panda Robot Arm, we developed a pick and place sequence using an Intel Realsense camera for identifying april tags. Robot Arm Control The Franka Emika Panda Robot Arm has a ROS package franka_ros which allows us to control the robot arm using the Moveit API. The Moveit API offers services and action servers to faciliate trajectory planning, end-effector operation, and motion profiling.

Thursday, December 12, 2024 | 2 minutes Read

Educational Robot Arm

An educational kit designed to teach the fundamentals of kinematics and dynamics. The kit is intended to accompany the course ME3460: Robot Dynamics & Control at Northeastern University. The kit is a physical representation of the final project, which is currently done purely through MATLAB simulation. Kit Design The robot is a 3-link planar manipulator with 3 degrees of freedom. The end effector is a simple gripper that can grab small objects.

Friday, April 12, 2024 | 3 minutes Read

Napkin.AI

A fast, lightweight graph retrieval-augmented generation tool for navigating codebases. Update: Winner of the Best Developer Tool sponsored by Warp Collaborated with: Rishi Iyer, Henry Bloom, Arnav Brahmasandra 💡 Inspiration As developers, we are always exploring new projects and open-source repositories. However, the complexity and scale of these codebases serve as a major barrier to entry even for the most experienced developers. Even after spending countless hours reading over files, documentations, and (hopefully solved) issues, it’s easy to find yourself lost and intimidated. We wanted to create a tool that would help developers navigate these codebases more effectively, answering questions and providing context on both specific and broad aspects of the codebase.

Sunday, February 4, 2024 | 8 minutes Read

Automated Poker Table (B-I-K-E)

The B-I-K-E (Best in Kard Entertainment) Poker Table is a capstone project for the Northeastern University Electrical and Computer Engineering program. Collaborated with: Connor Nelson Chris Swagler Tyler Passerine Adin Moses Jackson Heun Professor Canek Fuentes An automated system that creates a seamless home-game experience by automating the tedious and repetitive tasks that occur during a game of cards while lowering the barrier of entry of knowledge for complex games, focusing on Texas Hold ‘Em Poker as a proof of concept.

Tuesday, April 11, 2023 | 5 minutes Read

Chess Robot

This robot utilized stepper motors, a camera, sensors, and a gantry style system to interact with the chess board and play against a human opponent. System Hardware The robot’s frame was assembled using aluminum extrusion and 3D printed parts. 2 motors were used to drive the X-axis, and 1 motor was used to drive the Y-axis (horizontally). Timing belts were used to transfer the motion from the stepper motors to the gantry sliding on linear rails. Limit switches were placed at the end of each axis to home the robot and ensure every movement was accurate during gameplay.

Wednesday, June 8, 2022 | 3 minutes Read

Robotech

A graphical simulation of a swarm of robotic agents cleaning algal bloom in a lake utilizing path planning algorithms such as A* and RRT*. Submitted for the RoboTech 2022 Hackathon. The Problem Statement One of the most detrimental factors to the environment is the pollution and destruction to the oceans and all marine life. Aquatic habitats are incredibly beneficial for the sustainability of the planet since they act as massive carbon sinks and lower overall greenhouse gas emissions.

Thursday, April 21, 2022 | 4 minutes Read