Projects
armctl
A unified interface for controlling a variety of industrial and hobbyist robots from different manufacturers. Supports Python and Command Line Interface (CLI) for seamless integration into existing workflows.
MGross21/armctl(opens GitHub in a new tab)MuJoCo Toolbox
A modern simulation wrapper for Google DeepMind's MuJoCo. Provides an intuitive and efficient interface for robotics simulation and reinforcement learning research. Features include environment management and rendering utilities.
MGross21/mujoco-toolbox(opens GitHub in a new tab)SpikeBLE
Cross-platform Bluetooth Low Energy (BLE) interface for programming LEGO® SPIKE™ devices using MicroPython. Enables wireless code deployment and real-time interaction with SPIKE™ robots from desktop. Extended functionality to iOS app to help users with learning difficulties access programming features.
MGross21/spikeble(opens GitHub in a new tab)Anonymous Local LLM
Provides a fully self-hosted AI stack for private, local language model inference and search, with a strong focus on privacy. Routes all internet requests through an anonymous proxy to prevent telemetry and tracking. Supports popular open-source LLMs like Llama 2 and Vicuna, with easy deployment via Docker.
MGross21/anon-ai(opens GitHub in a new tab)Pomodoro CLI
A simple pomodoro timer that runs in the terminal. Written in Rust and is less than 100 lines of code. Features customizable work/break intervals, desktop notifications, and sound alerts to help users stay focused and productive.
MGross21/pomodoro-cli(opens GitHub in a new tab)Glovebox Digital Twin
This project integrates the mujoco-toolbox and armctl repositories. Developed as part of a Master's thesis at Arizona State University. Integrates a unified robotic manipulator controller and YOLO OpenCV pipeline with a modern simulation wrapper for MuJoCo, enabling seamless simulation and control of various robotic arms within a constrained environment.
MGross21/glovebox-digitaltwin(opens GitHub in a new tab)SyncStride Wearable
SyncStride is a wrist-mounted wearable designed for individuals with Parkinson's Disease. It detects asymmetric arm swing in real-time and provides vibrotactile cues to encourage symmetrical gait patterns—helping reduce fall risk and improve mobility.
MGross21/syncstride-wearable(opens GitHub in a new tab)ASU DREAM Competition
Designed and built an artistic representation of human-robot interaction as a musical wind-up box. The robot and human animate together holding the college banner, symbolizing collaboration. The robot is controlled via an Arduino microcontroller and features LED lighting effects. Awarded 1st place and $1,000 at ASU's DREAM Competition.
MGross21/artisan-automation(opens GitHub in a new tab)Image Edge Mapping Research
Researched the intricacies of image edge mapping matrix operators (Prewitt, Sobel, Roberts Cross, etc.) and the applications of linear algebra in Autonomous Vehicle Object Detection algorithms. Developed MATLAB code to evaluate efficiency and effectiveness of said operators. Presented findings at the 2024 Innovation Showcase at Arizona State University.
MGross21/EdgeMappingResearchProject(opens GitHub in a new tab)Robotic Jumping Frog
The Robotic Jumping Frog project successfully modeled, optimized, and fabricated a bio-inspired five-bar leg mechanism to replicate frog-like jumping. Using MuJoCo simulation, the team refined link lengths, stiffness, and geometry to maximize propulsion efficiency, then translated the optimized design into a functional foldable-robotics prototype. Experimental validation confirmed the system's ability to achieve coordinated jumping motion, demonstrating a strong correlation between simulated and physical performance. The project achieved its goal of bridging physics-based simulation and real-world fabrication, providing a validated platform for studying bio-inspired locomotion.
MGross21/Robotic-Jumping-Frog(opens GitHub in a new tab)ArUco Tag AR Mesh Visualizer
The project uses computer vision to detect ArUco markers and determine their full six degrees of freedom, including position and orientation, in real time. A camera is calibrated using a chessboard pattern to ensure accurate pose estimation. Each detected tag can be linked to a 3D mesh defined in a metadata file, allowing automatic generation of STL models that visually represent the object. The system uses OpenCV to detect markers, compute their 6DoF poses, and display their axes, providing a foundation for real-time digital twin applications and adaptable robotic control systems.
MGross21/ArucoCV-6DoF-MeshVisualizer(opens GitHub in a new tab)