top of page

Pulse I: Sentry

- Configured a complex, industry-standard development workflow involving the NVIDIA Jetson Orin Nano, an external SSD, and a dedicated Docker container comprised with necessary CUDA, TensorRT, PyTorch, Torchvision, and object detection libraries and configurations

- Engineered and tuned a 3D vision system by integrating the Intel RealSense d435i to utilized the depth data to intelligently filter false positives and precisely control the arm's Z-axis (reach) distance from the target

- Employed a rigorous, methodical debugging approach, including systematic variable isolation, commanded motor tests, and a "level-based" version control system to ensure stability. Successfully resolved critical issues across hardware communication, software library conflicts, and complex control logic

- Integrated a stable, 3-motor system that achieved intelligent 3D face tracking on the X,Y, and Z axes in the Arduino Braccio Arm

project documentation

 

*Pulse I came to a premature termination due to a wornout servo horn in the Braccio base

edge AI on Braccio
Daniel Suh

edge AI on Braccio

final setup
wornout servo horn
extending system's virtual memory
a look at the d435i's thermal detection
first setup of NVIDIA jetson orin nano
nanoOWL library incorporation successful
bottom of page