Open Source Robotics

고품질 오픈소스 로봇 프로젝트, 도구 및 소프트웨어의 엄선된 컬렉션.

DexHand v2 커뮤니티 빌드

The Robot Studio의 오픈소스 덱스터러스 핸드 설계

라이선스

CERN-OHL-S

  • Full ROS 2 Stack + LLM-Controlled Dexterous Hand (Trent Shumay / IoT Design Shop): Not just a build — Trent Shumay and IoT Design Shop became the primary software and firmware contributor to the entire DexHand ecosystem. What They Built Starting with the mechanical build (SLA resin printing for higher resolution than FDM), they went on to develop: - Custom firmware for Arduino Nano RP2040 Connect with BLE and USB Serial interfaces - Complete ROS 2 package suite — URDF model, RVIZ2 visualization, gesture controller, USB serial bridge - LLM control demo connecting GPT-4 to the physical hand via voice commands — one of the earliest demonstrations of LLM-controlled dexterous manipulation (September 2023) - VR glove integration using MANUS Quantum Metagloves for intuitive teleoperation - Webcam hand tracking via MediaPipe for controller-free operation Why It Matters This build demonstrates the full potential of open-source robotics collaboration. What started as one person building a hand from STL files evolved into a comprehensive software ecosystem that was merged into the official DexHand project. The builder earned co-maintainer status, and their ROS 2 meta package (52 stars) is now the most popular repository in the DexHand ecosystem. The LLM control demo — giving GPT-4 the ability to control individual finger extensions and trigger animations via natural language — showed a compelling vision of how large language models can interface with physical robotics hardware. Technical Details - 19 DOF dexterous hand for approximately $300 in components - SLA printer remix opened the design to the resin printer community (higher detail than FDM) - Multi-interface control: BLE, USB serial, ROS 2 topics, voice/GPT-4, VR gloves (MANUS), webcam hand tracking (MediaPipe) - All code and configurations fully open source across 18 repositories

이 웹사이트는 JavaScript가 필요합니다. 오픈소스 로봇 프로젝트를 탐색하려면 JavaScript를 활성화하세요.