TNY-Robotics/TNY-360

TNY - 360 Robot source code and 3d models

What it solves

TNY-360 is an open-source quadruped robot dog designed for accessibility in robotics education and experimentation. It provides a complete hardware and software stack to build a compact, responsive robot capable of interacting with its environment through sensors and vision.

How it works

The robot is powered by an ESP32-S3 MCU using a dual-core architecture: Core 0 handles a high-speed (200Hz) kinematics and control loop for movement, while Core 1 manages high-level orchestration, UI, and WebSockets. It utilizes modified MG996R servos with a custom Op-Amp buffer PCB to achieve closed-loop position feedback, replacing standard blind PWM control.

Who it’s for

It is designed for makers, STEM students, and robotics enthusiasts who want to build their own quadruped robot using 3D printing and off-the-shelf components.

Highlights

  • Closed-Loop Control: High-speed position feedback via modified servos for more precise movement.
  • Dual-Core Processing: Segregated "Reflex" and "Brain" cores for stable control loops.
  • Modular Hardware: Dedicated PCBs for power, sensors, and brain, connected via standardized cables to eliminate cable clutter.
  • Coded for All: Supports both professional C++/ESP-IDF firmware and a block-based web app (TNY-Coder) for beginners.
  • Integrated Sensors: Equipped with an IMU, Time-of-Flight Lidar, OV2640 camera, and audio capabilities.

Related

  • Project
    Phyzicalorg/Phyzical_org

    A browser-based robot teleoperation platform that crowdsources human demonstration data for training Vision-Language-Action (VLA) models and imitation learning policies.

  • Project
    peridyno/peridyno

    PeriDyno is a CUDA-based parallel physics engine that provides real-time simulation of physical environments for intelligent agents.

  • Project
    OpenMOSS/EasyWAM

    EasyWAM is an open‑source, modular codebase for training and evaluating World Action Models (robotic video‑action predictors). It supports multiple backbones, LoRA fine‑tuning, FlashAttention, DeepSpeed, and provides ready‑to‑run recipes for benchmarks such as LIBERO, RoboDojo, and RoboCasa365.