AUTinyDane

AUTinyDane is an open-source, affordable quadruped robot platform developed at Aarhus University.
The project focuses on a complete real robot stack: mechanical design, servo actuation, inverse kinematics, gait generation, and PS4 teleoperation on a Raspberry Pi.

What this repository contains

Full Python control stack for a 12-DOF quadruped
Real-time gait/state logic for trot and standing control modes
Servo + IMU hardware interface for Raspberry Pi
PS4 controller integration for teleoperation
Calibration and utility scripts for bring-up and maintenance

System overview

Software and system overview

Hardware architecture

Controller: Raspberry Pi (single-board control)
Actuation: 12 hobby servos driven by PCA9685 boards
IMU: MPU6050
Chassis: Laser-cut + 3D-printed hybrid structure

Software architecture

Core runtime flow:

Read controller command inputs
Update movement/stance command state
Generate leg trajectories (swing/stance planners)
Solve inverse kinematics
Send commands to servo drivers

Main components:

src/GaitPlannerV2.py – trot cycle logic
src/StancePlannerV2.py, src/SwingPlannerV2.py – stance/swing trajectories
src/Kinematics.py – inverse kinematics
src/HardwareInterface.py – servo/IMU communication
src/PS4Controller/ – DualShock 4 mapping

Repository structure

AUTinyDane/
├── run_robot.py                 # Main runtime entry point
├── requirements.txt             # Python dependencies
├── src/                         # Core runtime modules
│   ├── Configuration.py         # Robot geometry and limits
│   ├── Command.py               # Command state and mode abstraction
│   ├── State.py                 # Runtime gait/state values
│   ├── Kinematics.py            # IK solver
│   ├── HardwareInterface.py     # PCA9685 + IMU interface
│   ├── ServoCalibration.py      # Servo neutral offsets and gains
│   ├── GaitPlannerV2.py         # Main gait planner
│   ├── StancePlannerV2.py       # Stance phase planning
│   ├── SwingPlannerV2.py        # Swing phase planning
│   └── PS4Controller/           # Controller input mapping/listener
├── tools/                       # Calibration and utility scripts
│   ├── calibrate_servos.py      # Interactive servo offset calibration
│   ├── ServoNeutral.py          # Move servos/legs to neutral
│   ├── TestServo.py             # Interactive servo testing helpers
│   └── IMU_calibration.py       # IMU calibration utility
├── tests/                       # Manual/experimental test scripts
└── legacy/                      # Older planners/simulators

Prerequisites

Hardware/runtime environment

Raspberry Pi with Python 3
I2C enabled on the Pi
PCA9685 board(s) connected on expected addresses (0x40–0x43 in current code)
MPU6050 IMU connected (0x68)
PS4 controller (for teleoperation)

Software

Python 3.10+ recommended
pip

Note: Several dependencies are Raspberry Pi / hardware specific and may not install or run on non-Pi machines.

Installation

git clone https://github.com/JensLajordMunk/AUTinyDane.git
cd AUTinyDane
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

If running on Raspberry Pi, also ensure system packages and I2C support are configured.

Quick start (robot runtime)

Verify wiring and power-off safety state
Confirm/update servo calibration in src/ServoCalibration.py
Pair and connect the PS4 controller (/dev/input/js0 expected)
Run:

python run_robot.py

Controller mapping

From src/PS4Controller/DogController.py:

X → TROT
Circle → ROTATE
Square → TRANSLATE
Triangle → NEUTRAL
Left stick (L3) → translational commands (forward/lateral + stand height)
Right stick (R3) → yaw/roll/pitch depending on mode

Current modes are defined in src/Command.py:

NEUTRAL
TROT
ROTATE
TRANSLATE

Calibration and utility scripts

Servo neutral/pose checks:
```
python tools/ServoNeutral.py
```
Interactive servo calibration (writes src/ServoCalibration.py):
```
python tools/calibrate_servos.py
```
Servo test helper:
```
python tools/TestServo.py
```

Run these carefully with the robot safely supported/off-ground.

Development and testing

This repository currently contains mostly hardware/manual test scripts under tests/ rather than a full automated CI-style unit test suite.

Basic discovery check:

python -m unittest discover -q

Useful manual scripts include:

tests/Test_kinematics.py
tests/gait_simulatorV2.py
tests/IMU_test.py

Troubleshooting

Controller not detected: ensure DS4 is connected and appears as /dev/input/js0.
No servo movement: verify I2C wiring and PCA9685 addresses.
Unstable pose/motion: re-run servo calibration and verify geometry values in src/Configuration.py.
Import/runtime issues on desktop OS: many dependencies target Raspberry Pi hardware.

CAD model

View CAD Model in Onshape

License

This project is licensed under the MIT License. See LICENSE.

This site is open source. Improve this page.