Spot Micro Quadruped Dog Robot is a small, four-legged robotic dog designed to mimic real canine movements. It serves various purposes, including research, education, and entertainment. Named "Spot Micro" due to its smaller size compared to Boston Dynamics' Spot robot, it's customizable for fun projects or serious applications like exploring rough terrain and search and rescue operations. Spot Micro showcases advancements in robotics with wide-ranging potential uses.
As students in the Spot Micro community, we developed a reliable and versatile PyBullet simulator for safe experimentation. It also serves as a Gym environment for Reinforcement Learning (RL) tasks.
Even without using the Gym environment, the simulator is useful. The main.py script accepts any gait implementation and provides a GUI for testing. For example, we've implemented a 12-point Bezier gait.
This simulator is a valuable tool for students, offering a safe and flexible platform for testing and development.
- X-axis Motion
- Y-axis Motion
- Z-axis Motion
- Roll Motion
- Pitch Motion
- Yaw Motion
More coming soon.....
First you will need to create an environment.
python -m venv <env-name>
To run this project, you will need to add the following to your environment
- Scipy
- Numpy
- PyBullet
- Gym
- Matplotlib
- Seaborn
- Setuptools
To install use command :
pip install <package-name>
After installing the required dependencies.
-
Start the environment by python :
.\env\Scripts\activate -
Navigate to the src folder :
cd .\simulation\src\ -
Run the main.py file :
python .\main.py
If you want to build the real world working model, here is the list of required parts:
Inspiration from : spot_mini_mini
Leg kinematics : Inverse kinematics
Lie Algebra : Code Snippets
Beizer curves : Beizer curves
Rotation : Rotation logic
Gym and env with the help of : Coumans, Erwin and Bai, Yunfei