You can directly try out PyBullet in your web browser, using Google Colab. !pip install pybullet takes about 15 seconds, since there is a precompiled Python wheel.
Here is an example training colab a PyBullet Gym environment using Stable Baselines PPO:
It is also possible to use hardware OpenGL3 rendering in a Colab, through EGL. Make sure to set the MESA environment variables. Here is an example colab.
Tagged a github release of Bullet Physics and PyBullet, both version 3.05. The release was used for our motion imitation research, and also includes various improvements for the finite-element-method (FEM) deformable simulation, by Xuchen Han and Chuyuan Fu .
https://github.com/bulletphysics/bullet3/releases
See also the related repository with model-predictive-control (MPC) quadruped locomotion and the motion imitation research at https://github.com/google-research/motion_imitation
Assistive Gym leverages PyBullet for physical human-robot interaction and assistive robotics. Assistive Gym currently supports four collaborative robots and six physically assistive tasks.
It also supports learning-based control algorithms, and includes models of human motion, human preferences, robot base pose optimization, and realistic pose-dependent human joint limits
Paper Link: https://ras.papercept.net/proceedings/ICRA20/1572.pdf Zackory Erickson, Vamsee Gangaram, Ariel Kapusta, C. Karen Liu, and Charles C. Kemp, “Assistive Gym: A Physics Simulation Framework for Assistive Robotics”, IEEE International Conference on Robotics and Automation (ICRA), 2020.
In “Learning Agile Robotic Locomotion Skills by Imitating Animals”, we present a framework that takes a reference motion clip recorded from an animal (a dog, in this case) and uses RL to train a control policy that enables a robot to imitate the motion in the real world. By providing the system with different reference motions, we are able to train a quadruped robot to perform a diverse set of agile behaviors, ranging from fast walking gaits to dynamic hops and turns. The policies are trained primarily in simulation, and then transferred to the real world using a latent space adaptation technique that can efficiently adapt a policy using only a few minutes of data from the real robot. All simulations are performed using PyBullet.
See also https://ai.googleblog.com/2020/04/exploring-nature-inspired-robot-agility.html and https://arxiv.org/abs/2004.00784
Robots can use meta-learning to quickly adapt from simulation to various real world conditions.
See paper at https://arxiv.org/abs/2003.01239
Become the squishy glowing creature you’ve always wanted to be. Take control of the worlds inhabitants to solve puzzles. Overthrow the Mastermote! Check it out at https://store.steampowered.com/app/791240/Lumote
Check out the Wired article about the Alphabet ‘Everyday Robot’.
PyBullet and Bullet Physics is used in the collaboration, as discussed in this “Speeding up robot learning by 100x with simulation” paper and described in those sim-to-real slides and the “Challenges of Self-Supervision via Interaction in Robotics” slides.
Facebook AI Habitat is a new open source simulation platform created by Facebook AI that’s designed to train embodied agents (such as virtual robots) in photo-realistic 3D environments. The latest version adds Bullet Physics.
The github repo is here: https://github.com/facebookresearch/habitat-sim
Also check the online demo (no physics): aihabitat.org/demo
This Robot Table Tennis project shows some very exciting research by Reza Mahjourian. It uses PyBullet and its Virtual Reality physics server support. Check out https://sites.google.com/corp/view/robottabletennis and his PhD thesis, Arxiv paper is here: https://arxiv.org/abs/1811.12927
“Hierarchical Policy Design for Sample-Efficient Learning of Robot Table Tennis Through Self-Play”
TossingBot, a new paper by Google Robotics (Andy Zeng, Shuran Song, Johnny Lee, Alberto Rodriguez, Thomas Funkhouser
) using PyBullet.
We investigate whether a robot arm can learn to pick and throw arbitrary objects into selected boxes quickly and accurately. Throwing has the potential to increase the physical reachability and picking speed of a robot arm. However, precisely throwing arbitrary objects in unstructured settings presents many challenges: from acquiring reliable pre-throw conditions (e.g. initial pose of object in manipulator) to handling varying object-centric properties (e.g. mass distribution, friction, shape) and dynamics (e.g. aerodynamics). In this work, we propose an end-to-end formulation that jointly learns to infer control parameters for grasping and throwing motion primitives from visual observations (images of arbitrary objects in a bin) through trial and error. Within this formulation, we investigate the synergies between grasping and throwing (i.e., learning grasps that enable more accurate throws) and between simulation and deep learning (i.e., using deep networks to predict residuals on top of control parameters predicted by a physics simulator). The resulting system, TossingBot, is able to grasp and throw arbitrary objects into boxes located outside its maximum reach range at 500+ mean picks per hour (600+ grasps per hour with 85% throwing accuracy); and generalizes to new objects and target locations.
See https://arxiv.org/abs/1903.11239 and a video here: https://www.youtube.com/watch?v=f5Zn2Up2RjQ&feature=youtu.be