gym-panda

OpenaAI Gym Franka Emika Panda robot grasping environment implemented with PyBullet

Links

• GitHub: https://github.com/mahyaret/gym-panda
• PyPI: https://pypi.org/project/gym-panda/
• Documentation: https://www.etedal.net/2020/04/pybullet-panda_2.html
• Issue Tracker:https://github.com/mahyaret/gym-panda/issues
• Download: https://pypi.org/project/gym-panda/#files

Install

Install with pip:

pip install gym-panda

Or, install from source:

git clone https://github.com/mahyaret/gym-panda.git
cd gym-panda
pip install .

Basic Usage

Running an environment:

import gym
import gym_panda
env = gym.make('panda-v0')
env.reset()
for _ in range(100):
    env.render()
    obs, reward, done, info = env.step(
        env.action_space.sample()) # take a random action
env.close()

Running a PD control HACK!

import gym
import gym_panda

env = gym.make('panda-v0')
observation = env.reset()
done = False
error = 0.01
fingers = 1
object_position = [0.7, 0, 0.1]

k_p = 10
k_d = 1
dt = 1./240. # the default timestep in pybullet is 240 Hz  
t = 0

for i_episode in range(20):
   observation = env.reset()
   fingers = 1
   for t in range(100):
       env.render()
       print(observation)
       dx = object_position[0]-observation[0]
       dy = object_position[1]-observation[1]
       target_z = object_position[2] 
       if abs(dx) < error and abs(dy) < error and abs(dz) < error:
           fingers = 0
       if (observation[3]+observation[4])<error+0.02 and fingers==0:
           target_z = 0.5
       dz = target_z-observation[2]
       pd_x = k_p*dx + k_d*dx/dt
       pd_y = k_p*dy + k_d*dy/dt
       pd_z = k_p*dz + k_d*dz/dt
       action = [pd_x,pd_y,pd_z,fingers]
       observation, reward, done, info = env.step(action)
       object_position = info['object_position']
       if done:
           print("Episode finished after {} timesteps".format(t+1))
           break
env.close()

Development

• clone the repo:

git clone https://github.com/mahyaret/gym-panda.git
cd gym-panda

• Create/activate the virtual environment:

pipenv shell --python=python3.6

• Install development dependencies:

pipenv install --dev