Doubts about setJointMotorControl2 implementation

Official Python bindings with a focus on reinforcement learning and robotics.
Post Reply
locked21
Posts: 2
Joined: Mon Oct 26, 2020 9:34 pm

Doubts about setJointMotorControl2 implementation

Post by locked21 »

Hi, first of all thanks for making this wonderful platform available.

I've been working on some car navigation problems for reinforcement learning. Before using PyBullet, I was using a kinematic system of differential equations (simplified bycicle model) to simulate the car motion.

From what I understand from the documentation, PyBullet seems to control a motor by defining a reference input position or velocity to the motor.

Of course, in the real world, when we define a reference velocity for a PID controller, for example, the actual velocity does not follow immediately the reference value (like the angular speed vs. time plot in figure attached).

PyBullet takes that into account? The simulation looks completely fine, I just need this information for documentation purposes.

Thank you.
Attachments
step_reponse.png
step_reponse.png (41.73 KiB) Viewed 30216 times
User avatar
Erwin Coumans
Site Admin
Posts: 4221
Joined: Sun Jun 26, 2005 6:43 pm
Location: California, USA
Contact:

Re: Doubts about setJointMotorControl2 implementation

Post by Erwin Coumans »

PyBullet setJointMotorControl2 also has regular PD control. When using velocity control, it is up to the maximum force how fast the target is reached.
Post Reply