Comments for studywolf

Comment on Dynamic movement primitives part 1: The basics by travisdewolf

travisdewolf — Mon, 18 Nov 2024 15:31:02 +0000

In reply to Dagothar.

It definitely can be simply a linearly decaying value (and i believe later implementations of canonical systems changed to linear dynamics for simplicity). The reason to use canonical systems at all is just so you can control the temporal dynamics of the system easily. For example, slowing them down, speeding them up, tying the speed of the playback to the system error during execution (i.e., if the arm is having a hard time following the trajectory, slow the trajectory down)

Comment on Building models in Mujoco by travisdewolf

travisdewolf — Mon, 18 Nov 2024 15:28:32 +0000

In reply to ruoshiwen.

Thank you very much! Glad you found it useful!

Comment on Full body obstacle collision avoidance by travisdewolf

travisdewolf — Mon, 18 Nov 2024 15:28:15 +0000

In reply to Dharmik.

This is using the operational space control algorithm! I describe it in detail in several other posts, including this one: https://studywolf.wordpress.com/2013/09/17/robot-control-4-operation-space-control/

Comment on Using VREP for simulation of force-controlled models by travisdewolf

travisdewolf — Mon, 18 Nov 2024 15:24:33 +0000

In reply to Enes Emre Öztürk.

Oof, I apologize, it’s been a long time since i’ve used CoppeliaSim and I don’t recall offhand how to set this (but i suspect it’s google-able). I switched over to using Mujoco a while ago and haven’t looked back!

Comment on Deep learning for control using augmented Hessian-free optimization by travisdewolf

travisdewolf — Mon, 18 Nov 2024 15:23:22 +0000

In reply to Bryson.

ty!

Comment on Dynamic movement primitives part 1: The basics by travisdewolf

travisdewolf — Mon, 18 Nov 2024 15:23:09 +0000

In reply to Bai Jiang.

Hello! Sorry, I don’t quite follow. I talk about incorporating the error during execution of the path into the DMP equations in the “incorporating system feedback” section of the 2nd DMPs post (https://studywolf.wordpress.com/2013/12/05/dynamic-movement-primitives-part-2-controlling-a-system-and-comparison-with-direct-trajectory-control/)
Is this what you’re wondering about? If not could you rephrase for me?

Comment on Dynamic movement primitives part 1: The basics by Bai Jiang

Bai Jiang — Fri, 15 Nov 2024 03:17:09 +0000

Hi,I have a question about the"Imitating a desired path".whether I can selected the W to control the error between the acutal and the imitating to achieve arbitrarily modifying the shape of the curve we wanted.

Comment on Deep learning for control using augmented Hessian-free optimization by Bryson

Bryson — Sun, 27 Oct 2024 08:24:42 +0000

Great reading your posst

Comment on Using VREP for simulation of force-controlled models by Enes Emre Öztürk

Enes Emre Öztürk — Wed, 27 Mar 2024 11:04:59 +0000

In reply to travisdewolf.

how did you set the step configuration to 'custom'?

Also are you still using CoppeliaSim. İ am at the very beginning stage and i need help but i can't find :((

Comment on Full body obstacle collision avoidance by Dharmik

Dharmik — Thu, 14 Dec 2023 09:01:02 +0000

Which motion/path algorithm was used in the GIF of the 2D robot at the bottom? Im new to robotics and haven’t seen many algorithms and the one I’m using would have to account for obstacles from the start instead of during runtime.