Alert icon
We're changing our privacy policy. This stuff matters.  Learn more  Dismiss

Grasping simulaton with neural network in Matlab

Loading...

Sign in or sign up now!
19,082
Loading...
Alert icon
Sign in or sign up now!
Alert icon

Uploaded by on Sep 7, 2007

A neural network based agent using reinforcement learning determines the adequate posture of a simulated arm to grasp an obect while avoiding obstacles.

Category:

Film & Animation

Tags:

License:

Standard YouTube License

  • likes, 4 dislikes

Link to this comment:

Share to:

Uploader Comments (arbolte)

  • How many total DOF were active in the learning?

  • The arm has 7 DOFs and each finger has 4 DOFs. 3 fingers are used in the proposed example and therefore 19 DOFS are active.

  • Indeed, only three fingers are involved in the task

Top Comments

  • arm needs some valium

see all

All Comments (10)

Sign In or Sign Up now to post a comment!
  • How does the hand "see" the object?

  • @03ricr Thank you for your interest. The algorithm works as well for five fingers.

  • @MACDeCorum

    that is interesting i would say the simulation might involve the other fingers if a lifting motion was simulated where all fingers must be involved just to lift the weight. otherwise three fingers works well for grasping.

  • ok, im guessing the seizures was the "robot" learning to use its arm, but did anyone else notice that 1 finger that didnt get used?

  • seizure time xD

  • That's one of the funnies simulations I've seen around here.

Loading...

0 / 00Unsaved Playlist Return to active list
    1. Your queue is empty. Add videos to your queue using this button:
      or sign in to load a different list.
    Loading...Loading...Saving...
    • Clear all videos from this list
    • Learn more