Artificial Intelligence Depot
Visiting guest. Why not sign in?
News, knowledge and discussion for the AI enthusiast.
FEATURES COMMUNITY KNOWLEDGE SEARCH  
Natural Language Understanding
Delivers a synthesis of major modern techniques in natural language processing. The approach is unique in its coverage of semantic interpretation and discourse alongside the foundational material in syntax.
More information at Amazon US UK

Reply to Message

Not registered yet?

The AI Depot has a focused community of friendly users. Rather than let anyone abuse the site at the brink of promiscuity, we prefer to let only those with an active interest participate... this simply requires registering.

Why not sign up!

Joining the site's community is completely free. You can then post messages freely, and customise your personal profile at will. Specific privileges will also be granted to you, like being able to access printer-friendly articles without restrictions. So, why not register?

Username:
Password:
Subject:
Email me when someone replies.
Body:

Parent Message

Visually guided reaching

One useful task for a humanoid robot is to be able to reach towards objects within its immediate vacinity and try to grasp them. Physical manipulation helps the system to learn about the properties of objects, and to determine whether the object is passive (inanimate) or active (animate).

Rodney doesn't have any hands or pressure sensors yet, but I've been trying to get him to reach for objects which are close to the cameras. For this task the physical structure of the robot allows some crude shortcuts to be taken. When rodney observes moving objects his cameras pan and tilt to try to maintain the object within the centre of vision. When rodney's cameras are tilted downwards they can only view a limited area within the immediate vacinity of the robot, so its safe to assume that any movement observed is within close proximity. One of the arms is then moved to try to intercept the object, using the already calculated centre of motion value to guide it into the right area. This seems to be a crude but effective mechanism allowing the robot to reach for objects. A more sophisticated method would be to use the stereo matching to calculate distance and shape more accurately.

- Bob

http://www.fuzzgun.btinternet.co.uk/rodney18.jpg

136 posts.
Saturday 05 January, 19:01
Reply
Cunning!


> When rodney's cameras are tilted downwards they can only view a
> limited area within the immediate vacinity of the robot

A truly ingenious way of determining the distance ;)

Stereoscopy would be expensive though... It might be worth having a few sonar sensors, but I don't know how much they're worth.

935 posts.
Wednesday 09 January, 04:25
Reply
Wither ultrasonics?

I do actually have a set of six ultrasonic sensors from an earlier (now defunct) robot. It's quite possible that I could use these on Rodney, but I'd prefer to keep the system simple and efficient if possible.

I think approximate distance measurement should be possible with stereo vision, and this is the approach I'll be following. If the problem turns out to be absolutely insoluble only then would I resort to using the ultrasonics.

- Bob

136 posts.
Wednesday 09 January, 13:37
Reply

Back to the Artificial Intelligence Depot.