 | | RE: A robot who knows his position (fwd)
|
|
(...) Just for reference: The optical mouse pads made by Mouse Systems, and shipped with Sun workstations for a long time, used two detectors in the mouse. One used a red LED, and the other one looked the same, but was dark. I suspect it was an IR (...) (28 years ago, 28-Apr-99, to lugnet.robotics)
|
| |
 | | Re: Og Needs More Inputs (fwd)
|
|
(...) Actually, I retract what I said about the light sensor and the touch sensor, but not on the grounds you stated, or at least on clearer grounds than you seemed to explain. (That you can configure the touch sensor or bias the photodetector seems (...) (28 years ago, 28-Apr-99, to lugnet.robotics)
|
| |
 | | Re: Inverting the robot position problem...
|
|
(...) When one color tape overlaps the other color it may be possible for the sensor to miss the bottom color when it passes over the intersection. Maybe you could use a square piece of tape of a third color at the intersection? Like you said, this (...) (28 years ago, 28-Apr-99, to lugnet.robotics)
|
| |
 | | Re: Inverting the robot position problem...
|
|
(...) That was what I had intended to use -- 1/8" tape spaced at 1" centers on a white background. I was thinking black and red should be sufficiently distinguishable from each other, but that will probably take experimentation. John A. Tamplin (...) (28 years ago, 28-Apr-99, to lugnet.robotics)
|
| |
 | | Re: Og Needs More Inputs (fwd)
|
|
(...) To add to this, the effect of placing a touch sensor in parallel with a rotation sensor is as follows: When touch sensor is not pressed, rotation sensor works fine When touch sensor is pressed repeatedly, rotation sensor drifts When touch (...) (28 years ago, 28-Apr-99, to lugnet.robotics)
|
| |
 | | Re: Inverting the robot position problem...
|
|
(...) You are right about the limitations of image based maze followers, but with you don't even need to get that complicated- any bot that can follow a wall using touch sensors can do a maze reasonably well. Picture a blind man walking along a (...) (28 years ago, 28-Apr-99, to lugnet.robotics)
|
| |
 | | Inverting the robot position problem...
|
|
(...) This brings out a VERY interesting application. Most bots have trouble navigating a real maze because of obvious slippage problems and generally poor resolution of sensors. What about inverting the problem and printing (using your favorite (...) (28 years ago, 28-Apr-99, to lugnet.robotics)
|
| |
 | | Re: A robot who knows his position (fwd)
|
|
(...) I have been thinking about a different approach. I have not actually built anything to use it yet, but I believe it should work reasonably well. The basic idea is similar to the old optical mouse pads -- draw horizontal and vertical lines on (...) (28 years ago, 28-Apr-99, to lugnet.robotics)
|
| |
 | | Re: A robot who knows his position (fwd)
|
|
(...) Looking forward to seeing them :) (...) Dead-reckoning should be used BETWEEN two consecutive (artificial) landmark sensing. When you get any external reference point you can zero your accumulated errors. (...) moving (...) Seems interesting. (...) (28 years ago, 28-Apr-99, to lugnet.robotics)
|
| |
 | | Re: Og Needs More Inputs (fwd)
|
|
(...) connected (...) Let's assume it is the LEGO rotation sensor. Which is not an 8-bit sensor. It is an incremental encoder that sends values back as 4 analog values. If a touch sensor is put in parallel with a rotation sensor the analog values (...) (28 years ago, 28-Apr-99, to lugnet.robotics)
|