Mouse sensor cluster for position and orientation detection

Four Laser-Mouse sensor+PIC33+PC=Position and Orientation sensing.
Screen: Top: output of each sensor(4x2)
Middle: Pos&Dir (X,Y,Th), Vel(Vx,Vy,omega) **FS of Th=180deg
Bottom: Calculated velocity distribution
25Hz sampling(max transmission via 115.2kbps=250Hz for 8ch)
Sensor resolution 2000DPI.
Where to check? Response and Repeatability of X,Y,Th (come back to almost zero)

Category:

Science & Technology

Tags:

• Robotics
• Sensor
• Sensing
• Mouse

[makumagai]

Hi,

One largest reason is that the system detect 'velocity' of the unit, which means we need to integrate the velocity into position. On the integration, a small error in direction cause large error in position after motion. The same problem is often discussed to estimate motion of wheeled robot.

Repeatability seems to be good and I didn't find lost count in particular direction. The surface is of course important (as computer mice), but for all direction.

Best regards.

[stormfox1]

very cool, what are you going to do with it?

[makumagai]

@stormfox1 This is a test implementation for other `secret':-) project. I'd like to upload a video of that project someday.

I can share how to make this, if someone want this kind of sensing.

[Nomoreidsleft]

@makumagai Yes, can you let me know on the theory and algorithm? I'm trying to implement position hold using mouse sensors on my quadcopter. Right now I have to combine sonar ranging and an optical mouse sensor to get proper velocity calculation, but you seem to be able to do it with just multiple sensors.

[makumagai]

@Nomoreidsleft Hi,

Can you access to IEEE explore?

ttp://ieeexplore.ieee.org/sear­ch/freesrchabstract.jsp?tp=&ar­number=5979899

Is the published document.

If not, let me know which e-mail address to send PDF.

But note that the mouse sensor can only measure surface speed (not left from the surface) because gap between the sensor and the target must be kept small (1or2mm) and constant. So I think it is not easy to apply to a quadcopter.

[cpu4edu]

おみごと！

[makumagai]

@cpu4edu ありがとうございます。

[wktodd]

@makumagai

Thanks :) That's very helpful :)

Regards, Bill

[wktodd]

Nice demo :)

I have been thinking about using a mouse sensor to measure linear displacement.

Have you been able to find out what causes the small return to zero (counting) errors? Is the accuracy and repeatability speed, or surface dependant? Do you get lost counts on reversal?

Regards, Bill