Mindboards

aswin, did you already a median filter of 3 for closer objects?
to my experience, 4 or at least 5-255 cm are fine then.

Doc,

You know I love filters. So, Yaeh, I use median filters with my US sensors. But mainly to filter out 255 values when they are invalid. I don't bother to use filters to get good short distance data as I build rovers, "if you can read this sign, you're too close". I can imagine that for other purposes you might need short distance data and filtering would help.

For short ranges I advise to usethe minimum filter as it will filter these samples {255, 4, 255} to become 4. The median filter would return 255.

Aswin

For short ranges I advise to use the minimum filter as it will filter these samples {255, 4, 255} to become 4. The median filter would return 255.

good idea!

In the absence of sufficient sensor ports I now tried it with 6 USS at 2 HT sensor Muxers but I get really bad cross-over echoes in mazes.
I really need to drive them in a single shot mode - but no way to do this with the existing NXT stuff.


I think I have to wait for the EV3 and a real good USB network protocol for sensor and motor remote-controlling...

doc-helmut wrote: I really need to drive them in a single shot mode - but no way to do this with the existing NXT stuff.

Doc,

do you mean you cannot use single shot mode? Is this because of the sensor mux? The US sensor itself supports single shot mode.

Aswin

USS at HTSMux in single shot mode ??
how??

edit:
oh - I think I misunderstood you -
yes, at 1 single NXT port of course 1 USS is working in SSM, but I only have 2 ports for overall 6 USS
(or, quite exactly: 4 ports for 16 sensors + 1 motor muxer for 3 extra motors)