converge values to range [-pi,pi] in matlab, NOT using wraptopi

0
votes

The assignment is: Reset the values stored in the matrix accordingly, so that they lie within this range [-pi,pi]. Take care not to use any non-standard Matlab function(s) to do this.

(of course are all angles theta + 2n*pi the same, for any integer n. Therefore for example: an angle of 1.5*pi should be reset to -????0.5*pi.)

The nonstandard matlabfunction wraptopi does this (I think), but I am not allowed to use this function. I've got the feeling that I could use modulo to do this, but I don't know how.

Could anyone help me please?

Thanks in advance

1
matlabpi
Modulo will work. When you want to wrap onto a range that doesn't start at zero [0... N), you can add an offset, do modulo, then subtract the offset out again. – Ben Voigt
Thanks a lot! I'm almost there. However, when I compare my result with what wraptopi does, my values are on avarage a bit to big... This is what I did: %Start off with matrix D offset=max(max(D(:,:))); % determine the offset: I take the max value of the matrix D_with_offset=D+offset; % add the offset D_mod=mod(D_with_offset,(2*pi)); % take the modulo of all values, with 2pi D_back_within_range=D_mod-pi; % remove the offset again Could someone tell me where this goes wrong? – rasa
Your offset should just be pi, not the max of the matrix. Same value you subtract out at the end. – Ben Voigt
Thanks a lot! It works! :) – rasa

1 Answers

4
votes

To expand on @Ben Voight, you could use modulo-style operations this way:

To wrap to [0, 2*pi], you'd do this:

angle_rad = angle_rad - 2*pi*floor(angle_rad/(2*pi));

To wrap to [-pi, +pi], you'd add another term

angle_rad = angle_rad - 2*pi*floor( (angle_rad+pi)/(2*pi) );