pmean.m v1.1 (Sep 2009)

MEAN and STD of angles, polar, cylindrical, spherical or cartesian coordinates!

Description

This function calculates the mean and standard deviation of vectors in polar coordinates (and others), that is, from angles and magnitudes.

The usage is very simple:

For polar or cylindrical: Direction, Radius, Height
            [mD,eD] = pmean(D);
      [mD,mR,eD,eR] = pmean(D,R);
[mD,mR,mH,eD,eR,eH] = pmean(D,R,H);

For spherical coordinates: Azimuth, Elevation, Radius
            [mA,eA] = pmean(A);
      [mA,mE,eA,eE] = pmean(A,E,'s');
[mA,mE,mR,eA,eE,eR] = pmean(A,E,R,'s');

For cartesian coordinates:
            [mX,eX] = pmean(X,'c');
      [mX,mY,eX,eY] = pmean(X,Y,'c');
[mX,mY,mZ,eX,eY,eZ] = pmean(X,Y,Z,'c');

where "m" stands for MEAN's, and "e" for STD's.

Input angles should be in radians.

It ignores NaN's.

Enjoy it!
Any comments or bug reports will be very appreciated!

• MATLAB release: MATLAB 7.7 (R2008b)

[Michael Porter]

Thank you for this code. I'd like to know how you make the estimations of the std's?

[Carlos Adrian Vargas Aguilera]

Hi Michael.

They are the root of the sum of the squared deviations, defined as (dyj/dxi)dxi, for each xi cartesian coordinate. yj=yj(x1,x2,x3) are then the transformation from cartesian to other coordinate system.

Cheers!

[jichao zhao]

It is not accurate. Try this example
Say, D = zeros(2,1);
R = ones(2,1);
D(1) = 330; D(2) = 0;
 [mD,mR,eD,eR] = pmean(D*pi/180,R);
 polar(D*pi/180,R,'yo')
hold on
 polar(mD*pi/180,mR,'ro')