How do you calculate the average of a set of circu

Question

I want to calculate the average of a set of circular data. For example, I might have several samples from the reading of a compass. The problem of course is how to deal with the wraparound. The same algorithm might be useful for a clockface.

The actual question is more complicated - what do statistics mean on a sphere or in an algebraic space which "wraps around", e.g. the additive group mod n. The answer may not be unique, e.g. the average of 359 degrees and 1 degree could be 0 degrees or 180, but statistically 0 looks better.

This is a real programming problem for me and I'm trying to make it not look like just a Math problem.

Answer 1

I solved the problem with the help of the answer from @David_Hanak. As he states:

The angle that points "between" the two others while staying in the same semicircle, e.g. for 355 and 5, this would be 0, not 180. To do this, you need to check if the difference between the two angles is larger than 180 or not. If so, increment the smaller angle by 360 before using the above formula.

So what I did was calculate the average of all the angles. And then all the angles that are less than this, increase them by 360. Then recalculate the average by adding them all and dividing them by their length.

        float angleY = 0f;
        int count = eulerAngles.Count;

        for (byte i = 0; i < count; i++)
            angleY += eulerAngles[i].y;

        float averageAngle = angleY / count;

        angleY = 0f;
        for (byte i = 0; i < count; i++)
        {
            float angle = eulerAngles[i].y;
            if (angle < averageAngle)
                angle += 360f;
            angleY += angle;
        }

        angleY = angleY / count;

Works perfectly.

Answer 2

Compute unit vectors from the angles and take the angle of their average.

Answer 3

I have a different method than @Starblue that gives "correct" answers to some of the angles given above. For example:

• angle_avg([350,10])=0
• angle_avg([-90,90,40])=13.333
• angle_avg([350,2])=356

It uses a sum over the differences between consecutive angles. The code (in Matlab):

function [avg] = angle_avg(angles)
last = angles(1);
sum = angles(1);
for i=2:length(angles)
    diff = mod(angles(i)-angles(i-1)+ 180,360)-180
    last = last + diff;
    sum = sum + last;
end
avg = mod(sum/length(angles), 360);
end

Answer 4

You can use this function in Matlab:

function retVal=DegreeAngleMean(x) 

len=length(x);

sum1=0; 
sum2=0; 

count1=0;
count2=0; 

for i=1:len 
   if x(i)<180 
       sum1=sum1+x(i); 
       count1=count1+1; 
   else 
       sum2=sum2+x(i); 
       count2=count2+1; 
   end 
end 

if (count1>0) 
     k1=sum1/count1; 
end 

if (count2>0) 
     k2=sum2/count2; 
end 

if count1>0 && count2>0 
   if(k2-k1 >= 180) 
       retVal = ((sum1+sum2)-count2*360)/len; 
   else 
       retVal = (sum1+sum2)/len; 
   end 
elseif count1>0 
    retVal = k1; 
else 
    retVal = k2; 
end 

Answer 5

Python function:

from math import sin,cos,atan2,pi
import numpy as np
def meanangle(angles,weights=0,setting='degrees'):
    '''computes the mean angle'''
    if weights==0:
         weights=np.ones(len(angles))
    sumsin=0
    sumcos=0
    if setting=='degrees':
        angles=np.array(angles)*pi/180
    for i in range(len(angles)):
        sumsin+=weights[i]/sum(weights)*sin(angles[i])
        sumcos+=weights[i]/sum(weights)*cos(angles[i])
    average=atan2(sumsin,sumcos)
    if setting=='degrees':
        average=average*180/pi
    return average

Answer 6

Let's represent these angles with points on the circumference of the circle.

Can we assume that all these points fall on the same half of the circle? (Otherwise, there is no obvious way to define the "average angle". Think of two points on the diameter, e.g. 0 deg and 180 deg --- is the average 90 deg or 270 deg? What happens when we have 3 or more evenly spread out points?)

With this assumption, we pick an arbitrary point on that semicircle as the "origin", and measure the given set of angles with respect to this origin (call this the "relative angle"). Note that the relative angle has an absolute value strictly less than 180 deg. Finally, take the mean of these relative angles to get the desired average angle (relative to our origin of course).