DistributedPositioning

TODO improve difficulty:

generalize to 3d
increase time between each communication
non-zero-mean noise

TODO improve accuracy estimate:

use number of knownNodes and their accuracy to estimate own accuracy
if knownNodes <= 2 then cost should be 0 and can't be used for estimating accuracy
stddev(posEst) could be a good estimate of accuracy if it is close to correct position
switch from scalar 'accuracy' to covariance matrix for each direction

TODO misc:

find why system does not converge in beginning
improve search algortihm in findNode