Local Minima

Whenever you attempt to minimize a set of complex equations, the issue of local minima arises. Because you start from an initial guess and seek to minimize the error until you can minimize no more, you do not know that you are at the lowest possible minima for the entire solution space.

For this reason, it is possible for the bundle adjustment algorithms to find a solution that, while it is a minimum, is not the minimum you were searching for. SpatialAnalyzer contains a full 3-D graphical mod- el of the measurement environment to help decipher some of the local minima. It is important, however, to start the instruments near to their optimal location (i.e. use a good guess) to minimize the chance of a local minima.