Hi there, i have the following equation Ax=0 where x is a vector with 9 unknown elements. Now i draw samples from a very noisy distribution and each sample gives me a row of A, i.e. if i draw less than 9 samples the problem is ill-posed and if i draw more it becomes over-determined. Assuming we start from an estimated x=x0, i'm now looking for a online solution which updates x with every new arriving sample. It has to be computationally efficient, e.g. simply adding a new Sample to A (which contains all previous samples) and then solving Ax=0 and repeating this for every new sample won't work... thanks.