I have two time dependent signal sources X & Y. Both can be modeled as having a linear combination of time dependent individual components and common components; so X(t)=a(t)+C(t)+noise, Y(t)=b(t)+C(t)+noise. For both signals I have multiple 'runs' of data, but only ~10 time samples per run. Which Method/techniques should I use to recover a(t) and b(t)? I have tried SVD, but it seems that a(t) and b(t) are similar enough to where their contribution gets 'mixed' in the U vector. The best results I have obtained have been through a High Pass filter because C(t) tends to move faster, but even that hasn't been as robust as I would like. I appreciate your input, thank you.