This paper addresses the problem of anchor-free multi-agent collaborative localization. We discuss three different coarse localization schemes: 1) Intuitive Coarse Localization (ICL), 2) Multidimensional Scaling (MDS) and 3) Semidefinite Programming (SDP). Then, a unified set of sequential and parallel LS techniques are applied to modify these coarse estimates. An outlier detection procedure is also introduced for localization with the presence of outliers. The numerical comparisons yield important insights to practitioners.