I’m using TREK1000 kit for my thesis work and I want to evaluate different 3d localization alorithms.
Does anyone know which is the algorithm used in the RTLS software?
Do you know other [color=#333333][size=large]3D[/size][/color] localization algorithms?

If you can suggest me some articles about this topic Ii would be very grateful.

The TREK uses a trilateration function based on 3 ranges as a solver This is described in section 2.3.1 in the DECARANGERTLS PC SOURCE CODE. The description is attached but it is also included in teh TREK package downloadable from Decawave.com
Reading the source code you can also conclude that TREK1000 supports max 4 anchors, but it will always base the location on three ranges.

In another example, 3D could use spherical intersection algorithm as the solver. That is an least mean square fit to the point on intersection of the hyperboloids defined by the measured TDOAs. A detailed description of the algorithm is given in the attached paper. It is not that same as the trilateration used in TREK but it is not a Kalman solver either,

In general for Kalman filters one should do some research and implement their own variant algorithm.
A good few customers for example use a form of extended Kalman filtering , while others use a graph based method, namely gtsam: http://borg.cc.gatech.edu/?destination=node%252F299

Maybe others in the forum can help with other algorithms for 3D