This is a small but powerful tool for multivariable subspace identification. It implements the so called Multivariable Output Error State Space algorithm.

The identification is carried out in two steps by using the nested function technique. The main function returns a score vector for the subspace obtained from the input and output data. From the score vector, users can easily determine an appropriate order for the model to be identified. Then, by calling the function handle, also returned by the main function, with the determined order, the state space matrices are obtained.

An example is included to show how to use this tool.