Robot identification using fractional subspace method

This paper is concerned with fractional identification of state space model of continuous time MIMO systems. The methodology used in this paper involves a continuous-time fractional operator allowing to find fractional derivatives of the stochastic input - output data which are treated in time domain and identifying the state space matrices of the system using QR factorization. There are many advantages in describing a physical system using fractional CT models in that the dynamic behavior of the system is, in actuality, inherently fractional. The efficacy of the approach is examined by comparing with other approaches using integer identification.