Estimation-based recursive algorithm for the identification of a discontinuous output sampling system

System identification is usually based on continuous sampling presumption, but discontinuous sampling may happen in some real applications. Each sample value is to be used by some of the equations. Without the equations due to discontinuation, the total number of usable equations will reduce significantly, and the identification calculation will become less efficient. When the rate of lost sample increases to a criticality value, there is no equation left to do the traditional identification calculation. A new estimation-based recursive algorithm is suggested. In the algorithm the estimated output sample values are used instead of the lost ones, and appropriate equations are established, so that the efficiency will not decrease too much by the discontinuation of sample. By means of the appropriate equations, the calculation can be done even when the critical lost rate is reached. The analysis shows that the algorithm converges with probability in the vicinity of the real value. Simulation shows that it is efficient and practical