Iterative learning identification using quantized observations

This paper develops a novel iterative learning parameter identification algorithm for a class of single parameter systems with multi-threshold quantized observations. The identification algorithm is constructed along the iteration axis and it can incorporate the parameter identification ability and the learning ability to deal with unknown time-varying parameters. Based on the recursive form of the estimation error along the iteration axis, it is proved that the convergence of parameter estimation can be guaranteed over the whole finite time interval. A numerical example is given to demonstrate the effectiveness of the algorithms.