Multi-innovation Identification Algorithms For Dual-rate Sampled-data Systems

The Dual-rate sampled-data systems are abundant in the industry, because thesystem is slow output and fast input, we can not use the traditional Single-ratesystem identification method to identify the Dual-rate system. We need to findeffective algorithms to overcome the effects of Dual-rate system data integrity to theidentification. In this paper proposed several effective algorithms to solve theidentification of such a system. Research results have theoretical and paracticalvalue. The main contribution are as follows:1For the AR Dual-rate system, a Modified Least Square Algorithm is proposedbased on polynomial technology. The simulation illustrates the algorithm theadvantage of high identification precision and samll computation.2For the Dual-rate Output Error Model, a intermediate variable is introducedto estimate the unmeasured data by using the idea of the auxiliary model, so theDual-rate system is equivalent to the single-rate system. In order to overcome theerror of the equivalent model to the identification, the Multi-innovation GradidentIdentification Algorithm based on the auxiliary model is presented. Simulationresults show that the multi-innovation algorithm can greatly improve the recognitionperformance.3The Modified Iteration Algorithm is established based on the limitedlength.The algorithm takes the advantages of the iterative algorithm that make fulluse of measured data and makes the information vector array in each iteration nottoo big. Therefore, the algorithm has the advantages of small computational. Thesimulation illustrates the effectiveness.4Polynomial Technology in the dual-rate variance error model brought achange in the noise model, as well as the increase in the number of parameters needto identify. In order to overcome this problem, Dual-rate system Recursive LeastSquare algorithm based model transformation and Gradient Modification Iterationalgorithm are given based Model transformation.