The Parameters And States Estimation For Dual-rate System By The Swarm Intelligence Optimization Algorithm

In the past, the research of multirate system identification and estimation has been used basedon traditional recursive identification algorithms. With the development of control theory andidentification theory, many complicated system identification algorithms have received muchresearch attention in quick convergence, high accuracy and strong robustness in several decades.Therefore, the research of system identification based on new algorithm has significant value intheory and application. The paper is based on biological optimization theory, combiningcharacteristics of linear system, and presents new parameter and state estimation algorithm for statespace models from the mulirate. Through modern control theory, the convergence and parameterssetting rule of the algorithm is analyzed. The author reads and researches a lot of relevantreferences, and studies the identification of state systems. The innovation research results areobtainedin thethesisas follows:1. For the weaknesses of the standard particle swarm optimization algorithm, the paperpresents a modified cooperative particle swarm optimization algorithm (MCPSO) and deduces theiterative formula of the algorithm. According to stability theory for standard second order system,the convergence and parameters setting rule of the algorithm is analyzed and a good optimizationperformance is shown. Finally, a valuable simulation example is given. The optimization of testfunctionsisprovided toshowthevalidity androbustnessoftheproposedalgorithm.2. For the dual-rate system state space models is set based on lifted technique. Basedintelligent optimization identification algorithm a residual is derived. Minimizing the covariancematrix of the estimation error may get parameters estimation. For the dual-rate system white noisemodel and color noise model, the identification algorithm of MCPSO is compared with that ofrecursive least square. A valuable simulation is provided to show the convergence and robustnessofthe MCPSOalgorithm.3. Based on the canonical state space models for non-uniformly sampled-data system,MCPSO algorithm is used to estimate the parameters and states. An example is given to validatetheeffectivenessofthenewstateestimation algorithmfornon-uniformlysystem. A simple conclusion is given in the end. The further research and difficulties of the thesis areoutlined in the end, for example, the parameter identification and state estimation problem, the bestparameterssetting,thetheoreticalproofoftheproposed algorithmand soon.