【摘要】 基于方块脉冲函数逼近的线性连续回归模型的参数估计及其应用赵明旺（武汉钢铁学院）ＰＡＲＡＭＥＴＥＲＥＳＴＩＭＡＴＩＯＮＦＯＲＬＩＮＥＡＲＣＯＮＴＩＮＵＯＵＳＲＥＧＲＥＳＳＩＶＥＳＹＳＴＥＭＳＶＩＡＢＬＯＣＫＰＵＬＳＥＦＵＮＣＴＩＯＮＳＡＮＤＩＴＳＡＰＰ...更多还原

【Abstract】 Abstract Parameter estimation for linear continuous regressive systems is studied via Legendre polynomials approximation, and a least-squares algorithm is presented.Then, the application of this method to parameter estimation for dynamic systems in systems and control is discussed. A correlative analysis of the BPF approximating values of the continuous Wiener process is made and the biased property of the least-squares method for stochastic continuous dynamic systems is pointed out. Finally, the instrumental variable method for consistent estimation and three methods for choosing instrumental variable systems are proposed. computer simulations show that these methods are effective and practical.更多还原