摘要
为了改善固定遗忘因子RLS(Recursive least-square)算法在时变系统中的跟踪性能,提出了一种改进的RLS算法。改进的RLS算法结合了可变遗忘因子的RLS算法和自扰动RLS算法,既克服了固定遗忘因子RLS算法中跟踪速度和参数失调的矛盾,而且也避免了当参数估值趋向于参数真值时,卡尔曼增益趋于零,从而RLS算法失去对时变系统的跟踪能力的问题。最后,在MATLAB平台下,对改进后的RLS算法进行了仿真验证。仿真结果表明,算法具有较快的收敛速度和跟踪速度以及较小的稳态误差。
In order to improve the tracking performance of the fixed forgetting factor RLS algorithm in the time - varying system, a modified RLS algorithm is proposed, which combines variable forgetting factor RLS algorithm with self- perturbing RLS algorithm. It overcomes the contravention between the tracking velocity and parameters＇ misadjustment in the fixed forgetting factor RLS algorithm. In addition, Kalman gain will tend to zero as the parameter estimates approach their true values. As a result, RLS algorithm will eventually lose its tracking ability in the time - varying system. Now the proposed algorithm can solve this problem. In the end of this paper, the modified RLS algorithm is simulated in the MATLAB platform. The simulation results prove that this algorithm has high convergence velocity, high tracking velocity and small stationary error.
出处
《计算机仿真》
CSCD
北大核心
2009年第8期345-348,共4页
Computer Simulation
