论文部分内容阅读
针对模糊辨识中采用迭代和人为决策法确定模糊规则数时易受噪声和人为因素的影响,而导致算法鲁棒性较差和计算量较高的问题,提出一种基于改进客观聚类分析的模糊辨识方法.首先引入并改进了客观聚类分析法,克服了迭代导致的规则数冗余,降低了人为因素对聚类结果的影响,从而减小了计算量并提高了鲁棒性;然后结合模糊聚类和稳态卡尔曼滤波法,分别辨识了前提和结论参数;最后通过Box-Jenkins仿真实例验证了所提方法的有效性.
Aiming at the problem that fuzzy and fuzzy rules are easily influenced by noise and human factors in iterative and artificial decision making method, the robustness of the algorithm is poor and the computational complexity is high. A new algorithm based on improved objective clustering analysis Fuzzy identification method.Firstly, objective clustering analysis method is introduced and improved to overcome the rule number redundancy caused by iteration and reduce the influence of human factors on the clustering results, thus reducing the computational complexity and improving the robustness. Then Combined with fuzzy clustering and steady-state Kalman filtering, the preconditions and the conclusion parameters are identified respectively. Finally, the effectiveness of the proposed method is verified by Box-Jenkins simulation.