Fuzzy Sliding Mode Controllers Synthesis through Genetic Optimization
Fuzzy Sliding Mode Controllers Synthesis through Genetic Optimization
复制标题
通过遗传优化综合模糊滑模控制器
DOI：
10.1007/978-94-010-0324-7_23
复制
发表时间：
2002
期刊：
影响因子：
--
通讯作者：
D. Naso
中科院分区：
文献类型：
--
作者：
M. Dotoli;B. Maione;D. Naso
本文介绍了一种应用遗传算法（GAs）的模糊滑模（FSM）控制器的设计过程。有限状态机控制器成功地减少了滑模控制器的计算量和典型的抖振现象。然而，有限状态机控制器的综合是困难的，因为没有一般的标准来选择定义整体控制速度的滑动参数，以及输入/输出增益缩放控制器的输入/输出信号。目前，由于缺乏以适当标准为基础的准则，这些参数仍然是通过试错程序调整的。本文提出了一种自动化的方法，使用遗传算法来设计有限状态机控制器。该方法说明了两个不同的有限状态机控制器的倒立摆的设计。在解决方案空间与经验选择的界限的遗传搜索导致的解决方案，大大提高了手动调整获得的性能。使用一种新型的遗传算法与自适应分辨率编码保证了最终解决方案的质量，平衡的计算工作。
The paper describes a procedure applying genetic algorithms (GAs) to the design of fuzzy sliding mode (FSM) controllers. FSM controllers are successful in reducing computational effort and typical chattering phenomena of sliding mode (SM) controllers. However, the synthesis of FSM controllers is difficult, because there are no general criteria for choosing the sliding parameter defining the overall control speed, and the input/output gains scaling the input/output signals of the controller. At present, due to the lack of guidelines based on appropriate criteria, these parameters are still tuned by means of a trial and error procedure. This paper proposes an automated approach using GAs to design FSM controllers. The method is illustrated with the design of two different FSM controllers for an inverted pendulum. A genetic search in a solution space with bounds selected empirically leads to a solution improving considerably the performance obtained with manual tuning. The use of a new type of GA with adaptive resolution encoding guarantees a quality of the final solution, balancing the computational efforts.