基于粒度计算的二级倒立摆模糊控制策略

Fuzzy Control Strategy of Double Inverted Pendulum Based on Granular Computing

二级倒立摆系统是一个多变量、非线性、不稳定的快速系统,在对其进行传统的模糊控制时,由于多变量输入而容易造成'规则爆炸'.为解决这一问题,提出了基于粒度计算的二级倒立摆模糊控制策略.根据粒度空间结构,将误差相空间划分为不同的粒元,从而构成粗粒度层和细粒度层,在不同的粒度层设计不同的控制策略,即在粗粒度层次上采用模糊控制实现系统的鲁棒性;在细粒度层次上采用PID(Proportion-Integral-Derivative)控制对系统进行进一步微调,逐步求精,最终获得理想的控制性能.实验证明:该控制策略不仅可以解决模糊控制规则呈指数爆炸的问题,而且还可以兼顾倒立摆系统的快速性和准确性指标.

Double inverted pendulum is a kind of multi-variable, nonlinear and fast system. When it is controlled by fuzzy controllers, ‘rule explosion＇ is caused by many inputs. To solve the problem, fuzzy control strategy is provided based on granular computing. According to the granular theory, error space is divided into different granular elements to constitute coarser granularity and finer granularity. Fuzzy control at the coarser level is adopted to realize robustness, while Proportional-Integral-Derivation （PID） control at the finer level to realize delicate adjustment. The simulations testify that the strategy is effective not only solving the problem of‘rule explosion’ , but also achieving the ideal performance of stable state and transient state. Key words：