帕金森状态的慢变量反馈模糊控制  

Closed-loop fuzzy control of Parkinsonian state based on slow variable

作　　者：刘晨[1,3] 王江[1] 邓斌[1] 魏熙乐[1] 于海涛[1] 李会艳[2] 

机构地区：[1]天津大学电气与自动化工程学院,天津300072 [2]天津职业技术师范大学自动化与电气工程学院,天津300222 [3]Department of Electrical Engineering and Computer Science,Case Western Reserve University,Cleveland 44106,USA

出　　处：《中国科学：信息科学》2015年第3期439-456,共18页Scientia Sinica(Informationis)

基　　金：国家自然科学基金(批准号:61172009,61374182);国家留学基金委国家建设高水平大学公派研究生项目资助

摘　　要：闭环深度脑刺激是实现帕金森状态控制的有效方法,但是如何获得反馈信号和控制技术是尚未解决的问题.本文提出基于慢变量反馈的变论域模糊控制策略去实现帕金森状态的控制.针对高度非线性神经系统,本文以控制信号的能量优化为控制目标,旨在通过低能耗的外部刺激手段改善丘脑神经元的放电模式,提出了无香卡尔曼滤波器和变论域模糊控制相结合的控制策略,实现了丘脑神经元帕金森状态的闭环控制.采用慢变量作为反馈变量,大大降低了控制信号的波动性和能量损耗.定性分析和定量指标均证明了基于慢变量反馈的变论域模糊控制策略的有效性.Closed-loop deep brain stimulation is an effective method for controlling the Parkinsonian state. However, the twin issues of how to obtain a suitable feedback variable and design a high-performance control strategy are still unresolved. This paper proposes a variable universe fuzzy closed-loop control method based on slow variable to modulate the abnormal Parkinsonian state. For highly nonlinear neural systems, in order to achieve energy optimization of the control signal, this paper designs a closed-loop control strategy of thalamic neurons by combining unscented Kalman filter with variable universe fuzzy control, with the objective of improving the firing patterns of thalamic neurons via external stimuli with lower energy consumption. Using a slow variable as the feedback variable significantly decreases the fluctuations and energy expenditure of the stimuli. Qualitative and quantitative analyses conducted demonstrate that the proposed variable universe fuzzy closed-loop control strategy based on slow variable is effective.

关 键 词：帕金森状态 变论域 模糊控制 慢变量 无香卡尔曼滤波器 

分 类 号：TP273.4[自动化与计算机技术—检测技术与自动化装置]