胰岛素基础率是人工胰腺系统实现人体血糖闭环控制的基准,但该变量在临床治疗中难以准确确定.针对这一问题,本文设计了一种基于胰岛素基础率动态估计的人工胰腺自抗扰控制方法,通过扩张状态观测器(Extended state observer,ESO)实时估...胰岛素基础率是人工胰腺系统实现人体血糖闭环控制的基准,但该变量在临床治疗中难以准确确定.针对这一问题,本文设计了一种基于胰岛素基础率动态估计的人工胰腺自抗扰控制方法,通过扩张状态观测器(Extended state observer,ESO)实时估计血糖代谢过程中的内部与外界干扰,构建具备参数自适应能力的反馈控制律和胰岛素注射安全约束,实现血糖闭环调控能力的有效改善.在此基础上,本文设计了基于移动设备和蓝牙模块的人工胰腺软件系统,并通过美国食品药品监督管理局(Food and Drug Administration,FDA)接受的UVA/Padova T1DM仿真平台完成算法的比较仿真与功能测试.本文的工作将为后续人工胰腺临床试验的开展提供方法基础和技术支持,也为我国糖尿病患者血糖管理的改善提供精准医学治疗手段.展开更多
A simple identification method based on a closed-loop experiment is proposed to measure the infinity norm of sensitivity function.A chirp signal,modified to have desired band-limited characteristic and finite duration...A simple identification method based on a closed-loop experiment is proposed to measure the infinity norm of sensitivity function.A chirp signal,modified to have desired band-limited characteristic and finite duration,is used as the excitation in the experiment,and the sensitivity function is calculated using Fourier transform of input and error signals before the infinity norm is evaluated through maximization of the magnitude of sensitivity function.With an additional feature of providing values of gain margin and phase margin at a little extra effort,this method can be used in the identification step of a controller auto-tuning procedure,as having been supported by simulation results showing its capability of providing fast and accurate estimates for a large variety of processes.展开更多
A plant-friendly proportional-integral-derivative （PID） controller optimization framework is proposed to make tradeoffs among set-point response,controller output variations and robustness.The objective function is ...A plant-friendly proportional-integral-derivative （PID） controller optimization framework is proposed to make tradeoffs among set-point response,controller output variations and robustness.The objective function is chosen as the weighted sum of the integral of squared time-weighted error and the integral of squared timeweighted derivative of the control variable with respect to set-point response,while the robustness of the system is guaranteed by constraints on gain and phase margins.Due to the complex structure of the constraints,the problem is solved by genetic algorithms.Simulation analysis show the proposed method could efficiently reduce the controller output variations while maintaining a short settling time.Based on the simulation results,iterative tuning rules for the weighting factor in the objective function are obtained,which allows efficient simple proportional-integral（PI） tuning formulae to be derived.展开更多
文摘胰岛素基础率是人工胰腺系统实现人体血糖闭环控制的基准,但该变量在临床治疗中难以准确确定.针对这一问题,本文设计了一种基于胰岛素基础率动态估计的人工胰腺自抗扰控制方法,通过扩张状态观测器(Extended state observer,ESO)实时估计血糖代谢过程中的内部与外界干扰,构建具备参数自适应能力的反馈控制律和胰岛素注射安全约束,实现血糖闭环调控能力的有效改善.在此基础上,本文设计了基于移动设备和蓝牙模块的人工胰腺软件系统,并通过美国食品药品监督管理局(Food and Drug Administration,FDA)接受的UVA/Padova T1DM仿真平台完成算法的比较仿真与功能测试.本文的工作将为后续人工胰腺临床试验的开展提供方法基础和技术支持,也为我国糖尿病患者血糖管理的改善提供精准医学治疗手段.
基金Sponsored by the Key Construction Program of the"985"Program (1010012047201)
文摘A simple identification method based on a closed-loop experiment is proposed to measure the infinity norm of sensitivity function.A chirp signal,modified to have desired band-limited characteristic and finite duration,is used as the excitation in the experiment,and the sensitivity function is calculated using Fourier transform of input and error signals before the infinity norm is evaluated through maximization of the magnitude of sensitivity function.With an additional feature of providing values of gain margin and phase margin at a little extra effort,this method can be used in the identification step of a controller auto-tuning procedure,as having been supported by simulation results showing its capability of providing fast and accurate estimates for a large variety of processes.
基金Sponsored by the Key Construction Program of the"985"Program (1010012047201)
文摘A plant-friendly proportional-integral-derivative （PID） controller optimization framework is proposed to make tradeoffs among set-point response,controller output variations and robustness.The objective function is chosen as the weighted sum of the integral of squared time-weighted error and the integral of squared timeweighted derivative of the control variable with respect to set-point response,while the robustness of the system is guaranteed by constraints on gain and phase margins.Due to the complex structure of the constraints,the problem is solved by genetic algorithms.Simulation analysis show the proposed method could efficiently reduce the controller output variations while maintaining a short settling time.Based on the simulation results,iterative tuning rules for the weighting factor in the objective function are obtained,which allows efficient simple proportional-integral（PI） tuning formulae to be derived.
