分区频域卡尔曼滤波(Partitioned block frequency domain Kalman filtering,PBFDKF)因其收敛速度快、稳态误差小的优势被应用在自适应滤波声反馈抑制(Adaptive feedback cancellation,AFC)。然而,当声反馈路径发生突变时,卡尔曼滤波会...分区频域卡尔曼滤波(Partitioned block frequency domain Kalman filtering,PBFDKF)因其收敛速度快、稳态误差小的优势被应用在自适应滤波声反馈抑制(Adaptive feedback cancellation,AFC)。然而,当声反馈路径发生突变时,卡尔曼滤波会进入锁死状态,难以再次跟踪。本文提出一种融合神经网络的卡尔曼滤波啸叫抑制状态检测算法(Kalman⁃filter⁃based AFC with state detection model,KFSD)。该系统将卡尔曼滤波声反馈抑制系统的传声器采集信号、残差信号和滤波器更新量作为输入特征,通过神经网络对卡尔曼滤波的状态误差协方差矩阵进行修正,从而实现路径突变情况下的再次跟踪和收敛。仿真实验结果验证了所提算法具有较高的正判率、较低的虚警率和较短的延迟帧数,算法同时具备快速再跟踪性能,提高了声反馈抑制效果。展开更多
An adaptive output feedback neural network tracking controller is designed for a class of unknown output feedback nonlinear time-delay systems by using backstepping technique. Neural networks are used to approximate u...An adaptive output feedback neural network tracking controller is designed for a class of unknown output feedback nonlinear time-delay systems by using backstepping technique. Neural networks are used to approximate unknown time-delay functions. Delay-dependent filters are introduced for state estimation. The domination method is used to deal with the smooth time-delay basis functions. The adaptive bounding technique is employed to estimate the upper bound of the neural network reconstruction error. Based on Lyapunov-Krasoviskii functional, the semi-global uniform ultimate boundedness (SGUUB) of all the signals in the closed-loop system is proved. The arbitrary output tracking accuracy is achieved by tuning the design parameters and the neural node number.The feasibility is investigated by an illustrative simulationexample.展开更多
文摘分区频域卡尔曼滤波(Partitioned block frequency domain Kalman filtering,PBFDKF)因其收敛速度快、稳态误差小的优势被应用在自适应滤波声反馈抑制(Adaptive feedback cancellation,AFC)。然而,当声反馈路径发生突变时,卡尔曼滤波会进入锁死状态,难以再次跟踪。本文提出一种融合神经网络的卡尔曼滤波啸叫抑制状态检测算法(Kalman⁃filter⁃based AFC with state detection model,KFSD)。该系统将卡尔曼滤波声反馈抑制系统的传声器采集信号、残差信号和滤波器更新量作为输入特征,通过神经网络对卡尔曼滤波的状态误差协方差矩阵进行修正,从而实现路径突变情况下的再次跟踪和收敛。仿真实验结果验证了所提算法具有较高的正判率、较低的虚警率和较短的延迟帧数,算法同时具备快速再跟踪性能,提高了声反馈抑制效果。
文摘An adaptive output feedback neural network tracking controller is designed for a class of unknown output feedback nonlinear time-delay systems by using backstepping technique. Neural networks are used to approximate unknown time-delay functions. Delay-dependent filters are introduced for state estimation. The domination method is used to deal with the smooth time-delay basis functions. The adaptive bounding technique is employed to estimate the upper bound of the neural network reconstruction error. Based on Lyapunov-Krasoviskii functional, the semi-global uniform ultimate boundedness (SGUUB) of all the signals in the closed-loop system is proved. The arbitrary output tracking accuracy is achieved by tuning the design parameters and the neural node number.The feasibility is investigated by an illustrative simulationexample.
