.png){kind=icon}
.png){kind=icon}
Return
Adaptive Dual-Loop Feedforward Control for Microvibration Environment Realization in High-Precision Sensor Testing Systems
.png){kind=icon}
PRE
AI
S
孙
J
K
X
DOI：10.1109/jsen.2026.3706361.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
High-precision sensors, such as accelerometers and inertial measurement units, require stringent microvibration environments to evaluate their performance under realistic operating conditions. Artificially establishing a controlled microvibration environment below the inherent ground vibration level is crucial for sensor calibration and characterization. To address this requirement, this article proposes an approach for creating a microvibration environment using an adaptive dual-loop feedforward control (ADLFC). The first loop employs a ground-vibration feedforward controller, which utilizes the filtered-x least mean square (Fx-LMS) algorithm to update the coefficients of a fixed-structure infinite impulse response (IIR) filter. This loop is designed to enhance the rejection of low-frequency vibration disturbances beyond what is achievable through feedback control alone. The second loop is an input-command feedforward controller that incorporates a radial basis function neural network (RBFNN)-based compensator. This compensator integrates prior physical information with fuzzy logic inference (PIFLI) to enhance the tracking efficiency and accuracy in a microvibration environment. Feedback control using active damping is implemented to suppress the structural resonant peak and ensure closed-loop stability. Furthermore, to mitigate the influence of closed-loop dynamics on the accuracy of microvibration realization, a zero-phase filtering input shaper (ZPFIS) is adopted to preprocess the input commands. Experimental results demonstrate that the proposed control strategy successfully achieves the desired microvibration performance across various test environments within the allowable errors, confirming its potential applicability in high-precision sensor testing and calibration.
Keywords:
Active vibration control
adaptive feedforward control
filtered-x least mean square (Fx-LMS) algorithm
microvibration environment
radial basis function neural network (RBFNN)
Journal
IF:
4.5
Papers:
2.1W
Citations:
7.3W
Organization
Cited Papers