Filtered Extended Observer for Building-like structure Under Disturbed Sampled Outputs
摘要
Building-like structures under seismic excitation are often monitored using sensors with limited sampling rates and noisy measurements. In this context, an accurate reconstruction of both the full state vector and unknown external disturbances is critical for health monitoring and displacement control, however it remains challenging when only sampled and noisy outputs are available.
PurposeThis work aims to design and analyze a continuous-time filtered extended state observer for multiple-input multiple-output (MIMO) linear systems with sampled and noisy output measurements, with special emphasis on building-like structures subjected to seismic excitation. The goal is to achieve simultaneous estimation of the system states and an unknown external input without requiring system discretization or switching schemes.
MethodsA continuous-time filtered extended state observer is proposed that incorporates sampled output measurements into a filtering and estimation structure. The observer considers the unknown external input as an extended state and reconstructs it together with the original state vector. A Lyapunov-based analysis is performed to derive sufficient conditions that guarantee that the observation error converges in a bounded region, explicitly characterizing its dependence on the sampling interval and the limits of the measurement disturbances. Robustness with respect to bounded noise and varying sampling rates is formally established.
ResultsThe proposed methodology is evaluated in a two-story building model subjected to seismic excitation in different sampling scenarios. The numerical simulations show an accurate reconstruction of both the structural states and the seismic disturbance, even in the presence of significant measurement noise and relatively long sampling periods. The results confirm that the error limits increase in accordance with the sampling interval and the amplitude of the disturbance, in accordance with the theoretical analysis.
ConclusionsThe filtered extended state observer provides an alternative for the simultaneous continuous estimation of state and disturbances in linear MIMO systems using only sampled and noisy outputs, without the need to discretize the system or apply switching strategies. Theoretical and numerical results indicate that the approach is robust to measurement noise and low sampling frequencies, making it a promising tool for structural state monitoring and control in building-like systems subjected to seismic excitation.