A Comprehensive Multiobjective Optimal Sensor Placement Method for Response Reconstruction

In aeronautical engineering, vibration response reconstruction is a critical yet challenging task. Despite some progress in recent years, existing methods still perform poorly due to a lack of consideration of effective sensor placement, especially when multiple responses need to be reconstructed simultaneously. In this article, we propose a cost-efficient and effective scheme, sparse parameter decomposition-adaptive response reconstruction (sdaRR) method, which performs optimal sensor placement while reconstructing the multitarget response. First, we propose a multitask learning module to reconstruct multiple responses simultaneously. We also meticulously design a simple yet effective adaptive weighting method to address the issue of reconstruction imbalance in multitask learning. Second, sparse regularization and parameter decomposition modules are introduced into the response reconstruction framework to precisely select the optimal sensors. Specifically, we explicitly separate the parameters into task-shared and task-specific regularization terms, which not only identify task-shared sensors but also task-specific sensors. This enhances the interpretability and reliability of multitarget response reconstruction, further improving accuracy. Extensive experiments conducted on both the standard aircraft and the large passenger aircraft models validate the effectiveness of our method. Compared with state-of-the-art methods, it demonstrates more robust reconstruction performance while maintaining strong interpretability. This can provide reliable insights into optimal sensor placement in response reconstruction.