Deep-Learning-Driven Noise-Adaptive Filtering for Multisensor Fusion under Nonstationary Noise Environments

In dynamic systems, complex time-varying noise characteristics impose challenges for multisensor data fusion. This article proposes a deep-learning-driven noise-adaptive filtering approach for multisensor data fusion. A parallel gated recurrent unit (GRU) and 1-D convolutional neural network (1-D-CNN) architecture jointly extracts temporal dependencies and local spatial patterns from raw measurement sequences to estimate sensor noise variances. Each sensor then uses these variance estimates for more accurate adaptive local filtering. Based on these refined local estimates, the fusion center employs an event-triggered scheme to drastically cut communications while preserving fusion accuracy. Comprehensive experiments in a high-fidelity Unreal Engine-AirSim quadrotor uncrewed aerial vehicle (UAV) simulation under nominal, drift, abrupt, and extreme noise modes confirm that our pipeline achieves superior global state estimation while dramatically reducing data transmission burden.