The key to successful positioning of autonomous mobile robots in complicated indoor environments lies in the strong anti-interference of the positioning system and accurate measurements from sensors. Inertial navigation systems (INS) are widely used for indoor mobile robots because they are not susceptible to external interferences and work properly, but the positioning errors may be accumulated over time. Thus ultra wideband (UWB) is usually adopted to compensate the accumulated errors due to its high ranging precision. Unfortunately, UWB is easily affected by the multipath effects and non-line-of-sight (NLOS) factor in complex indoor environments, which may degrade the positioning performance. To solve above problems, this paper proposes an effective system framework of INS/UWB integrated positioning for autonomous indoor mobile robots, in which our modeling approach is simple to implement and a Sage–Husa fuzzy adaptive filter (SHFAF) is proposed. Due to the favorable property (i.e., self-adaptive adjustment) of SHFAF, the difficult problem of time-varying noise in complex indoor environments is considered and solved explicitly. Moreover, outliers can be detected and corrected by the proposed sliding window estimation with fading coefficients. This facilitates the positioning performance improvement for indoor mobile robots. The benefits of what we propose are illustrated by not only simulations but more importantly experimental results.

Remote sensing scene classification (RSSC) faces severe challenges from data scarcity and complex background clutter. To overcome these limitations, this paper draws inspiration from few-shot fine-grained image classification (FSFGIC) to filter noise and capture subtle details. However, existing methods often process global context and local features separately, which limits their ability to suppress background noise in complex scenes. Consequently, the Guided Relational Cross-Attention Dual-branch Network (GRCD-Net) is proposed. Its core Guided Relational Cross-Attention (GRC) block leverages global semantics to filter local background noise prior to bidirectional feature interaction. Additionally, Iterative Global Relation (IGR) and Patch-level Dual-Metric (PDM) modules are integrated to robustly refine global relations and capture local similarities. Extensive experiments demonstrate that GRCD-Net consistently outperforms baselines by 2–4% on standard FSFGIC benchmarks. Notably, on the challenging NWPU-RESISC45 RSSC dataset, it achieves an 81.39% one-shot accuracy and exceeds current state-of-the-art methods by 7.55%, validating its efficacy for complex Earth observation.