Data streams with varying feature spaces have received extensive attention recently, while the common concept drift in them remains underexplored. Unsupervised concept drift detectors can report potential drifts without class labels, making them suitable for practical scenarios where labeling is usually costly and difficult. However, existing unsupervised detectors usually operate under fixed feature spaces. To address this limitation, a Matching Degree Histogram-based unsupervised detector for data streams with Varying Feature Spaces (MDH-VFS) is proposed. Changes in input features are refined into four scenarios, specifying the sources of concept drifts in such data streams. Based on this, MDH-VFS monitors the distribution of each feature independently using the fix-slide windows model. A matching degree-based histogram (MD-Histogram) supporting online updating is proposed to model data distribution. MD-Histogram requires no prior distributions and captures data change more sensitively than traditional histograms. The dissimilarity between two MD-Histograms is measured by the Hellinger distance, and drift is detected using an adaptive thresholding strategy. Both the drift positions and drift features can be reported. Experimental results show that MDH-VFS can not only effectively detect drifts in data streams with varying feature spaces (achieving average F1-score/MCC above 77% and outperforming nine existing detectors with improvements of at least 43%), but also improve the classification performance of downstream learning algorithms (reaching a maximum average accuracy of 88% and yielding up to 7.23% improvement).

With the continuous development of science and technology, navigation and positioning technology has been applied to all aspects of society. The ultra-wide band (UWB) based positioning technology has real-time indoor and outdoor accurate tracking ability and high positioning accuracy, which has a wide range of military and civilian applications. Despite that, the data will have abnormal fluctuations in the case of strong interference due to the complex and changeable indoor environment, which may affect the accuracy of positioning and even cause serious accidents. In this paper, UWB precise positioning under signal interference is studied. A two-stream 1D residual network (TS-1DRN) model learning location features from multimodal data is proposed where the main network structure is based on ResNet2D, and a precise positioning model based on the two-stream deep residual network with fusion utilization of multimodal data is applied to accurate positioning in abnormal scenarios. Considering that the anchor coordinates and distance can be used to obtain the tag coordinates in physical model, distance data are further added with anchor coordinates as the neural network training inputs into the two-stream network compared with previous studies. The effectiveness of the proposed model is verified by comparing with the classical algorithms commonly used for UWB positioning. The positioning accuracy under NLOS is improved by about 150% in the 3D space, and it also performs well in other dimensions, with the minimum positioning error reduced to 34.9952mm. Furthermore, the data in normal scenarios were also used for training and testing, and the experimental results are also significantly improved, indicating the robustness of the proposed model.