The specific characteristics of remote sensing images, such as large directional variations, large target sizes, and dense target distributions, make target detection a challenging task. To improve the detection performance of models while ensuring real-time detection, this paper proposes a lightweight object detection algorithm based on an attention mechanism and YOLOv5s. Firstly, a depthwise-decoupled head (DD-head) module and spatial pyramid pooling cross-stage partial GSConv (SPPCSPG) module were constructed to replace the coupled head and the spatial pyramid pooling-fast (SPPF) module of YOLOv5s. A shuffle attention (SA) mechanism was introduced in the head structure to enhance spatial attention and reconstruct channel attention. A content-aware reassembly of features (CARAFE) module was introduced in the up-sampling operation to reassemble feature points with similar semantic information. In the neck structure, a GSConv module was introduced to maintain detection accuracy while reducing the number of parameters. Experimental results on remote sensing datasets, RSOD and DIOR, showed an improvement of 1.4% and 1.2% in mean average precision accuracy compared with the original YOLOv5s algorithm. Moreover, the algorithm was also tested on conventional object detection datasets, PASCAL VOC and MS COCO, which showed an improvement of 1.4% and 3.1% in mean average precision accuracy. Therefore, the experiments showed that the constructed algorithm not only outperformed the original network on remote sensing images but also performed better than the original network on conventional object detection images.

A new Precise Point Positioning (PPP) service, called the PPP-B2b service, has been implemented in the BeiDou-3 Navigation Satellite System (BDS-3), which brings new opportunities for time transfer. However, the solution using the traditional PPP method with the PPP-B2b correction still absorbs some unknown errors and needs reconverging when there exist abnormal data. We developed a new receiver clock model to improve PPP time transfer using the PPP-B2b correction. The traditional PPP time transfers using PPP-B2b with BDS-3, Global Positioning System (GPS), and BDS-3/GPS (Scheme1) are compared with the corresponding time transfer with the proposed clock model (Scheme2). The results show that GPS-only PPP is not recommended because of low accuracy of 2 ns. BDS-3 or BDS-3/GPS PPP time transfers in Scheme1 can realize about 0.2 ns accuracy. When the new clock model is applied, the accuracy can be improved by up to 45% and 39.8% for BDS-3 and BDS-3/GPS PPP, respectively. The proposed clock model can significantly improve the short-term frequency stability by 57.4%, but less for the long-term stability.

With the increasing demand for location-based services (LBS) in daily production and life, indoor positioning has gradually become one of the current research hotspots. The most pressing challenge in indoor positioning is achieving high-precision location in complex indoor environments. Aiming at the problem that severe non-line-of-sight (NLOS) interference with UWB positioning accuracy in complex indoor environments, this article proposes a novel indoor UWB NLOS correction positioning method based on anchor LOS/NLOS map, which utilizes prior environmental information. After anchors are deployed, the LOS/NLOS map is generated using the prior environmental information. Then the line-of-sight (LOS) anchors are selected for position solution, and the observations under the influence of the NLOS anchors are corrected. Then the trajectory is optimized using a Map-Modified Kalman filtering algorithm, and it adopts a method to solve the position of two LOS anchors by using trajectory constraints, which can deal with more complex indoor environment. The experimental results show that the NLOS correction method in this article has better accuracy than the traditional filtering algorithm, especially in NLOS environment with few anchors. Compared with the Standard Kalman filtering algorithm, the average positioning accuracy is improved by 69.30%, and compared with the robust Kalman filtering algorithm, the average positioning accuracy is improved by 33.77%.

Ultra-wideband (UWB) time-of-flight (TOF)-based ranging information in a non-line-of-sight (NLOS) environment can display significant forward errors, which directly affect positioning performance. NLOS has been a major factor limiting the improvement of UWB positioning accuracy and its application in complex scenarios. Therefore, in order to weaken the influence of the indoor complex environment on the NLOS environment of UWB and to further improve the performance of positioning, in this paper, we first analyze the factors and characteristics of NLOS formation in an indoor environment. The NLOS is divided into fixed NLOS influenced by spatial structure and dynamic random NLOS influenced by human occlusion. Then, the anchor LOS/NLOS information map is established by making full use of indoor spatial a priori information. On this basis, a robust adaptive extended Kalman filtering algorithm based on the anchor LOS/NLOS information map is designed, which is not only effectively able to exclude the influence of spatial NLOS, but can also optimize the random error. The proposed algorithm was validated in different experimental scenarios. The experimental results show that the positioning accuracy is better than 0.32 m in complex indoor NLOS environments.

This work presents a novel RGB-D dynamic Simultaneous Localisation and Mapping (SLAM) method that improves the precision, stability, and efficiency of localisation while relying on lightweight deep learning in a dynamic environment compared to the traditional static feature-based visual SLAM algorithm. Based on ORB-SLAM3, the GCNv2-tiny network instead of the ORB method, improves the reliability of feature extraction and matching and the accuracy of position estimation; then, the semantic segmentation thread employs the lightweight YOLOv5s object detection algorithm based on the GSConv network combined with a depth image to determine potentially dynamic regions of the image. Finally, to guarantee that the static feature points are used for position estimation, dynamic probability is employed to determine the true dynamic feature points based on the optical flow, semantic labels, and the state in last frame. We have performed experiments on the TUM datasets to verify the feasibility of the algorithm. Compared with the classical dynamic visual SLAM algorithm, the experimental results demonstrate that the absolute trajectory error is greatly reduced in dynamic environments, and that the computing efficiency is improved by 31.54% compared with the real-time dynamic visual SLAM algorithm with close accuracy, demonstrating the superiority of DLD-SLAM in accuracy, stability, and efficiency.