To meet the realistic demands of autonomous transportation for quadrotor suspended-payload systems, it is necessary to investigate load trajectory tracking control using only onboard sensing. This paper proposes a nonlinear payload trajectory tracking strategy based on the force sensor. The system employs a force sensor to measure the tension on the cable, in conjunction with the position of the quadrotor, to provide a real-time estimate of the payload position. A payload velocity observer is designed to ensure its status is always available. Furthermore, in order to meet the needs of transporting objects of varying weights, the system utilizes data feedback from the force sensor to autonomously adapt to various payload masses. Based on the payload state estimation, a payload trajectory tracking controller that only relies on onboard sensing to achieve outdoor operation is designed. And multiple sets of simulations and real-world outdoor experiments are performed to validate the effectiveness of the controller. Experimental results show that the proposed control strategy achieves satisfied performance for different payloads and different trajectories.

Global navigation satellite system (GNSS), inertial measurement unit (IMU), and light detection and ranging (LiDAR), as common sensors in navigation systems, are widely used in fusion simultaneous localization and mapping (SLAM) to improve the performance. However, strong robustness and high efficiency are not received much attention, posing challenges for ubiquitous application of robotics systems. This article presents a robust and fast fusion odometry with inertial navigation system (INS)-centric multiple modalities. Specifically, our framework is based on two subsystems: the GNSS-IMU subsystem estimates coarse states and provides global constraints, the LiDAR-IMU subsystem refines states and constructs the 3-D map. The state estimation is performed by iterated error state Kalman filter without interruption even if partial sensor fails. Moreover, a fast and dynamic frame transformation method and distributed backward propagation are designed for linking world-frame states with global measurements and coping with asynchronism data. Extensive qualitative and quantitative experiments are also conducted on public datasets and in real-world scenarios. The results show that our system exhibits competitive accuracy and efficiency, with strong robustness against sensor degradation and failure.

Hyperspectral image change detection (HSIs-CD) often faces the challenge of limited sample sizes, and labeling data is both time-consuming and labor-intensive. Foundation models leverage extensive unlabeled data for self-supervised generative pre-training, allowing the model to learn rich data representations. However, few models have been specifically designed for HSIs, and existing methods often rely on pre-training datasets that are limited to data from a small number of satellite sensors. This limitation affects generalization, especially when there are significant differences between data from different sensors. Moreover, the difference map (DMP) of bi-temporal HSIs is often used as input to the networks. While the DMP-based approach reduces FLOPs, it may lead to information loss compared to dual-branch networks. In this letter, we propose a mini-patch-based generative pre-trained spectral-spatial transformer (GPSST) for semi-supervised HSIs-CD. We begin by collecting public HSIs datasets and dividing them into thousands of patches. Each patch is then split into spectral-spatial tokens, with a portion of these tokens masked and used as input for the GPSST. We then design a spectral-spatial masked autoencoder as the backbone of GPSST for self-supervised generative learning. Finally, we fine-tune the GPSST encoder using a small number of labeled patches and design a principal component analysis (PCA) branch to compensate for the information loss caused by the DMP. Our experiments demonstrate that GPSST outperforms existing methods, achieving superior accuracy in HSIs-CD.

Change detection (CD) is a technique widely used in remote sensing for identifying the differences between data acquired at different times. Most existing 3D CD approaches voxelize point clouds into 3D grids, project them into 2D images, or rasterize them into digital surface models due to the irregular format of point clouds and the variety of changes in three-dimensional (3D) objects. However, the details of the geometric structure and spatiotemporal sequence information may not be fully utilized. In this article, we propose PBFormer, a transformer network with Siamese architecture, for directly inferring pointwise changes in bi-temporal 3D point clouds. First, we extract point sequences from irregular 3D point clouds using the k-nearest neighbor method. Second, we uniquely use a point transformer network as an encoder to extract point feature information from bitemporal 3D point clouds. Then, we design a module for fusing the spatiotemporal features of bi-temporal point clouds to effectively detect change features. Finally, multilayer perceptrons are used to obtain the CD results. Extensive experiments conducted on the Urb3DCD benchmark show that PBFormer outperforms other excellent approaches for 3D point cloud CD tasks.