This paper proposes a new method to detect and identify foreign matter mixed in a plastic bottle filled with transfusion solution. A spin-stop mechanism and mixed illumination style are applied to obtain high contrast images between moving foreign matter and a static transfusion background. The Gaussian mixture model is used to model the complex background of the transfusion image and to extract moving objects. A set of features of moving objects are extracted and selected by the ReliefF algorithm, and optimal feature vectors are fed into the back propagation (BP) neural network to distinguish between foreign matter and bubbles. The mind evolutionary algorithm (MEA) is applied to optimize the connection weights and thresholds of the BP neural network to obtain a higher classification accuracy and faster convergence rate. Experimental results show that the proposed method can effectively detect visible foreign matter in 250-mL transfusion bottles. The misdetection rate and false alarm rate are low, and the detection accuracy and detection speed are satisfactory.

An autonomous place recognition system is essential for scenarios where GPS is useless, such as underground tunnels. However, it is difficult to use existing algorithms to fully utilize the small number of effective features in underground tunnel data, and recognition accuracy is difficult to guarantee. In order to solve this challenge, an efficient point cloud position recognition algorithm, named Dual-Attention Transformer Network (DAT-Net), is proposed in this paper. The algorithm firstly adopts the farthest point downsampling module to eliminate the invalid redundant points in the point cloud data and retain the basic shape of the point cloud, which reduces the size of the point cloud and, at the same time, reduces the influence of the invalid point cloud on the data analysis. After that, this paper proposes the dual-attention Transformer module to facilitate local information exchange by utilizing the multi-head self-attention mechanism. It extracts local descriptors and integrates highly discriminative global descriptors based on global context with the help of a feature fusion layer to obtain a more accurate and robust global feature representation. Experimental results show that the method proposed in this paper achieves an average $F_1$ score of 0.841 on the SubT-Tunnel dataset and outperforms many existing state-of-the-art algorithms in recognition accuracy and robustness tests.

Aiming at the problem of high cost-effectiveness ratio and redundant resource, a new multi-target coverage algorithm for UAV swarms communication coverage is proposed in this study. Firstly, the proposed algorithm takes each target point as the center and the distance of a single UAV communication coverage as the radius. Then, the number of target points in each circle is calculated and the circle with the maximum number is retained. Finally, using means centering and quadratic perturbation, the target points and circle center within this circle are redetermined. For the uncovered target points, the above operations are repeated until all target points are covered or the number of clusters are more than the UAVs. The experiment results show that the proposed algorithm achieves a coverage of nearly 90% random target points, which is 16% higher than conventional algorithms. Therefore, the proposed algorithm is well applied to cooperative communication support and target area coverage in the combat.

In military operational environments, satellite navigation is typically unavailable, and unmanned aerial vehicle swarms must rely on inter-node communication links for distance measurement to achieve relative positioning. When the absolute positions of a subset of nodes are known, this approach can also enable absolute positioning of the entire swarm. Conventional relative positioning methods often overlook the inherent difference in ranging ability, assuming the same distance errors across all nodes. This oversimplification leads to significant positioning inaccuracies and slow convergence. To address these limitations, this paper proposes a distance error variance adjustable relative positioning method, DEVARP. The proposed method assigns varying confidence weights to inter-node distance measurements based on link quality, and solves it by continuous numerical iteration so as to achieve high-speed and high-precision positioning. In the simulation experiments involving a 32-node UAV swarm, the results show that the confidence-adjustable method achieves an absolute positioning error below 1 m with a probability of 91.05%, compared to 76.57% for traditional methods—a 14.48% improvement. Additionally, the proposed method converges within seven iterations with a probability of 94.10%. In contrast, conventional methods achieve the same result in seven steps with a probability of 78.83% —a 17.53% performance gap—which confirms both the accuracy and computational efficiency of the method.

A highly effective control method is very important to guarantee the safety of the formation of flying missions for multiple unmanned aerial vehicles (UAVs), especially in the presence of complex flying environments and actuator constraints. In this regard, this paper investigates the formation tracking control problem of multiple UAVs in the presence of actuator saturation. Firstly, a brand-novel finite-time anti-saturated control scheme is proposed for multiple UAVs to track the desired position commands, wherein the tracking performance is tuned by introducing a logarithmic function-based state-mapping policy. Then, an adaptive scheme based on projection rules is devised to compensate for the negative effects brought by the actuator saturation. Based on the proposed formation tracking controller, the finite-time formation tracking performance tuning and control saturation problems can be addressed simultaneously with a comparatively allowable system robustness. Finally, three groups of illustrative examples are organized to verify the effectiveness of the proposed formation tracking control scheme.