This article studies the energy efficiency optimization of cargo uncrewed aerial vehicle (UAV) pickup systems, with constraints on on-board energy and load capacity. In the UAV-enabled cargo pickup system, minimizing the total energy consumption and ensuring the safe flight of the cargo UAV is a problem to be solved. However, due to building blockages, the channel between the UAV and ground base stations (GBSs) frequently switches between line-of-sight (LoS) and non-line-of-sight (NLoS), thereby affecting the UAV’s communication quality. This effect is further aggravated by environmental noise interference. Moreover, limited by the on-board energy, it is unrealistic for the UAV to pick up all the cargo in a single flight without charging or replacing the battery. To address the above-mentioned challenges, we propose a UAV pickup system energy efficiency optimization (UPSEEO) framework. In this framework, the UAV’s trajectory between any two pickup points is optimized via the A${}^{*}$algorithm to ensure the stability of the UAV communication link. Next, we employ the particle swarm optimization (PSO) algorithm to optimize both task allocation and flight speed to minimize the total energy consumption, subject to constraints on UAV on-board energy limits and payload capacity. Numerical results show that the proposed framework can ensure the UAV’s communication quality in any spatial topology, with an improvement in energy efficiency of approximately 5% to 50% compared to the comparison experiment.

Classical Transformer-based line segment detection methods have delivered impressive results. However, we observe that some accurately detected line segments are assigned low confidence scores during prediction, causing them to be ranked lower and potentially suppressed. Additionally, these models often require prolonged training periods to achieve strong performance, largely due to the necessity of bipartite matching. In this paper, we introduce RANK-LETR, a novel Transformer-based line segment detection method. Our approach leverages learnable geometric information to refine the ranking of predicted line segments by enhancing the confidence scores of high-quality predictions in a posterior verification step. We also propose a new line segment proposal method, wherein the feature point nearest to the centroid of the line segment directly predicts the location, significantly improving training efficiency and stability. Moreover, we introduce a line segment ranking loss to stabilize rankings during training, thereby enhancing the generalization capability of the model. Experimental results demonstrate that our method outperforms other Transformer-based and CNN-based approaches in prediction accuracy while requiring fewer training epochs than previous Transformer-based models.

In this paper, we propose a novel transformer-based end-to-end real-time vanishing point detection method, which is named Vanishing Point TRansformer (VPTR). The proposed method can directly regress the locations of vanishing points from given images. To achieve this goal, we pose vanishing point detection as a point object detection task on the Gaussian hemisphere with region division. Considering low-level features always provide more geometric information which can contribute to accurate vanishing point prediction, we propose a clear architecture where vanishing point queries in the decoder can directly gather multi-level features from CNN backbone with deformable attention in VPTR. Our method does not rely on line detection or Manhattan world assumption, which makes it more flexible to use. VPTR runs at an inferring speed of 140 FPS on one NVIDIA 3090 card. Experimental results on synthetic and real-world datasets demonstrate that our method can be used in both natural and structural scenes, and is superior to other state-of-the-art methods on the balance of accuracy and efficiency.

Radar signal interpretation plays a crucial role in remote detection and ranging. With the gradual display of the advantages of neural network technology in signal processing, learning-based radar signal interpretation is becoming a research hot-spot and made great progress. And since radar semantic segmentation (RSS) can provide more fine-grained target information, it has become a more concerned direction in this field. However, the temporal information, which is an important clue for analyzing radar data, has not been exploited sufficiently in present RSS frameworks. In this work, we propose a novel temporal information learning paradigm, i.e., data-driven temporal information aggregation with learned target-history relations. Following this idea, a flexible learning module, called Temporal Relation-Aware Module (TRAM) is carefully designed. TRAM contains two main blocks: i) an encoder for capturing the target-history temporal relations (TH-TRE) and ii) a learnable temporal relation attentive pooling (TRAP) for aggregating temporal information. Based on TRAM, an end-to-end Temporal-Aware RSS Network (TARSS-Net) is presented, which has outstanding performance on publicly available and our collected real-measured datasets. Code and supplementary materials are available at https://github.com/zlw9161/TARSS-Net.

In this paper we propose a new named entity recognition (NER) model that combines negative sampling and reinforcement learning to reduce the impact of unlabeled entities and make full use of the labeled entities. Generally, there are a small number of correct entities in the training data which are not labeled inevitably. These unlabeled entities will greatly affect the performance of NER models. So in our model reinforcement learning is exploited to select the negative instances correctly and avoid those unlabeled entities for NER classifier because unlabeled entities are noisy data for negative instances. Then a NER classifier is trained on these negative instances and the manually labeled entities. Thus our model can be trained in the right direction under the guidance of the labeled entities and eliminate the influence of unlabeled entities. To evaluate our model, experiments are conducted on three public datasets, BioNLP11EPI, BioNLP13CG and BioNLP13GE, and with two kinds of embeddings, BERT and word2vec. Experiments show that our model can outperform the most of other models, and achieve comparable performance to the PFT model. Notably PFT need collect a large number of related sentences from unlabeled documents to pre-fine-tune BERT while our model need no additional corpus for training. Meanwhile when some entities are masked randomly, the performance of our model is significantly better than negative sampling model.