Genetic algorithm (GA) is an effective method for path planning problems. As a powerful variant of GA, island genetic algorithm (IGA) has considerable improvement in performance. In this paper, a new island model of G...
Genetic algorithm (GA) is an effective method for path planning problems. As a powerful variant of GA, island genetic algorithm (IGA) has considerable improvement in performance. In this paper, a new island model of GA is proposed to avoid the premature phenomenon and achieve better efficiency. First, a new method of creating subpopulations is presented based on K-means to expand the searching area for the optima. Meanwhile, recombining subpopulations is proposed as a new strategy to improve the diversity of populations and save computational time. Moreover, a method is designed based on Monte Carlo sampling to handle the uncertainty of maps. Comparative experiments are presented to verify the efficiency of the proposed algorithm. Then, a proper number of samples is found by simulation to balance the accuracy and the time cost of Monte Carlo sampling.
Underwater supporting robots serving as a relay of energy supplements and communication for other underwater equipment are promising for ocean exploration, development, and protection. This paper proposes a novel auto...
Underwater supporting robots serving as a relay of energy supplements and communication for other underwater equipment are promising for ocean exploration, development, and protection. This paper proposes a novel autonomous docking system centered on a designed supporting robotic fish named ‘CourierFish’. Specifically, CourierFish is capable of docking with a surface dock station for supplying itself and docking with a seafloor platform for supporting equipment in the platform. A visual navigation scheme integrating LED and ArUco markers is presented for accurate localization. The control approach for docking motion is also illustrated. Simulations and aquatic experiments are performed to verify the feasibility of the proposed docking system. The obtained results lay a solid foundation for the construction of various underwater equipment and robot networks.
This paper studies the lack of systematic, teaching content and the obsolete teaching mode of the principle of computer composition in most colleges and universities in China. Proposed system-oriented teaching methods...
详细信息
This paper studies the lack of systematic, teaching content and the obsolete teaching mode of the principle of computer composition in most colleges and universities in China. Proposed system-oriented teaching methods and modes. This paper expounds the teaching goal of cultivating system competence based on classroom teaching system, supported by gradient experiment teaching, combining MOOC platform and SPOC teaching mode, and taking the second class as the teaching implementation system expanded.
Networks are ubiquitous in the real world such as social networks and communication networks, and anomaly detection on networks aims at finding nodes whose structural or attributed patterns deviate significantly from ...
详细信息
Single image super-resolution (SISR) algorithms reconstruct high-resolution (HR) images with their low-resolution (LR) counterparts. It is desirable to develop image quality assessment (IQA) methods that can not only ...
详细信息
To solve the problem that the flame area of small targets on ships can not be easily segmented and the flame boundary is obviously affected by other background such as smoke, this experiment improves DeepLabV3+ algori...
详细信息
To solve the problem that the flame area of small targets on ships can not be easily segmented and the flame boundary is obviously affected by other background such as smoke, this experiment improves DeepLabV3+ algorithm from the following three aspects. Firstly, Dense Atrous Spatial Pyramid Pooling (DASPP) was improved to enhance the extraction and utilization of flame boundary and small target flame features. Secondly, in order to avoid the influence of background such as smoke around the flame, the high-level flame features and shallow flame features obtained by DASPP were combined with the improved Attention Mechanism (PE-SCAM) to obtain more context information in spatial and channel dimensions, so as to enhance the processing effect of the model on the flame and its boundaries. Finally, the traditional cross loss function is optimized to reduce the influence caused by the imbalance of positive and negative samples between the flame and the background. Experimental results show that the MIou of the proposed algorithm on the ship flame data set reaches 88.43%, which can realize the accurate segmentation of small target flames and flame boundaries on the basis of ensuring high real-time performance.
In this paper, we focus on the use of multi-modal data to achieve a semantic segmentation of aerial imagery. Thereby, the multi-modal data is composed of a true orthophoto, the Digital Surface Model (DSM) and further ...
详细信息
In this paper, we focus on the use of multi-modal data to achieve a semantic segmentation of aerial imagery. Thereby, the multi-modal data is composed of a true orthophoto, the Digital Surface Model (DSM) and further representations derived from these. Taking data of different modalities separately and in combination as input to a Residual Shuffling Convolutional Neural Network (RSCNN), we analyze their value for the classification task given with a benchmark dataset. The derived results reveal an improvement if different types of geometric features extracted from the DSM are used in addition to the true orthophoto.
Relation detection plays a crucial role in Knowledge Base Question Answering (KBQA) because of the high variance of relation expression in the question. Traditional deep learning methods follow an encoding-comparing p...
详细信息
Bat algorithm is a novel swarm intelligent algorithm inspired by the echolocation behavior of bats with varying pulse rates of emission and loudness. In this paper, a new variant which is called adaptive bat algorithm...
详细信息
Hand gesture recognition (HGR) based on multimodal data has attracted considerable attention owing to its great potential in applications. Various manually designed multimodal deep networks have performed well in mult...
详细信息
暂无评论