堆叠覆盖环境下的机械臂避障抓取是一个重要且有挑战性的任务。针对机械臂在堆叠环境下的避障抓取任务,本文提出了一种基于图像编码器和深度强化学习(deep reinforcement learning,DRL)的机械臂避障抓取方法Ec-DSAC(encoder and crop fo...堆叠覆盖环境下的机械臂避障抓取是一个重要且有挑战性的任务。针对机械臂在堆叠环境下的避障抓取任务,本文提出了一种基于图像编码器和深度强化学习(deep reinforcement learning,DRL)的机械臂避障抓取方法Ec-DSAC(encoder and crop for discrete SAC)。首先设计结合YOLO(you only look once)v5和对比学习网络编码的图像编码器,能够编码关键特征和全局特征,实现像素信息至向量信息的降维。其次结合图像编码器和离散软演员-评价家(soft actor-critic,SAC)算法,设计离散动作空间和密集奖励函数约束并引导策略输出的学习方向,同时使用随机图像裁剪增加强化学习的样本效率。最后,提出了一种应用于深度强化学习预训练的二次行为克隆方法,增强了强化学习网络的学习能力并提高了控制策略的成功率。仿真实验中Ec-DSAC的避障抓取成功率稳定高于80.0%,验证其具有比现有方法更好的避障抓取性能。现实实验中避障抓取成功率为73.3%,验证其在现实堆叠覆盖环境下避障抓取的有效性。展开更多
The aim of this study is to investigate microbial structures and diversities in five active hydrothermal fields＇ sediments along the Eastern Lau Spreading Centre （ELSC） in the Lau Basin （southwest Pacific）. Micro...The aim of this study is to investigate microbial structures and diversities in five active hydrothermal fields＇ sediments along the Eastern Lau Spreading Centre （ELSC） in the Lau Basin （southwest Pacific）. Microbial communities were surveyed by denatured gradient gel electrophoresis （DGGE） and clone library analysis of 16S rRNA genes. The differences in microbial community structures among sediment samples from the five deep-sea hydrothermal sites were revealed by DGGE profiles. Cluster analysis of DGGE profiles sepa- rated the five hydrothermal samples into two groups. Four different 16S rRNA gene clone libraries, repre- senting two selected hydrothermal samples （19-4TVG8 and 19-4TVG11）, were constructed. Twenty-three and 32 phylotypes were identified from 166 and 160 bacterial clones respectively, including Proteobacteria, Bacteroidetes, Firmicutes, Nitrospirae and Planctomycetes. The phylum Proteobacteria is dominant in both bacterial libraries with a predominance of Gamma-Proteobacteria. A total of 31 and 25 phylotypes were obtained from 160 and 130 archaeal clones respectively, including Miscellaneous Crenarchaeotic Group, Marine Group Ⅰ and Ⅲ, Marine Benthic Group E, Terrestrial Hot Spring Crenarchaeota and Deep-sea Hy- drothermal Vent Euryarchaeota. These results show a variety of clones related to those involved in sulfur cycling, suggesting that the cycling and utilization of sulfur compounds may extensively occur in the Lau Basin deep-sea hydrothermal ecosystem.展开更多
Pseudoalteromonas sp. SM9913 is a phychrotmphic bacterium isolated from the deep-sea sediment. The genes encoding chaperones DnaJ and DnaK of P. sp. SM9913 were cloned by normal PCR and TAIL - PCR （GenBank accession ...Pseudoalteromonas sp. SM9913 is a phychrotmphic bacterium isolated from the deep-sea sediment. The genes encoding chaperones DnaJ and DnaK of P. sp. SM9913 were cloned by normal PCR and TAIL - PCR （GenBank accession Nos DQ640312, DQ504163 ）. The chaperones DnaJ and DnaK from the strain SM9913 contain such conserved domains as those of many other bacteria, and show some cold-adapted characteristics in their structures when compared with those from psychro-, meso-and themophilic bacteria. It is indicated that chaperones DnaJ and DnaK of P. sp. SM9913 may be adapted to low temperature in deep-sea and function well in assisting folding, assembling and translocation of proteins at low temperature. This research lays a foundation for the further study on the cold-adapted mechanism of chaperones DnaJ and DnaK of cold-adapted microorganisms.展开更多
The extraction of rolling bearing fault features using traditional diagnostic methods is not sufficiently comprehensive and the features are often chosen subjectively and depend on human experience. In this paper, an ...The extraction of rolling bearing fault features using traditional diagnostic methods is not sufficiently comprehensive and the features are often chosen subjectively and depend on human experience. In this paper, an improved deep convolutional process is used to extract a set of features adaptively. The hidden multi-layer feature of deep convolutional neural networks is also exploited to improve the extraction features. A deterministic detection of low-confidence samples is performed to ensure the reliability of the recognition results and to decrease the rate of false positives by evaluating the diagnosis of the deep convolutional neural network. To improve the efficiency of the continuous learning elements of the rolling bearing fault diagnosis, a clone learning strategy based on cloning and mutation operations is proposed. The experimental results show that the proposed deep convolutional neural network model can extract multiple rolling bearing fault features, improve classification and detection accuracy by reducing the false positive rate when diagnosing rolling bearing faults, and accelerate learning efficiency when using low-confidence rolling bearing fault samples.展开更多
文摘堆叠覆盖环境下的机械臂避障抓取是一个重要且有挑战性的任务。针对机械臂在堆叠环境下的避障抓取任务,本文提出了一种基于图像编码器和深度强化学习(deep reinforcement learning,DRL)的机械臂避障抓取方法Ec-DSAC(encoder and crop for discrete SAC)。首先设计结合YOLO(you only look once)v5和对比学习网络编码的图像编码器,能够编码关键特征和全局特征,实现像素信息至向量信息的降维。其次结合图像编码器和离散软演员-评价家(soft actor-critic,SAC)算法,设计离散动作空间和密集奖励函数约束并引导策略输出的学习方向,同时使用随机图像裁剪增加强化学习的样本效率。最后,提出了一种应用于深度强化学习预训练的二次行为克隆方法,增强了强化学习网络的学习能力并提高了控制策略的成功率。仿真实验中Ec-DSAC的避障抓取成功率稳定高于80.0%,验证其具有比现有方法更好的避障抓取性能。现实实验中避障抓取成功率为73.3%,验证其在现实堆叠覆盖环境下避障抓取的有效性。
基金Foundation item: The China Ocean Mineral Resources Research and Development Association under contract No. DYXM-115-02-2-07the State Oceanic Administration of People’s Republic of China under contract No. 200805032the National Natural Science Foundation of China under contract Nos 50621063 and 40646029
文摘The aim of this study is to investigate microbial structures and diversities in five active hydrothermal fields＇ sediments along the Eastern Lau Spreading Centre （ELSC） in the Lau Basin （southwest Pacific）. Microbial communities were surveyed by denatured gradient gel electrophoresis （DGGE） and clone library analysis of 16S rRNA genes. The differences in microbial community structures among sediment samples from the five deep-sea hydrothermal sites were revealed by DGGE profiles. Cluster analysis of DGGE profiles sepa- rated the five hydrothermal samples into two groups. Four different 16S rRNA gene clone libraries, repre- senting two selected hydrothermal samples （19-4TVG8 and 19-4TVG11）, were constructed. Twenty-three and 32 phylotypes were identified from 166 and 160 bacterial clones respectively, including Proteobacteria, Bacteroidetes, Firmicutes, Nitrospirae and Planctomycetes. The phylum Proteobacteria is dominant in both bacterial libraries with a predominance of Gamma-Proteobacteria. A total of 31 and 25 phylotypes were obtained from 160 and 130 archaeal clones respectively, including Miscellaneous Crenarchaeotic Group, Marine Group Ⅰ and Ⅲ, Marine Benthic Group E, Terrestrial Hot Spring Crenarchaeota and Deep-sea Hy- drothermal Vent Euryarchaeota. These results show a variety of clones related to those involved in sulfur cycling, suggesting that the cycling and utilization of sulfur compounds may extensively occur in the Lau Basin deep-sea hydrothermal ecosystem.
基金The work was supported by the Hi-Tech Research and Development Program of China under contract Nos 2006AA09Z414 and 2007AA091903；the China Ocean Mineral Resources R ＆ D Association under contract No. DYXM - 115 - 02 - 2 - 6；the National Natural Science Foundation of China under contract No. Z2004D02；the Natural Science Foundation of Shandong Province of China under contract No. Z2004D02；the Foundation for Young Excellent Scientists in Shandong Province of China under contract No. 2006BS02002；the Program for New Century Excellent Talents in University under contract No. NCET - 06 - 0578.
文摘Pseudoalteromonas sp. SM9913 is a phychrotmphic bacterium isolated from the deep-sea sediment. The genes encoding chaperones DnaJ and DnaK of P. sp. SM9913 were cloned by normal PCR and TAIL - PCR （GenBank accession Nos DQ640312, DQ504163 ）. The chaperones DnaJ and DnaK from the strain SM9913 contain such conserved domains as those of many other bacteria, and show some cold-adapted characteristics in their structures when compared with those from psychro-, meso-and themophilic bacteria. It is indicated that chaperones DnaJ and DnaK of P. sp. SM9913 may be adapted to low temperature in deep-sea and function well in assisting folding, assembling and translocation of proteins at low temperature. This research lays a foundation for the further study on the cold-adapted mechanism of chaperones DnaJ and DnaK of cold-adapted microorganisms.
基金supported by the National Natural Science Foundation of China (No. 61472271)
文摘The extraction of rolling bearing fault features using traditional diagnostic methods is not sufficiently comprehensive and the features are often chosen subjectively and depend on human experience. In this paper, an improved deep convolutional process is used to extract a set of features adaptively. The hidden multi-layer feature of deep convolutional neural networks is also exploited to improve the extraction features. A deterministic detection of low-confidence samples is performed to ensure the reliability of the recognition results and to decrease the rate of false positives by evaluating the diagnosis of the deep convolutional neural network. To improve the efficiency of the continuous learning elements of the rolling bearing fault diagnosis, a clone learning strategy based on cloning and mutation operations is proposed. The experimental results show that the proposed deep convolutional neural network model can extract multiple rolling bearing fault features, improve classification and detection accuracy by reducing the false positive rate when diagnosing rolling bearing faults, and accelerate learning efficiency when using low-confidence rolling bearing fault samples.
