In this work,we constructed a neural network proxy model(NNPM)to estimate the hydrodynamic resistance in the ship hull structure design process,which is based on the hydrodynamic load data obtained from both the poten...In this work,we constructed a neural network proxy model(NNPM)to estimate the hydrodynamic resistance in the ship hull structure design process,which is based on the hydrodynamic load data obtained from both the potential flow method(PFM)and the viscous flow method(VFM).Here the PFM dataset is applied for the tuning,pre-training,and the VFM dataset is applied for the fine-training.By adopting the PFM and VFM datasets simultaneously,we aim to construct an NNPM to achieve the high-accuracy prediction on hydrodynamic load on ship hull structures exerted from the viscous flow,while ensuring a moderate data-acquiring workload.The high accuracy prediction on hydrodynamic loads and the relatively low dataset establishment cost of the NNPM developed demonstrated the effectiveness and feasibility of hybrid dataset based NNPM achieving a high precision prediction of hydrodynamic loads on ship hull structures.The successful construction of the high precision hydrodynamic prediction NNPM advances the artificial intelligence-assisted design(AIAD)technology for various marine structures.展开更多
分心驾驶是交通事故发生的主要原因之一.针对目前分心驾驶检测手段单一、检测种类少、检测效率低的问题,提出一种基于轻量化网络与嵌入式的分心行为协同检测系统.首先,结合Ghost模块和通道注意力机制提出一种轻量化目标检测网络YOLO-Gho...分心驾驶是交通事故发生的主要原因之一.针对目前分心驾驶检测手段单一、检测种类少、检测效率低的问题,提出一种基于轻量化网络与嵌入式的分心行为协同检测系统.首先,结合Ghost模块和通道注意力机制提出一种轻量化目标检测网络YOLO-Ghost,采用CSPGBottleck构建GhostDarknet作为主干网络,同时构建一种具有多尺度注意力机制的多特征融合模块SE-FPN来进行特征融合,根据固有检测场景进行检测头优化,以CIOU(complete-IOU)作为损失函数.采用YOLO-Ghost识别和定位局部特征,提出APJ(anchor position judge)对手动分心行为进行判定;协同检测方面,利用MobileNetv3与YOLO-Ghost协同进行人脸关键点回归和视线估计;最后利用检测出的多模态信息对驾驶员当前行驶状态进行联合判定.实验结果表明,YOLO-Ghost的准确率和检测速度优于其他主流方法.将算法部署到嵌入式设备中,在NVIDIA Jetson TX1上实现了20FPS的实时检测性能,准确性和实时性均达到检测要求.展开更多
Due to complex mesoscopic and the distinct macroscopic evolution characteristics of ice,especially for its brittle-to-ductile transition in dynamic response,it is still a challenging task to build an accurate ice cons...Due to complex mesoscopic and the distinct macroscopic evolution characteristics of ice,especially for its brittle-to-ductile transition in dynamic response,it is still a challenging task to build an accurate ice constitutive model to predict ice loads during ship-ice collision.To address this,we incorporate the conventional multi-yield-surface plasticity model with the state-based peridynamics to simulate the stress and crack formation of ice under impact.Additionally,we take into account of the effects of inhomogeneous temperature distribution,strain rate,and pressure sensitivity.By doing so,we can successfully predict material failure of isotropic freshwater ice,iceberg ice,and columnar saline ice.Particularly,the proposed ice constitutive model is validated through several benchmark tests,and proved its applicability to model ice fragmentation under impacts,including drop tower tests and ballistic problems.Our results show that the proposed approach provides good computational performance to simulate ship-ice collision.展开更多
基金supported by a fellowship from China Scholar Council(No.201806680134).
文摘In this work,we constructed a neural network proxy model(NNPM)to estimate the hydrodynamic resistance in the ship hull structure design process,which is based on the hydrodynamic load data obtained from both the potential flow method(PFM)and the viscous flow method(VFM).Here the PFM dataset is applied for the tuning,pre-training,and the VFM dataset is applied for the fine-training.By adopting the PFM and VFM datasets simultaneously,we aim to construct an NNPM to achieve the high-accuracy prediction on hydrodynamic load on ship hull structures exerted from the viscous flow,while ensuring a moderate data-acquiring workload.The high accuracy prediction on hydrodynamic loads and the relatively low dataset establishment cost of the NNPM developed demonstrated the effectiveness and feasibility of hybrid dataset based NNPM achieving a high precision prediction of hydrodynamic loads on ship hull structures.The successful construction of the high precision hydrodynamic prediction NNPM advances the artificial intelligence-assisted design(AIAD)technology for various marine structures.
文摘分心驾驶是交通事故发生的主要原因之一.针对目前分心驾驶检测手段单一、检测种类少、检测效率低的问题,提出一种基于轻量化网络与嵌入式的分心行为协同检测系统.首先,结合Ghost模块和通道注意力机制提出一种轻量化目标检测网络YOLO-Ghost,采用CSPGBottleck构建GhostDarknet作为主干网络,同时构建一种具有多尺度注意力机制的多特征融合模块SE-FPN来进行特征融合,根据固有检测场景进行检测头优化,以CIOU(complete-IOU)作为损失函数.采用YOLO-Ghost识别和定位局部特征,提出APJ(anchor position judge)对手动分心行为进行判定;协同检测方面,利用MobileNetv3与YOLO-Ghost协同进行人脸关键点回归和视线估计;最后利用检测出的多模态信息对驾驶员当前行驶状态进行联合判定.实验结果表明,YOLO-Ghost的准确率和检测速度优于其他主流方法.将算法部署到嵌入式设备中,在NVIDIA Jetson TX1上实现了20FPS的实时检测性能,准确性和实时性均达到检测要求.
文摘密码子偏性是影响外源基因表达的重要因素,也是反映基因家族进化的关键线索之一。兰科植物是植物保护中的“旗舰”类群,且多数物种具有药用和观赏价值,但种子不易萌发,而ABI3(abscisic acid insensitive 3)基因是ABA信号转导的中心调控因子,是调控植物种子萌发的关键因子。因此,深入研究兰科ABI3基因密码子偏性,可为解析该基因功能、优化密码子组成奠定基础。本研究以拟南芥ABI3基因为探针,从公共数据库(NCBI)筛选出45条兰科植物ABI3基因,通过CodonW、SPSS、MEGA等生物信息学程序对其进行密码子使用模式和偏好性分析。相关参数分析结果表明,该基因的ENC(effective number of codon)值范围分布为40.84~58.46,其中大于52的占82.2%,CAI(codon adaptation index)平均值为0.203,远小于1,表明ABI3基因的密码子偏性相对较弱;中性绘图分析显示,GC12与GC3的回归曲线的斜率为0.6103,R2=0.7928,P<0.05,有极高的相关性,这暗示着碱基组成对密码子偏性具有重要的影响;ENC-plot分析中,所有基因ENC值都低于期望值,但相差不大,说明兰科ABI3基因密码子主要受碱基突变的影响,同时也受选择压力等其他因素影响。奇偶偏好性分析显示,大多数基因分布在平面图的右侧区域,密码子第3位A/T(U)的使用频率高于G/C。通过RSCU和ΔRSCU分析确定了26个最优密码子,其中以A/T(U)结尾的密码子有14个,以T(U)结尾的密码子有10个,说明兰科植物ABI3基因相对偏好使用A/U结尾的密码子,尤其偏好使用T(U)结尾的密码子,这与大多数单子叶植物密码子使用偏好不同。本研究可为兰科植物ABI3基因的系统发育、功能解析、提高ABI3基因表达效率提供理论基础,同时可为促进植物密码子生物学研究及外源基因改良提供参考。
文摘Due to complex mesoscopic and the distinct macroscopic evolution characteristics of ice,especially for its brittle-to-ductile transition in dynamic response,it is still a challenging task to build an accurate ice constitutive model to predict ice loads during ship-ice collision.To address this,we incorporate the conventional multi-yield-surface plasticity model with the state-based peridynamics to simulate the stress and crack formation of ice under impact.Additionally,we take into account of the effects of inhomogeneous temperature distribution,strain rate,and pressure sensitivity.By doing so,we can successfully predict material failure of isotropic freshwater ice,iceberg ice,and columnar saline ice.Particularly,the proposed ice constitutive model is validated through several benchmark tests,and proved its applicability to model ice fragmentation under impacts,including drop tower tests and ballistic problems.Our results show that the proposed approach provides good computational performance to simulate ship-ice collision.
