为了实现高光谱图像的快速训练、分类和超参数自适应寻优,提出基于深层残差3D卷积神经网络(3D-CNN)的高光谱图像识别分类算法.由于采用的3D特征提取算法更适合高光谱3D数据结构,使得网络可以快速地从完整的高光谱图像样本中同时提取丰...为了实现高光谱图像的快速训练、分类和超参数自适应寻优,提出基于深层残差3D卷积神经网络(3D-CNN)的高光谱图像识别分类算法.由于采用的3D特征提取算法更适合高光谱3D数据结构,使得网络可以快速地从完整的高光谱图像样本中同时提取丰富的空间和光谱特征;此外,通过对高光谱图像样本平面空间方向的旋转和翻转操作进行数据增强的方法;以及运用TPE超参数优化算法对设定的超参数选择空间自适应寻优的方法,都可以有效地提高分类准确率.通过在TensorFlow框架下对Pavia University, Indian Pines和KSC等标准高光谱数据集上的实验结果表明,与SSRN等其他算法相比,文中算法在加深网络结构的同时,提高了分类准确率;与人工设定超参数网络相比,以TPE自适应超参数优化算法优化的网络参数数量减少约一半,训练时间缩短约10%.展开更多
The joint optimization of detection threshold and waveform parameters for target tracking which comes from the idea of cognitive radar is investigated for the modified probabilistic data association(MPDA)filter.The tr...The joint optimization of detection threshold and waveform parameters for target tracking which comes from the idea of cognitive radar is investigated for the modified probabilistic data association(MPDA)filter.The transmitted waveforms and detection threshold are adaptively selected to enhance the tracking performance.The modified Riccati equation is adopted to predict the error covariance which is used as the criterion function,while the optimization problem is solved through the genetic algorithm(GA).The detection probability,false alarm probability and measurement noise covariance are all considered together,which significantly improves the tracking performance of the joint detection and tracking system.Simulation results show that the proposed adaptive waveform-detection threshold joint optimization method outperforms the adaptive threshold method and the fixed parameters method,which will reduce the tracking error.The average reduction of range error between the adaptive joint method and the fixed parameters method is about 0.6 m,while that between the adaptive joint method and the adaptive threshold only method is about 0.3 m.Similar error reduction occurs for the velocity error and acceleration error.展开更多
A comparative approach was performed between the response surface method(RSM) and the adaptive neuro-fuzzy inference system(ANFIS) to enhance the tensile properties, including the ultimate tensile strength and the ten...A comparative approach was performed between the response surface method(RSM) and the adaptive neuro-fuzzy inference system(ANFIS) to enhance the tensile properties, including the ultimate tensile strength and the tensile elongation, of friction stir welded age hardenable AA6061 and AA2024 aluminum alloys. The effects of the welding parameters, namely the tool rotational speed, welding speed, axial load and pin profile, on the ultimate tensile strength and the tensile elongation were analyzed using a three-level, four-factor Box-Behnken experimental design. The developed design was utilized to train the ANFIS models. The predictive capabilities of RSM and ANFIS were compared based on the root mean square error, the mean absolute error, and the correlation coefficient based on the obtained data set. The results demonstrate that the developed ANFIS models are more effective than the RSM model.展开更多
电网停电计划的排期结果关系到电网安全稳定运行和检修工作的开展,是电网运行方式业务的重要组成。目前,已有计划排期方法缺乏对计划间存在冲突这一场景的考虑,且算法效率较低,难以满足停电计划排期的实际需求。为此,该文以工作量不均...电网停电计划的排期结果关系到电网安全稳定运行和检修工作的开展,是电网运行方式业务的重要组成。目前,已有计划排期方法缺乏对计划间存在冲突这一场景的考虑,且算法效率较低,难以满足停电计划排期的实际需求。为此,该文以工作量不均衡度、停电计划时间调整量、停电经济成本为目标,涵盖计划关联关系判别和优先级排序等过程,建立了考虑冲突的电网停电计划优化求解模型。在此基础上,通过对NSGA II算法(the second generation of non-dominated sorting genetic algorithm,NSGAII)进行性能改进,提出了基于约束的自适应NSGAII算法(constraint-basedadaptive NSGAII,CA-NSGAII),并将其用于模型求解。最后,在IEEE-300输电系统模型中模拟了月停电计划排期过程,验证了该文所提模型与实际情况更为贴近,所提求解算法更加准确高效。展开更多
文摘为了实现高光谱图像的快速训练、分类和超参数自适应寻优,提出基于深层残差3D卷积神经网络(3D-CNN)的高光谱图像识别分类算法.由于采用的3D特征提取算法更适合高光谱3D数据结构,使得网络可以快速地从完整的高光谱图像样本中同时提取丰富的空间和光谱特征;此外,通过对高光谱图像样本平面空间方向的旋转和翻转操作进行数据增强的方法;以及运用TPE超参数优化算法对设定的超参数选择空间自适应寻优的方法,都可以有效地提高分类准确率.通过在TensorFlow框架下对Pavia University, Indian Pines和KSC等标准高光谱数据集上的实验结果表明,与SSRN等其他算法相比,文中算法在加深网络结构的同时,提高了分类准确率;与人工设定超参数网络相比,以TPE自适应超参数优化算法优化的网络参数数量减少约一半,训练时间缩短约10%.
基金Project(61171133) supported by the National Natural Science Foundation of ChinaProject(11JJ1010) supported by the Natural Science Fund for Distinguished Young Scholars of Hunan Province,China
文摘The joint optimization of detection threshold and waveform parameters for target tracking which comes from the idea of cognitive radar is investigated for the modified probabilistic data association(MPDA)filter.The transmitted waveforms and detection threshold are adaptively selected to enhance the tracking performance.The modified Riccati equation is adopted to predict the error covariance which is used as the criterion function,while the optimization problem is solved through the genetic algorithm(GA).The detection probability,false alarm probability and measurement noise covariance are all considered together,which significantly improves the tracking performance of the joint detection and tracking system.Simulation results show that the proposed adaptive waveform-detection threshold joint optimization method outperforms the adaptive threshold method and the fixed parameters method,which will reduce the tracking error.The average reduction of range error between the adaptive joint method and the fixed parameters method is about 0.6 m,while that between the adaptive joint method and the adaptive threshold only method is about 0.3 m.Similar error reduction occurs for the velocity error and acceleration error.
基金Sri Chandrasekharendra Saraswathi Viswa Maha Vidyalaya, Enathur, Kanchipuram, Tamilnadu for funding this research as a university minor research project
文摘A comparative approach was performed between the response surface method(RSM) and the adaptive neuro-fuzzy inference system(ANFIS) to enhance the tensile properties, including the ultimate tensile strength and the tensile elongation, of friction stir welded age hardenable AA6061 and AA2024 aluminum alloys. The effects of the welding parameters, namely the tool rotational speed, welding speed, axial load and pin profile, on the ultimate tensile strength and the tensile elongation were analyzed using a three-level, four-factor Box-Behnken experimental design. The developed design was utilized to train the ANFIS models. The predictive capabilities of RSM and ANFIS were compared based on the root mean square error, the mean absolute error, and the correlation coefficient based on the obtained data set. The results demonstrate that the developed ANFIS models are more effective than the RSM model.
文摘电网停电计划的排期结果关系到电网安全稳定运行和检修工作的开展,是电网运行方式业务的重要组成。目前,已有计划排期方法缺乏对计划间存在冲突这一场景的考虑,且算法效率较低,难以满足停电计划排期的实际需求。为此,该文以工作量不均衡度、停电计划时间调整量、停电经济成本为目标,涵盖计划关联关系判别和优先级排序等过程,建立了考虑冲突的电网停电计划优化求解模型。在此基础上,通过对NSGA II算法(the second generation of non-dominated sorting genetic algorithm,NSGAII)进行性能改进,提出了基于约束的自适应NSGAII算法(constraint-basedadaptive NSGAII,CA-NSGAII),并将其用于模型求解。最后,在IEEE-300输电系统模型中模拟了月停电计划排期过程,验证了该文所提模型与实际情况更为贴近,所提求解算法更加准确高效。
