An on-chip power-on reset circuit with a brown-out detection capability is implemented in a 0. 18 μm CMOS. A pF-order capacitor is charged with a proportional-to-absolute-temperature （PTAT） current from a bandgap r...An on-chip power-on reset circuit with a brown-out detection capability is implemented in a 0. 18 μm CMOS. A pF-order capacitor is charged with a proportional-to-absolute-temperature （PTAT） current from a bandgap reference with limited loop bandwidth and slow start-up feature, to generate a reset signal with high robustness and wide-range supply rise time. An embedded brown- out detector based on complementary voltage-to-current （V-to-I） conversion and current comparison can accurately respond to the brown-out event with high robustness over process and temperature when the supply is lower than 1.5 V and the brown-out duration is longer than 0. 1 ms. The presented design with embedded offset voltage cancellation consumes a quiescent current of 8. 5 μA from a 1. 8 V supply and works over ambient temperature of -40° to 120°.展开更多
面向车联网多用户通感一体系统,提出了一种感知辅助通信的大规模多输入多输出正交时频空鲁棒传输方案。由于实际雷达感知精度有限,基于感知参数重构的信道状态信息(CSI,channel state information)存在误差,系统的传输性能也会随之下降...面向车联网多用户通感一体系统,提出了一种感知辅助通信的大规模多输入多输出正交时频空鲁棒传输方案。由于实际雷达感知精度有限,基于感知参数重构的信道状态信息(CSI,channel state information)存在误差,系统的传输性能也会随之下降。对此,所提方案首先在发射端基于感知参数在时延多普勒域重构CSI,并考虑CSI误差设计鲁棒波束成形方案。其次,在接收端利用感知参数觉知用户间干扰及信道估计误差,并将所觉知的干扰误差以解析式的方式融入接收机中完成鲁棒设计。仿真结果表明,所提方案可以在CSI非理想情况下有效降低系统误码率,增加用户的数据接收速率,提升系统的整体性能。展开更多
目的针对第2代数字水印技术,提出一种基于Harris特征点和DWT-SVD的图像盲水印算法。方法提取归一化图像的Harris特征点;选取部分稳定特征点来确定要嵌入水印的特征区域;将特征区域作一次小波分解得到的低频子带,对低频子带进行分块,并...目的针对第2代数字水印技术,提出一种基于Harris特征点和DWT-SVD的图像盲水印算法。方法提取归一化图像的Harris特征点;选取部分稳定特征点来确定要嵌入水印的特征区域;将特征区域作一次小波分解得到的低频子带,对低频子带进行分块,并对每一块进行奇异值分解,通过对每块中最大奇异值进行加权的方法来嵌入水印信息。结果 PSNR值均大于45 d B,NC值接近于1,说明该算法具有可行性。结论该算法对剪切攻击具有很好的鲁棒性,同时该算法也能很好地抵抗噪声、中值滤波攻击、提高亮度攻击、降低亮度攻击、基本图像处理操作的攻击。展开更多
目标检测器现已被广泛应用在各类智能系统中,主要用于对图像中的物体进行识别与定位.然而,近年来的研究表明,目标检测器与DNNs分类器都易受数字对抗样本和物理对抗样本的影响.YOLOv3是实时检测任务中一种主流的目标检测器,现有攻击YOLOv...目标检测器现已被广泛应用在各类智能系统中,主要用于对图像中的物体进行识别与定位.然而,近年来的研究表明,目标检测器与DNNs分类器都易受数字对抗样本和物理对抗样本的影响.YOLOv3是实时检测任务中一种主流的目标检测器,现有攻击YOLOv3的物理对抗样本的构造方式大多是将生成的较大对抗性扰动打印出来再粘贴在特定类别的物体表面.最近的研究中出现的假阳性对抗样本(false positive adversarial example,FPAE)可通过目标模型直接生成得到,人无法识别出该对抗样本图像中的内容,但目标检测器却以高置信度将其误识别为攻击者指定的目标类.现有以YOLOv3为目标模型生成FPAE的方法仅有AA(appearing attack)方法一种,该方法在生成FPAE的过程中,为提升FPAE的鲁棒性,会在迭代优化过程中加入EOT(expectation over transformation)图像变换来模拟各种物理条件,但是并未考虑拍摄时可能出现的运动模糊(motion blur)情况,进而影响到对抗样本的攻击效果.此外,生成的FPAE在对除YOLOv3外的目标检测器进行黑盒攻击时的攻击成功率并不高.为生成性能更好的FPAE,以揭示现有目标检测器存在的弱点和测试现有目标检测器的安全性,以YOLOv3目标检测器为目标模型,提出RTFP(robust and transferable false positive)对抗攻击方法.该方法在迭代优化过程中,除了加入典型的图像变换外,还新加入了运动模糊变换.同时,在损失函数的设计上,借鉴了C&W攻击中损失函数的设计思想,并将目标模型在FPAE的中心所在的网格预测出的边界框与FPAE所在的真实边界框之间的重合度(intersection over union,IOU)作为预测的边界框的类别损失的权重项.在现实世界中的多角度、多距离拍摄测试以及实际道路上的驾车拍摄测试中,RTFP方法生成的FPAE能够保持较强的鲁棒性且迁移性强于现有方法生成的FPAE.展开更多
One key function of intelligent transportation systems is to automatically detect abnormal traffic phenomena and to help further investigations of the cause of the abnormality. This paper describes a robust principal ...One key function of intelligent transportation systems is to automatically detect abnormal traffic phenomena and to help further investigations of the cause of the abnormality. This paper describes a robust principal components analysis （RPCA）-based abnormal traffic flow pattern isolation and loop detector fault detection method. The results show that RPCA is a useful tool to distinguish regular traffic flow from abnormal traffic flow patterns caused by accidents and loop detector faults. This approach gives an effective traffic flow data pre-processing method to reduce the human effort in finding potential loop detector faults. The method can also be used to further investigate the causes of the abnormality.展开更多
基金Supported by the National Natural Science Foundation of China(6130603761201182)
文摘An on-chip power-on reset circuit with a brown-out detection capability is implemented in a 0. 18 μm CMOS. A pF-order capacitor is charged with a proportional-to-absolute-temperature （PTAT） current from a bandgap reference with limited loop bandwidth and slow start-up feature, to generate a reset signal with high robustness and wide-range supply rise time. An embedded brown- out detector based on complementary voltage-to-current （V-to-I） conversion and current comparison can accurately respond to the brown-out event with high robustness over process and temperature when the supply is lower than 1.5 V and the brown-out duration is longer than 0. 1 ms. The presented design with embedded offset voltage cancellation consumes a quiescent current of 8. 5 μA from a 1. 8 V supply and works over ambient temperature of -40° to 120°.
文摘面向车联网多用户通感一体系统,提出了一种感知辅助通信的大规模多输入多输出正交时频空鲁棒传输方案。由于实际雷达感知精度有限,基于感知参数重构的信道状态信息(CSI,channel state information)存在误差,系统的传输性能也会随之下降。对此,所提方案首先在发射端基于感知参数在时延多普勒域重构CSI,并考虑CSI误差设计鲁棒波束成形方案。其次,在接收端利用感知参数觉知用户间干扰及信道估计误差,并将所觉知的干扰误差以解析式的方式融入接收机中完成鲁棒设计。仿真结果表明,所提方案可以在CSI非理想情况下有效降低系统误码率,增加用户的数据接收速率,提升系统的整体性能。
文摘目的针对第2代数字水印技术,提出一种基于Harris特征点和DWT-SVD的图像盲水印算法。方法提取归一化图像的Harris特征点;选取部分稳定特征点来确定要嵌入水印的特征区域;将特征区域作一次小波分解得到的低频子带,对低频子带进行分块,并对每一块进行奇异值分解,通过对每块中最大奇异值进行加权的方法来嵌入水印信息。结果 PSNR值均大于45 d B,NC值接近于1,说明该算法具有可行性。结论该算法对剪切攻击具有很好的鲁棒性,同时该算法也能很好地抵抗噪声、中值滤波攻击、提高亮度攻击、降低亮度攻击、基本图像处理操作的攻击。
文摘目标检测器现已被广泛应用在各类智能系统中,主要用于对图像中的物体进行识别与定位.然而,近年来的研究表明,目标检测器与DNNs分类器都易受数字对抗样本和物理对抗样本的影响.YOLOv3是实时检测任务中一种主流的目标检测器,现有攻击YOLOv3的物理对抗样本的构造方式大多是将生成的较大对抗性扰动打印出来再粘贴在特定类别的物体表面.最近的研究中出现的假阳性对抗样本(false positive adversarial example,FPAE)可通过目标模型直接生成得到,人无法识别出该对抗样本图像中的内容,但目标检测器却以高置信度将其误识别为攻击者指定的目标类.现有以YOLOv3为目标模型生成FPAE的方法仅有AA(appearing attack)方法一种,该方法在生成FPAE的过程中,为提升FPAE的鲁棒性,会在迭代优化过程中加入EOT(expectation over transformation)图像变换来模拟各种物理条件,但是并未考虑拍摄时可能出现的运动模糊(motion blur)情况,进而影响到对抗样本的攻击效果.此外,生成的FPAE在对除YOLOv3外的目标检测器进行黑盒攻击时的攻击成功率并不高.为生成性能更好的FPAE,以揭示现有目标检测器存在的弱点和测试现有目标检测器的安全性,以YOLOv3目标检测器为目标模型,提出RTFP(robust and transferable false positive)对抗攻击方法.该方法在迭代优化过程中,除了加入典型的图像变换外,还新加入了运动模糊变换.同时,在损失函数的设计上,借鉴了C&W攻击中损失函数的设计思想,并将目标模型在FPAE的中心所在的网格预测出的边界框与FPAE所在的真实边界框之间的重合度(intersection over union,IOU)作为预测的边界框的类别损失的权重项.在现实世界中的多角度、多距离拍摄测试以及实际道路上的驾车拍摄测试中,RTFP方法生成的FPAE能够保持较强的鲁棒性且迁移性强于现有方法生成的FPAE.
基金Supported partly by the National Key Basic Research and Development (973) of China (No. 2006CB705506)the National High-Tech Research and Development (863) Program of China (Nos.2006AA11Z229 and 2007AA11Z222)the National Natural Science Foundation of China (Nos. 60374059 and 60534060)
文摘One key function of intelligent transportation systems is to automatically detect abnormal traffic phenomena and to help further investigations of the cause of the abnormality. This paper describes a robust principal components analysis （RPCA）-based abnormal traffic flow pattern isolation and loop detector fault detection method. The results show that RPCA is a useful tool to distinguish regular traffic flow from abnormal traffic flow patterns caused by accidents and loop detector faults. This approach gives an effective traffic flow data pre-processing method to reduce the human effort in finding potential loop detector faults. The method can also be used to further investigate the causes of the abnormality.
