公开(公告)号：US06649423B2

公开(公告)日：2003-11-18

申请号：US09971347

申请日：2001-10-04

Applicant: Thomas Anthony ,  Lung Tran ,  Manish Sharma

Inventor： Thomas Anthony ,  Lung Tran ,  Manish Sharma

IPC: H01L2100

CPC classification number: G11C11/16 ,  Y10S257/902

Abstract: A magnetic tunnel junction is fabricated by forming pinned and sense layers; and re-setting a magnetization vector of at least one of the layers.

Abstract translation: 通过形成固定和感应层制造磁性隧道结; 并重新设定至少一层的磁化矢量。

公开(公告)号：US06577529B1

公开(公告)日：2003-06-10

申请号：US10235011

申请日：2002-09-03

Applicant: Manish Sharma ,  Thomas C. Anthony ,  Lung T. Tran

Inventor： Manish Sharma ,  Thomas C. Anthony ,  Lung T. Tran

IPC: G11C1100

CPC classification number: G11C11/15 ,  G11C11/5607 ,  G11C2213/71

Abstract: A memory cell includes a conductor clad with ferromagnetic material; first and second spacer layers on opposite sides of the clad conductor; a first data layer on the first spacer layer; and a second data layer on the second spacer layer.

Abstract translation: 存储单元包括用铁磁材料包覆的导体; 在包层导体的相对侧上的第一和第二间隔层; 第一间隔层上的第一数据层; 以及第二间隔层上的第二数据层。

公开(公告)号：US06538917B1

公开(公告)日：2003-03-25

申请号：US09963851

申请日：2001-09-25

Applicant: Lung T. Tran ,  Manish Sharma

Inventor： Lung T. Tran ,  Manish Sharma

IPC: G11C1115

CPC classification number: G11C11/16

Abstract: A magneto-resistive device includes data and reference layers having different coercivities. Each layer has a magnetization that can be oriented in either of two directions. The memory device may be read by temporarily setting the magnetization of the reference layer to a known orientation, and determining a resistance state of the device.

Abstract translation: 磁阻装置包括具有不同矫顽力的数据和参考层。 每个层具有可以在两个方向中的任一方向上定向的磁化。 可以通过将参考层的磁化临时设置为已知取向来读取存储器件，并且确定器件的电阻状态。

公开(公告)号：US06434060B1

公开(公告)日：2002-08-13

申请号：US09917882

申请日：2001-07-31

Applicant: Lung T. Tran ,  Manish Sharma

Inventor： Lung T. Tran ,  Manish Sharma

IPC: G11C700

CPC classification number: G11C16/3486 ,  G11C17/16 ,  G11C17/18

Abstract: A method and circuit write a memory cell. The method applies a pulse to a write line connected to the memory cell. The duration of the pulse is not predetermined. The method compares a value on the input side of the cell to a reference value. The method discontinues the pulse on the write line, in response to the comparing step, preferably if the value on the write line exceeds the reference value. The circuit comprises a pulse generator and a comparator. The pulse generator has an output and an enable input. The output is connected to a write line connected to the memory cell. The output, when enabled, carries a pulse. The comparator has two inputs and an output. One of the inputs is connected to the write line. The other of the inputs is connected to a reference. The output is connected to the write line, whereby the pulse is disabled or enabled on the write line depending upon comparator output. A complete memory system comprises an array of memory cells, a write line, and a pulse generator and comparator as described above.

Abstract translation: 一种方法和电路写入一个存储单元。 该方法向连接到存储单元的写入线施加脉冲。 脉冲的持续时间不是预定的。 该方法将单元格输入端的值与参考值进行比较。 该方法响应于比较步骤中止写入线上的脉冲，优选地，如果写入线上的值超过参考值。 电路包括脉冲发生器和比较器。 脉冲发生器具有输出和使能输入。 输出连接到连接到存储单元的写入线。 输出，当使能时，会携带一个脉冲。 比较器有两个输入和一个输出。 其中一个输入连接到写入线。 另一个输入连接到引用。 输出连接到写入线，根据比较器输出，脉冲在写入线上被禁止或使能。 完整的存储器系统包括如上所述的存储器单元阵列，写入线以及脉冲发生器和比较器。

公开(公告)号：US06434048B1

公开(公告)日：2002-08-13

申请号：US09908901

申请日：2001-07-20

Applicant: Lung T. Tran ,  Manish Sharma

Inventor： Lung T. Tran ,  Manish Sharma

IPC: G11C700

CPC classification number: G11C7/1084 ,  G11C7/1006 ,  G11C7/1078

Abstract: A method and circuit write a memory cell. The method applies a pulse train to a write line connected to the memory cell. The number of pulses in the pulse train is not predetermined. The method compares a value on the input side of the cell to a reference value, wherein the input side of the memory cell provides an indication that a writing operation is complete. The method discontinues the pulse train on the write line, in response to the comparing step, preferably if the value on the write line exceeds the reference value. Preferably, the pulses are short in width and large in magnitude. The method may optionally count the number of pulses in the pulse train, and discontinue the pulse train on the write line and/or declare the cell as unusable if the number of pulses exceeds a predetermined maximum. The circuit comprises a pulse train generator and a comparator. The pulse train generator has an output and an enable input. The output is connected to a write line connected to the memory cell. The output, when enabled, carries a pulse train. The comparator has two inputs and an output. One of the inputs is connected to the write line. The other of the inputs is connected to a reference. The output is connected to the enable input of the pulse train generator, whereby the pulse train generator is disabled or enabled depending upon the comparator output. Optionally, the circuit further comprises a counter that counts pulses and disables the pulse train generator after a predetermined maximum number of pulses. A complete memory system comprises an array of memory cells, a write line, and a pulse train generator and comparator as described above.

Abstract translation: 一种方法和电路写入一个存储单元。 该方法将脉冲串应用于连接到存储单元的写入线。 脉冲串中的脉冲数不是预定的。 该方法将小区的输入侧的值与参考值进行比较，其中存储单元的输入侧提供写入操作完成的指示。 该方法响应于比较步骤中止写入线上的脉冲序列，优选地，如果写入线上的值超过参考值。 优选地，脉冲的宽度较短，而且幅度较大。 该方法可以可选地计数脉冲串中的脉冲数，并且如果脉冲数超过预定最大值，则在写入线上中止脉冲序列和/或将单元声明为不可用。 电路包括脉冲串发生器和比较器。 脉冲串发生器具有输出和使能输入。 输出连接到连接到存储单元的写入线。 输出，当使能时，携带脉冲串。 比较器有两个输入和一个输出。 其中一个输入连接到写入线。 另一个输入连接到引用。 输出连接到脉冲串发生器的使能输入端，根据比较器输出，脉冲串发生器被禁止或使能。 可选地，电路还包括计数器，其计数脉冲并且在预定的最大脉冲数之后禁用脉冲序列发生器。 完整的存储器系统包括如上所述的存储器单元阵列，写入线以及脉冲序列发生器和比较器。

公开(公告)号：US5862519A

公开(公告)日：1999-01-19

申请号：US827562

申请日：1997-04-01

Applicant: Manish Sharma ,  Richard J. Mammone

Inventor： Manish Sharma ,  Richard J. Mammone

IPC: G10L11/00 ,  G10L15/04 ,  G10L5/06 ,  G10L9/00

CPC classification number: G10L25/48 ,  G10L15/04

Abstract: The present invention relates to a method for segmenting speech into subword speech segments. Optimal boundary locations for each estimate of a number of segments are determined within an estimated range of the number of segments. In addition, an optimality criteria is found for each estimate of the number of segments within the range. Using the optimality criteria, the optimal number of subwords are determined. From the location of the boundaries and the optimal number of segments, data can be clustered or speech can be segmented. The method can be used in data processing systems, speaker verification, medium size vocabulary speech recognition systems, language identification systems and coarse subword level speech segmentation processes.

Abstract translation: 本发明涉及一种将语音分割为子词语音片段的方法。 在段的数量的估计范围内确定多个段的每个估计的最优边界位置。 此外，针对范围内的段数的每个估计找到最优性标准。 使用最优性标准，确定最佳子词数。 从边界的位置和段的最佳数量，可以对数据进行聚类或分割语音。 该方法可用于数据处理系统，扬声器验证，中等大小词汇语音识别系统，语言识别系统和粗略词级语音分割过程。

公开(公告)号：US5839103A

公开(公告)日：1998-11-17

申请号：US479012

申请日：1995-06-07

Applicant: Richard J. Mammone ,  Kevin Farrell ,  Manish Sharma ,  Devang Naik ,  Xiaoyu Zhang ,  Khaled Assaleh ,  Han-Seng Liou

Inventor： Richard J. Mammone ,  Kevin Farrell ,  Manish Sharma ,  Devang Naik ,  Xiaoyu Zhang ,  Khaled Assaleh ,  Han-Seng Liou

IPC: G10L15/10 ,  G10L15/02 ,  G10L15/08 ,  G10L15/20 ,  G10L17/00 ,  G10L21/02 ,  G10L5/00

CPC classification number: G10L17/10 ,  G10L17/20 ,  G10L17/04 ,  G10L17/14 ,  G10L17/18 ,  G10L25/03

Abstract: The present invention relates to a pattern recognition system which uses data fusion to combine data from a plurality of extracted features and a plurality of classifiers. Speaker patterns can be accurately verified with the combination of discriminant based and distortion based classifiers. A novel approach using a training set of a "leave one out" data can be used for training the system with a reduced data set. Extracted features can be improved with a pole filtered method for reducing channel effects and an affine transformation for improving the correlation between training and testing data.

Abstract translation: 本发明涉及使用数据融合来组合来自多个提取的特征的数据和多个分类器的模式识别系统。 可以通过基于判别式和基于失真的分类器的组合来准确地验证扬声器模式。 使用训练集“离开一个”数据的新颖方法可用于使用减少的数据集训练系统。 提取的特征可以通过用于减少信道效应的极点滤波方法和用于改善训练和测试数据之间的相关性的仿射变换来改进。

公开(公告)号：US09222978B2

公开(公告)日：2015-12-29

申请号：US13416912

申请日：2012-03-09

Applicant: Yu Huang ,  Wu-Tung Cheng ,  Ruifeng Guo ,  Manish Sharma ,  Liyang Lai

Inventor： Yu Huang ,  Wu-Tung Cheng ,  Ruifeng Guo ,  Manish Sharma ,  Liyang Lai

IPC: G01R31/28 ,  G01R31/3185

CPC classification number: G01R31/318536 ,  G01R31/318544

Abstract: Aspects of the invention relate to techniques of using two-dimensional scan architecture for testing and diagnosis. A two-dimensional scan cell network may be constructed by coupling input for each scan cell to outputs for two or more other scan cells and/or primary inputs through a multiplexer. To test and diagnose the two-dimensional scan cell network, the two-dimensional scan cell network may be loaded with chain patterns and unloaded with corresponding chain test data along two or more sets of scan paths. Based on the chain test data, one or more defective scan cells or defective scan cell candidates may be determined.

Abstract translation: 本发明的方面涉及使用二维扫描结构进行测试和诊断的技术。 可以通过将每个扫描单元的输入耦合到通过多路复用器的两个或更多个其它扫描单元和/或主要输入的输出来构造二维扫描单元网络。 为了测试和诊断二维扫描单元网络，二维扫描单元网络可以加载链模式并且沿着两组或更多组扫描路径卸载相应的链测试数据。 基于链测试数据，可以确定一个或多个有缺陷的扫描单元或有缺陷的扫描单元候选。

公开(公告)号：US08566032B2

公开(公告)日：2013-10-22

申请号：US12609582

申请日：2009-10-30

Applicant: Mahesh Chowdhary ,  Arun Kumar ,  Manish Sharma ,  Ashutosh Pande ,  Mahaveer Jain ,  Gagan Narula ,  Chinmay Agarwal ,  Kolin Paul

Inventor： Mahesh Chowdhary ,  Arun Kumar ,  Manish Sharma ,  Ashutosh Pande ,  Mahaveer Jain ,  Gagan Narula ,  Chinmay Agarwal ,  Kolin Paul

IPC: G01C21/00

CPC classification number: G01C21/005 ,  G01C21/20 ,  G01S19/49

Abstract: Methods and apparatuses for estimating a user's altitude with respect to the mean sea level are provided. According to some aspects, the present invention is able to estimate altitude in both open sky as well as in degraded GPS signal environments such as dense urban canyon environments where GPS performance is affected by fewer available satellites and/or multipath error. According to other aspects, the present invention uses data from a pressure sensor to estimate altitude, either with or without the use of GPS aiding data. According to further aspects, estimated altitude is integrated with other types of dead reckoning data to provide user context detection pertaining to changes of altitude.

Abstract translation: 提供了用于估计用户相对于平均海平面的高度的方法和装置。 根据一些方面，本发明能够估计开阔的天空以及劣化的GPS信号环境，例如密集的城市峡谷环境中的高度，其中GPS性能受到较少的可用卫星和/或多径误差的影响。 根据其他方面，本发明使用来自压力传感器的数据来估计高度，无论是否使用GPS辅助数据。 根据其他方面，估计高度与其他类型的航位推算数据集成，以提供与高度变化有关的用户上下文检测。

公开(公告)号：US20130199497A1

公开(公告)日：2013-08-08

申请号：US13368230

申请日：2012-02-07

Applicant: Manish Sharma ,  Gopichandra Surnilla

Inventor： Manish Sharma ,  Gopichandra Surnilla

IPC: F02B77/00

CPC classification number: F01N3/32 ,  F01L1/0532 ,  F01L2001/0537 ,  F01N3/22 ,  F01N3/34 ,  Y02T10/20

Abstract: Embodiments for injecting air into exhaust ports of an engine are provided. In one example, a multi-cylinder engine method comprises combusting a rich air-fuel mixture, injecting air into a first cylinder's exhaust port and not into a second cylinder's exhaust port during the first cylinder's exhaust stroke, and injecting air into the second cylinder's exhaust port and not into the first cylinder's exhaust port during the second cylinder's exhaust stroke. In this way, injection into each exhaust port may be timed with exhaust release into each exhaust port.

Abstract translation: 提供了将空气注入发动机的排气口的实施例。 在一个示例中，多气缸发动机方法包括在第一气缸的排气冲程期间燃烧富空燃混合物，将空气喷入第一气缸的排气口而不进入第二气缸的排气口，并将空气喷入第二气缸排气 在第二缸排气冲程期间不进入第一气缸的排气口。 以这种方式，每个排气口的喷射可以被排放到每个排气口中。