Inside_NAND_Flash_Memories.pdf
Inside_NAND_Flash_Memories.pdfRino micheloni· Luca Crippa· Alessia marelliInside nand flashMemoriesSpringerRino micheloniLuca CrippaIntegrated Device TechnologyForward InsightsAgrate BrianzaNorth yorkItalCanadarino. micheloni@ieee. orgluca.crippa @ieee.orgAlessia marelliIntegrated Device Technologyagrate brianzaItalylessiamarelli@gmail.comISBN97890-481-9430-8e-ISBN97890-481-9431-5DOI10.1007/978-90-481-9431-5Springer Dordrecht Heidelberg London New YorkLibrary of Congress Control Number: 2010931597O Springer Science+Business Media B.V. 2010No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or byany means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without writtenbeing entered and executed on a computer system, for exclusive use by the purchaser of the wor Ose ofpermission from the Publisher, with the exception of any material supplied specifically for the purpoCover design: eStudio Calamar SLPrinted on acid-free paperSpringerispartofSpringerScience+businessMedia(www.springer.com)PrefaceIn the last decade Flash cards became the most important digital storage supportIt is difficult to identify a single killer application(digital photography, MP3digital video, ...)or whether the success of this media support is due to itsusability in different applications. It is also difficult to state whether the recentlution in the infotainment area has been triggered by the availability of high-performance NAND Flash memories or if the extraordinary success of Flash cardsis a consequence of the establishment of new applicationsIn any case, independently of the cause-effect relationship linking new digitalapplications and Flash cards, to realize how NANd Flash memories entered in ourdaily life, it is sufficient to imagine as they would change our recent habits if theNAND memories disappeared suddenly. To take a picture it would be necessary tofind a film (as well as a traditional camera .), disks or even magnetic tapes wouldbe used to record a video or to listen a song, and a cellular phone would return tobe a simple mean of communication rather than a console for an extendedentertainmentThe development of nand Flash memories will not be set down on the mereevolution of these digital systems since a new killer application can trigger a furthersuccess for this memory support: the replacement of Hard Disk Drives(HDD)with Solid State Drives (SSD)The advantages of Ssd with respect to HDD are countless(higher read/writspeed, higher mechanical reliability, random access, silent operation, lower powerconsumption, lower weight, .. ) Nevertheless, the cost per gigabyte is stillsignificantly favorable to hDd and the success of ssd will depend only on theperformances that they will succeed in reachingThe present book, the fourth authored by rino micheloni and coworkers afterVLSI-Design of Non-volatiles Memories, Memories in Wireless Systems and ErrorCorrection Codes for Non-volatile Memories, all published by Springer, tacklesall the main aspects related to NANd Flash memories, from the physicaltechnological, and circuit issues to their use in Flash cards and SsdsAfter an extended market overview(Chap. 1), Chap 2 has been conceived as aguide for the entire book, so that the reader can directly reach the heart of theproblems that interest himThe following chapters deepen physical and technological aspects. In particularChaps. 3 and 4 tackle technological and reliability issues of traditional FloatingGate NAND memories, respectively, while the state-of-the-art of charge trappingtechnologies is effectively summarized in Chap 5Ⅴ i PrefaceThe central part of the book is dedicated to circuit issues: logic( Chap. 6),sensing circuits(Chaps. 8 and 9)and high voltage blocks( Chaps. 1l and 12)Other basic topics related to circuit aspects are also analyzed, such as theimplementation of Double Data Rate interfaces( Chap. 7), while multilevel storage(2 bits per cell) is presented in Chap. 10Techniques adopted to increase memory reliability and yield are discussedin Chaps. 13 and 14, the former dealing with redundancy, the latter with ErrorCorrection Codes. The test flux used by nand memories manufactures is describedin Chap 15Chapter 16 focus on nand devices storing 3 and 4 bits per cell that allowreaching the lower cost per gigabyte and that will conquer, in the next few years,the market of consumer applicationsThe last chapters are dedicated to the two most popular systems based onNAND memories: Flash cards( Chap. 17)and SsD(Chap. 18). The relationshipbetween Nand memories and system performances are highlightedFinally, Chap. 19 describes the effects of ionizing radiations on non-volatilememories, to understand whether NaNd memories can be safely used in spaceand military applicationsProf. Piero olivoDean of the Engineering FacultyUniversity of ferrara, ItalyAcknowledgementsAfter completing a project like a technical book, it is very hard to acknowledgeall the people who have contributed directly or indirectly with their work anddedicationFirst of all, we wish to thank all the authors of the contributed chaptersWe have to thank mark de Jongh for giving us the possibility of publishing thiswork and cindy zitter for her continuous supportLast but not least, we keep in mind all our present and past colleagues for theirsuggestions and fruitful discussionsRino. Luca and alessiaTable of contentsPrefaceAcknowledgementsvIII Market and applications for nand Flash memoriesGregory Wong2 NAND overview: from memory to systemsR. Micheloni, A Marelli and S. Commodaro3 Program and erase of Nand memory arrays55Cristoph Friederich4 Reliability issues of NAND Flash memoriesC. Zambelli. a. Chimenton and p. olivo5 Charge trap nand technologies115Alessandro grossi6 Control logic131A Marelli.R Micheloni andR. ravasio7 NAND DDR interface161Andrea silvagni8 Sensing circuits197L. Crippa and R. micheloni9 Parasitic effects and verify circuits235L. Crippa and R. michelonO MLC Storage261L Crippa and R Micheloni11 Charge pumps, voltage regulators and HV switchesR Micheloni and L Crippa
- 2020-12-08下载
- 积分：1
TDOA定位技术原理及算法
对TDOA定位技术的基本原理的描述，基于时间测量值的蜂窝无线定位算法第1期郭华:TDOA定位技术的基本原理和算法21值的不确定性度量。对于二维定位系统,、CEP定义为CEP≈(0.75s)ODOP(11)包含了一半以均值为中心的随机矢量实现的圆半G丑OP可作为从大量基站中选择所需定位基站的指径。如果定位估计器为元偏差的CEP即为MS相对标,选中的基站是使(nOP最小的基站,还可用于其真实位置的不确定性度量如图2所示。如果佔计建立新系统时作为选择基站位置的参考。器为有偏差的且以偏差B为界,则对于50%概率,MS的信计位置在距离B+CEP内,此时,CEP为一3基于COST259信道模型的算法仿复杂函数,通常用其近似表示,对于TDOA双曲线主位,CEP近似表示为CEP=0.75J2+02为了从分评估CHAN算法和 Taylor算法在不其中202分别为二维估计位置的方差同信道环境、不同测量条件下的定位性能,本文设定了多种仿真条件,考虑了实际应用中的可能出现的发射机位置情况,并且每次都在相同的条件下对MS做定位佔计。考核定位结果的均方根误差(RMSE)并以图表的形式直观的表示出米,结合理论分析,对这两和主CEP偏差矢量要的无线定位算法做评估。平均估计位置实际应用中,对尢线定位精度造成影响的因素有许多种。在本文中主要考虑以下因素:小区半径的大小、参与定位的基站数目、不同的信道参数、设图2圆误差概率CEP备的测量误差、定位基站的排列形状等。综合考虑2.3几何精度因子((丑OP)以上因素,仿真所需的主要条件如采用距离测量方法的定位系统准确率在很大程1)针对COS259模型中的A、B、C、D四种信度上取决于基站和待定位移动台MS之间的几何位道做仿真。不同的信道其参数T不同。在考察小置关系几何位置对定位准确率影响的度量即为几区半径和定位基站数目对定位性能的影响时主要考何精度因子(GDOP),定义为定位误差RMSE与测虑了A和B两种信道,考察基站的排列形状时主要距误差RMSE比率。GDOP表征了由于移动台与基考虑B和D两种信道。站几何位置关系对测距误差的放大程度。对于无偏(2)检测设备造成的测量误差:假设检测设备精差计器,GDOP为:度造成的TDOA误差服从均值为0,标准差为30mGDOP=Ntr[(AA(7)的高斯正态分布。在考察设备的测量误差对定位精度的影响时,标准差分别取:30m、60m、90m、120m、其中,tr()表示对结果矩阵求迹,即求矩阵主对角150m线兀素之和aA为根据某种特征测量值建立的线性3.1理想信道环境下测量设备误差对定位的影响方程组的系数矩阵,即仿真条件:小区半径R-2000m,参与定位的基Y=AX(8)站数目为3~7个,MS在1/12小区内均匀分布,假这里,基站位置Y是已知的M×1维向量,MS位置设信道为理想的LOS信道,由信道造成的NIOS误X是2×1维未知向量[xy],A是Mx矩阵,如果差为0。仅仅考由于检测设备的测量误差对算法AA是非奇异矩阵且M>2,则式(8)为方程数大于定位性能的影响。比较受限 Taylor算法与CHAN未知量数目的超定方程组,采用最小二乘LS)算算法的定位性能法获得MS估计位置定位结果如图3、图4和图5所示X=(AAAY(9)从图3到图5可以看出,当仅有理想高斯分布对于无偏差的估计器及二维双曲线定位系统的测量误差时,参与定位基站的数目已经不是很重GDOP可表小为要了,基站数目的变化不会影响两种算法的定位性GDOP=x +o v/ s(10)能。定位性能与测量误差直接相关(基本呈线性增这里s为测距误差标准差。GDOP与CEP有以长)。从3个图屮都可以发现CHAN算法的定位性下近似关系能要好与 Taylor算法。也就是说CHAN算法的抗C1994-2010chinaAcademicjOurnalElectronicPublishingHouseAllrightsreservedhttp://www.cnkinet22西安邮电学学报2007年1月高斯分冇;NLOS误差服从COST259模型。仿真结果如图6、图7和图8所示。图3基站数为3时理想高斯测量误差条件下, MTavlor算法与CHAN算法的定位性能比较图6基站数目为3时, TAylor算法和CHAN算法在BadUrban和 Urban环境下图4基站数为4时理想高斯测量误差条件下 MTavlor算法与CHAN算法的定位性能比较图7基站数目为4时, MAylor算法和CHAN算法在Barban和 Urban环境下的性能比画量柱准展图5基站数为7时理想高斯测量误差条件下 MAylor算法与CHAN算法的定位性能比较高斯噪声性能更好,较适用于LOS信道环境。图8基站数目为7时, TAylor算法和CHAN算法在3.2COsT259环境下小区大小定位基站数目及 BadUrban和Uban环境下的性能比排列对定位性能的影响标识说明: Badurban chan表小在 Badurban(T参与定位基站数目与小区大小对定位性能的=1)环境下CHAN算法的定位性能曲线; Urban-影响Chan表示在 Urban(T=0.4)环境下CHAN算法的仿真条件:参与定位的基站数目固定为(3~7),定位性能曲线; BadUrban Taylor表示在 Badurban小区半径大小由100400变化。MS在112(T=1)环境下 M Taylor算法的定位性能曲线;Ur小区内均匀分布,在 Badurban和Ubam环境下比 ban Taylor表示在 Urban(T=0.4)环境下 M Taylor较受限的 Taylor算法和CHN算法的定位性能。算法的定位性能曲线。Taylor算法的初始位置取MS的实际位置;设从图6到图8可以看出,在实际信道中,CHAN备的测量误差服从均值为0,标准差为30m的理想算法的定位性能要比 Taylor算法的性能差,这是o01994-2010ChinaaCademicjOurnalElectronicPublishingHouseAllrightsreservedhttp:/www.cnkinet第1期郭TDOA定位技术的基本原理和算法23由于CHAN算法本身就是针对運想高斯LoS环境想蜂窝状排列时 MAylor算法的定位性能曲线;提出米的,所以在实际信道中,由于NLOS误差的U Highway Chan表示在 Urban(T=0.4)环境下引入导致了CHAN算法性能的急剧下降。基站直线排列时CHAN算法的定位性能曲线我们也可以从图中发现,当NLOS误差分不相Highway Chan表示在 Rural(T=0.1)环境下同时,参与定位的基站数日的多少对算法的定位性基站直线排列时CHAN算法的定位性能曲线能没有多大的影响。但当小区半径大于3000m时,U Chan表示在 Urban(T=0.4)环境下基站理Taylor算法的性能呈直线下降,而且当小区半径增想蜂窝状排列时CHAN算法的定位性能曲线;大时, Maylor算法有可能出现不收敛情况。而CHan表示在Rurl(T-0.1)环境下基站理想CHAN算法性能变化较为平稳,且在任何情况下都蜂窝状排列时CHAN算法的定位性能由线能够给出定位结果。因此可以得出结论,在其它条件相同的条件下CHAN算法比 Taylor算法更适合与宏小区的无线非理想的基站分布对定位性能的影响仿真条件:小区半径从1000m~4000m,参与定位的基站数目为3。在类似高速公路的直线区域内基站一般呈直线分布。MS在1/12小区内均匀分布。比较受限 Taylor算法和CHAN算法在 Urban小区使加和Rura环境下的定位性能,以及基站直线分布与哩想蜂窝状分布下两种算法的定位性能。 Taylor图10基站直线排列与理想蜂窝状排列时在 Urban和算法中MS的初始位置取MS的实际位置。设备测Rurl环境下CHAN算法定位性能的比较量误差服从均值为0,标准差为30m的理想高斯分从图9和图10可以看出,基站的位置分布对两布。非视距误差NLOS满是COST259信道模型。种算法的定位性能影响都很人,其它条件不变时,当仿真结果如图9和图10所示基站呈直线分布时,定位性能都急剧下降。如果做樻向比较,在小、区半径较小时, Taylor算法有较好的定位性能。4小结通过以上的仿真与分析比较可以看出小区半径对CHAN算法的影响的程度比Taylor算法较小,在其它条件不变,只有小区半径增大的量准圣与时,CHAN算法的定位性能变化较为平稳。而Taylor算法的定位性能变化较为剧烈。说明在相同的图9基站直线排列与理想蜂窝状排列时,在 Urban和条件下,CHAN算法更适合与宏小区的定位。Rural环境下 TAylor算法定位性能的比较标识说明在LOS环境下,TDOA测量值的误差服从均U HighwayMTay表示在 Urban(T=0.4)环境值为0的埋想高斯分布CHAN算法的定位性能优于 Taylor算法的定位性能。所以CHAN算法更适下基站直线排列时 MAylor算法的定位性能曲线;合于LOS环境的定位。RHighwayM Tay表示在 Rural(T=0.1)环境下基站直线排列时 TAylor算法的定位性能曲线岀基站呈直线排列时,两种算法的定位性能与UMTay表示在 Urban(T=0.4)环境下基站理理想的蜂窝状分布时的定位性能相比急剧下降。在小区半径较小时, Taylor算法有较好的定位性能。想蜂窝状排列时 MAylor算法的定位性能曲线信道环境对定位性能有很大的影响,特别是对RMTav表示在 Rural(T=0.1)环境下基站理CIAN算法的影响更大一些。在郊区和远郊两种C1994-2010ChinaAcademicJournalElectronicPublishingHouseAllrightsreservedhttp://www.cnki.net4西安邮电学学报2007年1月算法都能取得较好的定位效果,基本能够满足FWrokshop j une 2001911的定位要求;但在城区的效果就明显下降,在闹21 J ames J. Caffery:Jr. and Gordon L. Stuber: Overvie市区几乎不能准确定位。因此,要在市区和闹市区of Radiolocation in CDMA Cellular Systems. TEEF获得较好的定位效果,就需要对定位算法进行改进,Communications Magazine April 1998 pp 38-45对如何取得更精确的TDOA测量值,更好的克服3]An Overview of Wireless Indoor geolocation Techniqueand Systems. Kaven Pahlavan. Xinrong Li. MikaNLOS误差进行深入旳研究。Ylianttila. Ranvir chana. And matti latva- aho. Mo-参考文献bile and wireless Communications Networks. IFIPTC6/ European commission NETWORKiNG 2000 Inter-[I Do menico Porcino. Philips Research Laboratories. Stannational Workshop. MWCN 2000. Paris France. Maydardisation of Location Technologies Mobile Location2000Principle and algorithm of doa location technologyGUO HuaDepartment of Electronic of Information Engineering Xi an University of Post and Telecommunications, Xi an 710121, China)Abstract: This thesis researches the wireless location algorithms based on time-related measurements in thWireless Cellular Network. By analyzing existing basic techniques and algorit hms in wireless location this thesisselects the TDOA algorithm for emp hases of research. First, we int roduce two typical al gorithms as Taylor algo-rithm and chan algorithm. Furthermore, we analyze and compare them by simulation under some commonmobile channel environment. In order to eval uate the performance of the two basic al gorithms in detail, the perfect channel and two typical mo bile communication channel model (CoST 259 and TiPl)are employed for future detailed simulation. In simulations, many relative parameters w hich may effect the performance of the loca-tion algorithms are examined, such as the cell size, the number of the base stations taking part in the locationservice, equip ment measurement errors, NLOS effect etcKey words: Time of Arrive (TOA); Difference Time of Arrive( TDOA); Angel of Arrive(AOA); NLOS error(上接第18页)Error code protection and scheme of Turbo code forwreless video frequency transmissionXU Hua, DONG Yirning, XIA YangCollege of Communication and Information EngineeringNanjing University of Posts and Telecommunications, Nanjing 210003, China)Abstract The characteristics of wireless channel such as time- variety and high BER requires not only strongerror correcting capability of channel coding method but also the ability to adjust bit rate according to the state ofwireless channel. The Rate Compatible Turbo code (RCP T)can fulfill such requirement, and protect video fre-quency stream with tr- s interleaver. The application of rCPt over the mo bile communication channel is intro-duced, and new schemes about adaptive coding system and unequal error protection are discussedKey words: Turbo codes; RCPT codes; puncture table, unequal error protectionC1994-2010ChinaAcademicJournalElectronicPublishingHouse.Allrightsreservedhttp://www.cnki.net
- 2020-11-28下载
- 积分：1
|
联系客服（QQ：3324461878）
