<strike id="ksmoc"></strike>
提出了一種改進的LSM-ALSM子空間模式識別方法，將LSM的旋轉策略引入ALSM，使子空間之間互不關聯的情況得到改善，提高了ALSM對相似樣本的區分能力。討論中以性能函數代替經驗函數來確定拒識規則的參數，實現了識別率、誤識率與拒識率之間的最佳平衡；通過對有限字符集的實驗結果表明，LSM-ALSM算法有效地改善了分類器的識別率和可靠性。關 鍵 詞 學習子空間; 性能函數; 散布矩陣; 最小描述長度在子空間模式識別方法中，一個線性子空間代表一個模式類別，該子空間由反映類別本質的一組特征矢量張成，分類器根據輸入樣本在各子空間上的投影長度將其歸為相應的類別。典型的子空間算法有以下三種[1, 2]：CLAFIC(Class-feature Information Compression)算法以相關矩陣的部分特征向量來構造子空間，實現了特征信息的壓縮，但對樣本的利用為一次性，不能根據分類結果進行調整和學習，對樣本信息的利用不充分；學習子空間方法(Leaning Subspace Method, LSM)通過旋轉子空間來拉大樣本所屬類別與最近鄰類別的距離，以此提高分類能力，但對樣本的訓練順序敏感，同一樣本訓練的順序不同對子空間構造的影響就不同；平均學習子空間算法(Averaged Learning Subspace Method, ALSM)是在迭代訓練過程中，用錯誤分類的樣本去調整散布矩陣，訓練結果與樣本輸入順序無關，所有樣本平均參與訓練，其不足之處是各模式的子空間之間相互獨立。針對以上問題，本文提出一種改進的子空間模式識別方法。子空間模式識別的基本原理1.1 子空間的分類規則子空間模式識別方法的每一類別由一個子空間表示，子空間分類器的基本分類規則是按矢量在各子空間上的投影長度大小，將樣本歸類到最大長度所對應的類別，在類x()iω的子空間上投影長度的平方為()211,2,,()argmax()jMTkkjpg===Σx (1)式中 函數稱為分類函數；為子空間基矢量。兩類的分類情況如圖1所示。
上傳時間： 2013-12-25
上傳用戶：熊少鋒
用單片機配置FPGA—PLD設計技巧 Configuration/Program Method for Altera Device Configure the FLEX Device You can use any Micro-Controller to configure the FLEX device–the main idea is clocking in ONE BITof configuration data per CLOCK–start from the BIT 0The total Configuration time–e.g. 10K10 need 15K byte configuration file•calculation equation–10K10* 1.5= 15Kbyte–configuration time for the file itself•15*1024*8*clock = 122,880Clock•assume the CLOCK is 4MHz•122,880*1/4Mhz=30.72msec
上傳時間： 2013-10-09
上傳用戶：a67818601
摘 要:用一種新的思路和方法,先計算低通、再計算高通濾波器的有關參數,然后組合成帶通濾波器.關鍵詞:濾波器;參數;新思路中圖分類號: TN713. 5 文獻識別碼:B 文章編號:1008 - 1666 (1999) 04 - 0089 - 03A New Consideration of the Band Filter’s CalculationGuo Wencheng( S hao Yang B usiness and Technology school , S haoyang , Hunan ,422000 )Abstract :This essay deals with a new Method of calculating the band filters - first calculatingthe relevant parameters of low - pass filters ,then calculating the ones of high - pass filters.Key words :filter ; parameters ;new considercation八十年代后,信息產業得到了迅猛發展. 帶通濾波器在微波通信、廣播電視和精密儀器設備中得到了廣泛應用. 帶通濾波器性能的優劣,對提高接收機信噪比,防止鄰近信道干擾,提高設備的技術指標,有著十分重要的意義.我在長期的教學實踐中,用切比雪夫型方法設計、計算出寬帶濾波器集中參數元件的數據. 該濾波器可運用在檢測微波頻率的儀器和其他設備中. 再將其思路和計算方法介紹給大家,供參考.
上傳時間： 2014-12-28
上傳用戶：Yukiseop
針對使用硬件描述語言進行設計存在的問題，提出一種基于FPGA并采用DSP Builder作為設計工具的數字信號處理器設計方法。并按照Matlab/Simulink/DSP Builder/QuartusⅡ設計流程，設計了一個12階FIR 低通數字濾波器，通過Quartus 時序仿真及嵌入式邏輯分析儀SignalTapⅡ硬件測試對設計進行了驗證。結果表明，所設計的FIR 濾波器功能正確，性能良好。 Abstract: Aiming at the problems in designing DSP using HDL，a Method of designing DSP based on FPGA which using DSP Builder as designed tool is pointed out.A 12-order low-pass FIR digital filter was designed according to the process of Matlab/Simulink/DSP Builder/QuartusⅡ, and the design was verified by the timing simulation based on QuartusⅡand practical test based on SignalTapⅡ. The result shows the designed filter is correct in function and good in performance.
上傳時間： 2013-11-17
上傳用戶：lo25643
為了擴大監控范圍，提高資源利用率，降低系統成本，提出了一種多通道視頻切換的解決方案。首先從視頻信號分離出行場信號，然后根據行場信號由DSP和FPGA產生控制信號，控制多路視頻通道之間的切換，從而實現讓一個視頻處理器同時監控不同場景。實驗結果表明，該方案可以在視頻監控告警系統中穩定、可靠地實現視頻通道的切換。 Abstract: To expand the scope of monitoring， improve resource utilization， reduce system cost， a multiple video channels signal switching Method is pointed out in this paper. First， horizontal sync signal and field sync signal from the video signal are separated， then control signal according to the sync signal by DSP and FPGA is generated to control the switching between multiple video channels. Thus， it achieves to make a video processor to monitor different place. Experimental results show that the Method can realize video channel switching reliably， and is applied in the video monitoring warning system successfully.
上傳時間： 2013-11-09
上傳用戶：不懂夜的黑
This application note describes how to retrieve user-defined data from Xilinx configurationPROMs (XC18V00 and Platform Flash devices) after the same PROM has configured theFPGA. The Method to add user-defined data to the configuration PROM file is also discussed.The reference design described in this application note can be used in any of the followingXilinx FPGA architectures: Spartan™-II, Spartan-IIE, Spartan-3, Virtex™, Virtex-E, Virtex-II,and Virtex-II Pro.
上傳時間： 2013-11-11
上傳用戶：zhouli
針對傳統集成電路（ASIC）功能固定、升級困難等缺點，利用FPGA實現了擴頻通信芯片STEL-2000A的核心功能。使用ISE提供的DDS IP核實現NCO模塊，在下變頻模塊調用了硬核乘法器并引入CIC濾波器進行低通濾波，給出了DQPSK解調的原理和實現方法，推導出一種簡便的引入?仔/4固定相移的實現方法。采用模塊化的設計方法使用VHDL語言編寫出源程序，在Virtex-II Pro 開發板上成功實現了整個系統。測試結果表明該系統正確實現了STEL-2000A的核心功能。 Abstract: To overcome drawbacks of ASIC such as fixed functionality and upgrade difficulty, FPGA was used to realize the core functions of STEL-2000A. This paper used the DDS IP core provided by ISE to realize the NCO module, called hard core multiplier and implemented CIC filter in the down converter, described the principle and implementation detail of the demodulation of DQPSK, and derived a simple Method to introduce a fixed phase shift of ?仔/4. The VHDL source code was designed by modularity Method , and the complete system was successfully implemented on Virtex-II Pro development board. Test results indicate that this system successfully realize the core function of the STEL-2000A.
上傳時間： 2013-11-06
上傳用戶：liu123
波長信號的解調是實現光纖光柵傳感網絡的關鍵，基于現有的光纖光柵傳感器解調方法，提出一種基于FPGA的雙匹配光纖光柵解調方法，此系統是一種高速率、高精度、低成本的解調系統，并且通過引入雙匹配光柵有效地克服了雙值問題同時擴大了檢測范圍。分析了光纖光柵的測溫原理并給出了該方案軟硬件設計，綜合考慮系統的解調精度和FPGA的處理速度給出了基于拉格朗日的曲線擬合算法。 Abstract: Sensor is one of the most important application of the fiber grating. Wavelength signal demodulating is the key techniques to carry out fiber grating sensing network， based on several existing Methods of fiber grating sensor demodulation inadequate， a two-match fiber grating demodulation Method was presented. This system is a high-speed， high precision， low-cost demodulation system. And by introducing a two-match grating effectively overcomes the problem of double value while expands the scope of testing. This paper analyzes the principle of fiber Bragg grating temperature and gives the software and hardware design of the program. Considering the system of demodulation accuracy and processing speed of FPGA，this paper gives the curve fitting algorithm based on Lagrange.
上傳時間： 2014-07-24
上傳用戶：caiguoqing
多遠程二極管溫度傳感器-Design Considerations for pc thermal management Multiple RDTS (remote diode temperature sensing) provides the most accurate Method of sensing an IC’s junction temperature. It overcomes thermal gradient and placement issues encountered when trying to place external sensors. PCB component count decreases when using a device that provides multiple inputs.Better temperature sensing improves product performance and reliability. Disk drive data integrity suffers at elevated temperatures. IBM published an article stating that a 5°C rise in operating temperature causes a 15% increase in the drive’s failure rate. The overall performance of a system can be improved by providing a more accurate temperature measurement of the most critical devices allowing them to run just a few degrees hotter.The LM83 directly senses its own temperature and the temperature of three external PN junctions. One is dedicated to the CPU of choice, the other two go to other parts of your system that need thermal monitoring such as the disk drive or graphics chip. The SMBus-compatible LM83 supports SMBus timeout and logic levels. The LM83 has two interrupt outputs; one for user-programmable limits and WATCHDOG capability (INT), the other is a Critical Temperature Alarm output (T_CRIT_A) for system power supply shutdown.
標簽： Considerat Design 遠程 二極管
上傳時間： 2014-12-21
上傳用戶：ljd123456
本文主要通過介紹PLC通訊的意義和三菱FX系列PLC的四種通訊方式，并重點介紹FX系列PLC與計算機無協議通訊，主要從無協議通訊的硬件、配線、數據寄存器設置、PLC與計算機無協議通訊的指令用法、PLC程序編寫和計算機VB程序的編寫來說明無協議通訊的過程和一般方法。 My dissertation introduces the significance of PLC communications and the four means of communication of Mitsubishi FX’s PLC, And highlights the no protocol communications of FX series PLC and computer, no protocol communications hardware, wiring, Register data set, and the usage of command about no protocol communications, How to write PLC program and computer VB program to illustrate the process of no protocol communications and general Method.
上傳時間： 2014-11-29
上傳用戶：Jerry_Chow
