發(fā)布時(shí)間：2018-11-02 17:07

【摘要】：DSP是以數(shù)字信號(hào)來分析大量信息的一種先進(jìn)的微處理器器件。其基本理論是將接收到的模擬信號(hào)轉(zhuǎn)換為1或0的數(shù)字信號(hào),然后對(duì)這個(gè)信號(hào)進(jìn)行弱化、修正、加強(qiáng),同時(shí)通過芯片進(jìn)行處理,把得到的數(shù)字信號(hào)數(shù)據(jù)轉(zhuǎn)換回模擬信號(hào)數(shù)據(jù)或者現(xiàn)實(shí)應(yīng)用中真實(shí)的環(huán)境格式。歷經(jīng)近數(shù)十年的飛速發(fā)展,利用DSP技術(shù)的商品已經(jīng)遍及人們?nèi)粘W(xué)習(xí)、工作和生活的方方面面。由此而產(chǎn)生的DSP測(cè)試系統(tǒng)設(shè)計(jì)已引起業(yè)界的關(guān)注�，F(xiàn)今無論在通訊、計(jì)算機(jī)行業(yè)中,甚至在人們?nèi)粘Ｉ畛Ｓ玫闹T多產(chǎn)品行業(yè)中,都已逐步運(yùn)用到這種集成度超高的由DSP芯片構(gòu)成的系統(tǒng)。實(shí)際應(yīng)用中測(cè)試系統(tǒng)往往因受到傳感器的工作頻帶過窄等原因,導(dǎo)致將某些有用的信號(hào)頻率分量畸形化,最終出現(xiàn)測(cè)試數(shù)據(jù)失真的問題。傳統(tǒng)的單片機(jī)或者集成電路甚至二者的結(jié)合,均無法實(shí)時(shí)修正上述原因產(chǎn)生的誤差�？梢�,測(cè)試系統(tǒng)的響應(yīng)時(shí)間、接收信號(hào)的瞬時(shí)性及真實(shí)性等參數(shù),均為考查該系統(tǒng)性能優(yōu)劣的重要指標(biāo)之一。為了能夠?qū)崟r(shí)修正誤差,則需要采用合適的數(shù)據(jù)處理芯片及算法使系統(tǒng)能夠具有誤差修正功能。本文采用TMS320F2812DSP芯片和自適應(yīng)神經(jīng)網(wǎng)絡(luò)動(dòng)態(tài)補(bǔ)償算法,針對(duì)TMS320F2812DSP芯片特有的性能和結(jié)構(gòu)特點(diǎn),通過設(shè)計(jì)其外圍硬件結(jié)構(gòu)并結(jié)合外擴(kuò)存儲(chǔ)器,設(shè)計(jì)了具有誤差修正功能和智能接口的DSP應(yīng)用系統(tǒng)。將信號(hào)的動(dòng)態(tài)輸入、輸出信號(hào)進(jìn)行校正,創(chuàng)建數(shù)學(xué)模型。并且同時(shí)將信號(hào)逆建模,也就是將信號(hào)的輸出看做是補(bǔ)償系統(tǒng)的輸入,信號(hào)的輸入信號(hào)看做是補(bǔ)償系統(tǒng)的輸出,基于神經(jīng)網(wǎng)絡(luò)擬合算法最后獲得相應(yīng)的輸入和輸出的線性關(guān)系。通過實(shí)驗(yàn)驗(yàn)證得知:該測(cè)試系統(tǒng)能夠正確地采集存儲(chǔ)數(shù)據(jù),并且可以高效彌補(bǔ)信號(hào)的動(dòng)態(tài)誤差,提高了信號(hào)調(diào)理的精度和實(shí)時(shí)性。論文的創(chuàng)新點(diǎn)如下:選用了核心器件TMS320F2812 DSP作為主要控制單元,并且基于其功能特性針對(duì)性地搭配性能良好的外圍電路。數(shù)據(jù)處理上,利用自適應(yīng)神經(jīng)網(wǎng)絡(luò)算法,進(jìn)行逆建模,有效的提高了信號(hào)調(diào)理的精度和實(shí)時(shí)性。并將DSP的高速性和誤差校正性與測(cè)試系統(tǒng)的智能拓展接口有效的整合在一起。本文設(shè)計(jì)的重點(diǎn)內(nèi)容:主要是針對(duì)TMS320F2812DSP芯片特有的性能進(jìn)行分析,針對(duì)性地設(shè)計(jì)相應(yīng)的外圍電路,同時(shí)對(duì)于外接輸入的數(shù)據(jù)進(jìn)行處理與研究。
[Abstract]:DSP is an advanced microprocessor device which uses digital signal to analyze a lot of information. The basic theory is to convert the received analog signal to a digital signal of 1 or 0, and then weaken, modify, strengthen, and process it by chip. Convert the obtained digital signal data back to analog signal data or real environmental format in real applications. After decades of rapid development, DSP technology has been used in all aspects of people's daily study, work and life. The design of DSP test system has attracted the attention of the industry. Nowadays, no matter in communication, computer industry, and even in many product industries commonly used in people's daily life, it has been gradually applied to this kind of system composed of DSP chips with high level of integration. In practical application, some useful signal frequency components are deformed due to the narrow working frequency band of the sensor, which leads to the distortion of the test data. Traditional microcontroller, integrated circuit or even the combination of the two, can not correct the error caused by the above reasons in real time. It can be seen that the response time of the test system, the instantaneous nature and authenticity of the received signal are all important indexes to test the performance of the system. In order to correct the error in real time, it is necessary to use the appropriate data processing chip and algorithm to make the system have error correction function. In this paper, the TMS320F2812DSP chip and the adaptive neural network dynamic compensation algorithm are adopted. According to the special performance and structural characteristics of the TMS320F2812DSP chip, the peripheral hardware structure is designed and combined with the extended memory. A DSP application system with error correction function and intelligent interface is designed. The dynamic input and output signals are corrected and the mathematical model is created. At the same time, the inverse model of the signal is modeled, that is, the output of the signal is regarded as the input of the compensation system, and the input signal of the signal is regarded as the output of the compensation system. Finally, the linear relationship between the input and the output is obtained based on the neural network fitting algorithm. The experimental results show that the system can collect and store data correctly, and can compensate the dynamic error of signal efficiently, and improve the precision and real-time of signal conditioning. The innovation of this paper is as follows: the core device TMS320F2812 DSP is selected as the main control unit, and based on its functional characteristics, the peripheral circuits with good performance are selected. In data processing, adaptive neural network algorithm is used for inverse modeling, which effectively improves the precision and real time of signal conditioning. The high speed and error correction of DSP are effectively integrated with the intelligent extended interface of the test system. The main content of this paper is to analyze the characteristic performance of TMS320F2812DSP chip, to design the corresponding peripheral circuit, and to process and study the external input data.
【學(xué)位授予單位】：上海交通大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2014
【分類號(hào)】：TP332;TP274

【參考文獻(xiàn)】

相關(guān)期刊論文 前2條

1 董健,蔣建偉,萬麗珍;CPLD與單片機(jī)在超壓存儲(chǔ)測(cè)試系統(tǒng)中的應(yīng)用[J];測(cè)試技術(shù)學(xué)報(bào);2005年01期

2 吳德會(huì);;基于最小二乘支持向量機(jī)的傳感器非線性動(dòng)態(tài)補(bǔ)償[J];儀器儀表學(xué)報(bào);2007年06期

相關(guān)碩士學(xué)位論文 前1條

1 嚴(yán)政;建模軟件包及通用動(dòng)態(tài)補(bǔ)償濾波器的研制[D];南京理工大學(xué);2002年

，

本文編號(hào)：2306416