After compared these methods , power spectral frequency band intensity , pca and linear network were choose to carry out the recognition system . finally , this paper provided software of gui as well as a group of simulink blocks to operate data and acquire results 在综合比较各种方法后,最终选择功率谱估计频带强度为分类特征, pca用于特征提取,线性神经网络作为分类器用于系统的实现。
The ball mill load - control apparatus adopted the assembly language and the c51 language to program and its concerned software was designed , including the data collection , the wave - percolation , the short - time average over - zero ratio , the estimation of the power chart and the average - period chart 磨机负荷控制仪软件使用汇编和c51编程，本文进行了相关软件的设计，包括数据采集、滤波、短时平均过零率、功率谱估计和平均周期图等。
In the end , the paper forecasts the trend in further research into digital ultrasonic signal processing , for example , how to use dsp to complete the data processing based on power spectral evaluation computing to reach a high speed , so that the problem of real - time during detection can be solved 展望了数字式超声信号处理技术进一步的研究方向，如采用dsp硬件来完成功率谱估计的数据处理以提高计算速度，更好地解决检测的实时性问题。
4 different types ’ features were generated , namely ar model parameters , power spectral frequency band intensity , energy for wavelet packet decomposition , wavelet packet entropy . every type of features were extracted respectively using pca and ica method and classified using linear neural network , knn and bp network 建立了ar模型参数、功率谱估计频带强度、小波包分解能量比率、小波包熵四种特征,分别使用pca与ica进行特征提取,采用线性神经网络、 k -紧邻法、 bp神经网络四种分类器进行分类。
To solve this problem , this thesis will focus on investigation of the following detection methods based on the features of the dsss signal and with the binary - phase - shift - keying dsss ( dsss / bpsk ) signal as an example . 1 . the power spectrum of the dsss / bpsk signal is discussed and the detection performance in each power spectrum estimation method is compared in terms of nonparametric methods , parametric methods and subspace methods in power spectrum estimation 本文针对这一问题,从dsss信号特征出发,以二进制相移键控的直接序列扩频( dsss / bpsk )信号为例,研究了以下检测方法: 1 .从功率谱估计的非参数方法、参数方法和子空间方法入手,讨论了通过不同类型方法估计的dsss / bpsk信号功率谱,并比较了各功率谱估计方法应用于dsss / bpsk信号检测的性能。
So it is essential to diminish the errors and enhance its anti - jamming ability . an innovative method of measuring pipeline thickness based on power spectral evaluation computing has been put forward . besides , experiments on the on - line measurement have yielded good results , proving the veracity and reliability of the measurement 介绍了为减小系统的测量误差、提高系统抗干扰能力而进行的对超声波回波信号的分析研究工作，创新地提出了一种基于对超声脉冲多次反射回波功率谱估计的超声测厚方法，并通过实验证明了该方法用于在线检测的准确性与可靠性。
It was demonstrated that this sort of load - control apparatus has superior performance . we used a electric transmitter as the transducer to collect the sound signal , using the daq toolbox of the matlab to achieve the collection of the sound signal , and using the dsp toolbox of the matlab and the sound - launch technology to analyze the time - domain characteristic and the frequency - domain characteristic of the signal , chiefly analyze its short - time average amplitude , its short - time over - zero ratio and its estimate of the power chart and the average - period chart 实验中使用了驻极体话筒作为球磨机声音拾取传感器，利用matlab中的数据采集( daq )工具箱实现对磨音的采集，利用matlab信号分析( dsp )工具箱和声发射技术对磨音进行时域和频域分析，主要分析磨音的短时平均幅度和能量、短时平均过零率、功率谱估计和平均周期图。
Through the inspecting , the author selected three methods of eradiate noise ' s feature extraction . the methods include the pedigree , e - filter , fractal . in those methods , the pedigree is based on frequency field , the e - filter is based on the energy field and the fractal is based on time field 本文作者经过考察，采用了三种前期研究中较为有效的水下目标特征提取方法? ?基于功率谱估计的线谱特征提取方法、基于小波分析的不同频段内能量特征提取方法和基于水声信号分维特性的特征提取方法。
The 10 % ~ 20 % international standard electrodes system contains 16 electrodes ’ position . sometimes they are fully redundant so we want to reduce them . after discussing the power spectral intensity and features ’ separability between the electrodes , the channels of 5 ~ 10 were decided finally 10 % ~ 20 %国际电极位置系统提供了16个电极的摆放位置,但是这些电极对于少数几种心理作业的识别存在很大冗余,通过对电极位置处功率谱估计以及特征可分性的讨论选择了5 ~ 10通道用于分类系统。
The recursive criterion of transient current protection based on mallat wavelet transform is presented in order to meet the real - time demands . the reserch shows deficiency existed in the prevenient principle of transient current protection , so a new improved scheme of transient current protection is proposed . the characteristics of transient protection at different bus distribution capacitance is also analysed in this thesis . the study shows that operate of transient protection is correct when the bus distribution capacitance change . other conclusion include : the trip of transient protection must be locked when the power circuit breaker of transimission line operated , because the trip is incorrect at this time ; the features of transient current and it ' s energy is different between the fault and thunder over - voltage in transimission line , and them can be identified ; the ratio of high frequency component of current in serise capaciters compensation line , svc and hvdc system is far little than fault transient current , so all of them do not affect the working of trasient current protection 应用功率谱估计方法讨论了故障高频暂态噪声信号在频率轴上的分布情况，利用功率谱分析结果实现的非参数化系统辨识方法研究了母线分布电容对电流高频暂态信号的衰减特性。论文的后半部分(第五,六,七章)着重分析了现有基于故障电流的暂态保护原理和判据存在的问题，提出了新的实现方案，并讨论了暂态电流保护的运行性能。利用多分辨分析和小波分解快速算法构成的递推暂态保护判据，可以准确区分故障线路和非故障线路，实现超高速暂态电流保护，计算更为简单方便，计算速度更快。
