WO2022006980A1 - Chirp信号Hammerstein模型系统辨识方法 - Google Patents

Chirp信号Hammerstein模型系统辨识方法 Download PDF

Info

Publication number
WO2022006980A1
WO2022006980A1 PCT/CN2020/102721 CN2020102721W WO2022006980A1 WO 2022006980 A1 WO2022006980 A1 WO 2022006980A1 CN 2020102721 W CN2020102721 W CN 2020102721W WO 2022006980 A1 WO2022006980 A1 WO 2022006980A1
Authority
WO
WIPO (PCT)
Prior art keywords
chirp signal
output
motor
amplitude
frequency domain
Prior art date
Application number
PCT/CN2020/102721
Other languages
English (en)
French (fr)
Inventor
曹仲晴
向征
郭璇
Original Assignee
瑞声声学科技(深圳)有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 瑞声声学科技(深圳)有限公司 filed Critical 瑞声声学科技(深圳)有限公司
Publication of WO2022006980A1 publication Critical patent/WO2022006980A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G01MEASURING; TESTING
    • G01RMEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/34Testing dynamo-electric machines
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/02Preprocessing
    • G06F2218/04Denoising
    • G06F2218/06Denoising by applying a scale-space analysis, e.g. using wavelet analysis

Definitions

  • the invention relates to the technical field of tactile perception, in particular to a method for identifying a Chirp signal Hammerstein model system.
  • linear motors are increasingly widely used in mobile terminals such as mobile phones. In order to achieve more precise control of linear motor systems, it is necessary to improve the motor modeling accuracy.
  • the invention provides a method for identifying the Hammerstein model system of Chirp signal, which realizes that the sequence frequency and amplitude can be arbitrarily designed like the step sequence signal.
  • the step sequence signal is in the form of a segment-by-segment single-frequency continuous signal.
  • the steps are the same; while the Chirp signal is a continuous signal with a continuously changing frequency, the advantage of the Chirp signal is used to shorten the identification time of the motor Hammerstein model system.
  • the present invention provides a kind of Chirp signal Hammerstein model system identification method including:
  • Step S10 Design and output the variable amplitude Chirp signal x(t);
  • Step S20 use the variable amplitude Chirp signal x(t) to excite the motor to obtain a first output, and use Fourier transform on the first output to obtain a frequency domain response;
  • Step S30 according to the frequency domain response, use the frequency domain analysis of the inverse signal of the variable amplitude Chirp signal x(t) to calculate the motor system response function H(t);
  • Step S40 using the motor system response function H(t) to obtain the kernel function Kernals through transformation matrix transformation;
  • Step S50 obtain the model output y_est from the variable amplitude Chirp signal x(t) and the kernel function Kernals;
  • Step S60 Comparing the model output y_est with the first output to obtain a difference model error.
  • variable amplitude Chirp signal x(t) is calculated by the following formula:
  • f 1 is the starting frequency of Chirp
  • a(t) is the amplitude
  • t is the time, that is, a(t) is a function of the frequency that changes with time
  • T is the duration of the Chirp signal
  • f 2 is the cut-off frequency of the Chirp signal.
  • Y is the frequency domain response
  • X_ is the frequency domain analysis of the inverse signal of the variable amplitude Chirp signal
  • a(t) is the amplitude
  • k i (t) is the kernel function Kernals
  • A is the transformation matrix
  • H i (t) represents the motor system response of the i-th harmonic response
  • i is a natural number.
  • t is the time
  • x(t) is the input variable amplitude Chirp signal
  • k i (t) is the i-th order kernel function Kernals
  • i is a natural number.
  • a 0 is the constant magnitude of a(t), where a(t)
  • a(t) a 0 ⁇ (t), which is a normalized variable voltage curve .
  • y is the first output.
  • the identification method of the Hammerstein model system of the Chirp signal provided by the present invention can realize the arbitrary design of the sequence frequency and amplitude as the step sequence signal, utilize the advantages of the Chirp signal to shorten the identification time of the Hammerstein model system of the motor, greatly improve the identification efficiency of the motor model, and further Improve the design of the motor system by modeling, and improve the user experience of tactile feedback.
  • FIG. 1 is a schematic flowchart of a method for identifying a Chirp signal Hammerstein model system according to an embodiment of the present invention
  • FIG. 2 is a schematic structural diagram of a limit voltage test platform for a linear motor provided by an embodiment of the present invention
  • FIG. 3 is a schematic diagram of a test result of a 6V variable-amplitude Chirp signal provided by an embodiment of the present invention
  • FIG. 4 is a schematic diagram of a test result of a 2V variable-amplitude Chirp signal provided by an embodiment of the present invention
  • FIG. 5 is a schematic diagram of a test result of a 1V variable-amplitude Chirp signal according to an embodiment of the present invention.
  • the present invention provides a method for identifying a Chirp signal Hammerstein model system, the method comprising:
  • Step S10 Design and output the variable amplitude Chirp signal x(t);
  • Step S20 use the variable amplitude Chirp signal x(t) to excite the motor to obtain a first output, and use Fourier transform on the first output to obtain a frequency domain response;
  • Step S30 according to the frequency domain response, use the frequency domain analysis of the inverse signal of the variable amplitude Chirp signal x(t) to calculate the motor system response function H(t);
  • Step S40 using the motor system response function H(t) to obtain the kernel function Kernals through transformation matrix transformation;
  • Step S50 obtain model output y_est by variable amplitude Chirp signal x(t) and described kernel function Kernals;
  • Step S60 Comparing the model output y_est with the first output to obtain a difference model error.
  • variable amplitude signal is as follows:
  • the frequency domain inverse filter of unequal amplitude signal is derived as follows:
  • ⁇ (t) is a convex function whose global minimum is in It can be seen that this is the same as the defined group delay.
  • n>2 can be omitted, and:
  • step S10 of the Chirp signal Hammerstein model system identification method provided by the present invention the specific calculation formula of the variable amplitude Chirp signal x(t) is:
  • f 1 is the starting frequency of Chirp
  • a(t) is the amplitude
  • t is the time, that is, a(t) is a function of the frequency that changes with time
  • T is the duration of the Chirp signal
  • f 2 is the cut-off frequency of the Chirp signal.
  • variable amplitude Chirp signal x(t) to excite the motor to obtain a first output
  • Fourier transform to obtain a frequency domain response for the first output
  • set the first output to be y
  • the frequency The domain response is Y
  • the formula for the Fourier change of the frequency domain response Y by the first output y is expressed as:
  • the motor system response function H(t) is analytically calculated in the frequency domain using the inverse signal of the variable amplitude Chirp signal x(t) according to the frequency domain response; the calculation formula of the motor system response function H(t) is:
  • Y is the frequency domain response
  • X_ is the frequency domain analysis of the inverse signal of the variable amplitude Chirp signal
  • a(t) is the amplitude
  • Y is the frequency domain response
  • X_ is the frequency domain analysis of the inverse signal of the variable amplitude Chirp signal
  • a(t) is the amplitude
  • k i (t) is the kernel function Kernals
  • A is the transformation matrix
  • H i (t) represents the motor system response of the i-th harmonic response
  • i is a natural number.
  • the amplitude a(t) is mostly a fixed value a 0 , and the amplitude is reduced only at the frequency points where some motors are easy to crack.
  • the model output y_est is obtained by the variable amplitude Chirp signal and the kernel function Kernals; the calculation formula of the model output y_est is:
  • t is the time
  • x(t) is the input variable amplitude Chirp signal
  • k i (t) is the i-th order kernel function Kernals
  • i is a natural number.
  • y is the first output.
  • variable-amplitude Chirp signal is tested using a limit voltage test platform of a linear motor.
  • the test platform includes a computer PC10, a motor 20, a tool 30, and a sponge.
  • the body 40 accelerometer 50, acquisition card 80, first amplifier 60 and second amplifier 70, wherein the PC10 is connected to the acquisition card 80, and the acquisition card 80 uses an NI-DAQ 4431 acquisition card;
  • the motor 20 is adhesively fitted on the tooling 30, the tooling 30 is placed and installed on the sponge body 40 to avoid the influence of the environment on the measurement results, and the motor 20 is a linear resonance driver (Linear resonance driver, LRA) ;
  • the accelerometer 50 is installed on the tooling 30 to measure the acceleration of the tooling 30 in the vibration direction of the motor 20, the accelerometer 50 is connected to the first amplifier 60, and the first amplifier 60 is connected to the The capture card 80, the second amplifier 70 is connected to the capture card 80 and the motor 20; specifically:
  • the generated digital signal is sent to the acquisition card 80 on the PC 10 for digital-to-analog conversion into an analog signal, and is amplified by the second amplifier 70 to excite the motor 20 to vibrate.
  • the vibration signal is collected and sent to the first amplifier 60, and the first amplifier 60 amplifies the vibration signal.
  • the collection card 80 simultaneously collects and measures the acceleration y(n) in the vibration direction and the voltage x(n) of the excitation motor.
  • Fig. 3, Fig. 4 and Fig. 5 are respectively the test results of the 6V variable amplitude Chirp signal, the 2V variable amplitude Chirp signal and the 1V variable amplitude Chirp signal tested by using the test platform, wherein the line A represents the actual output of the motor The signal, line B represents the model output y_est, and line C represents the difference model error result. It can be seen that the difference model error calculated by the variable amplitude Chirp signal is small, and it is easy to realize the identification of the motor Hammerstein model system, which greatly shortens the identification time of the motor Hammerstein model system and greatly improves the identification efficiency of the motor model.
  • the method for identifying the Hammerstein model system of the Chirp signal realizes that the sequence frequency and amplitude can be arbitrarily designed like the step sequence signal, and the advantages of the Chirp signal are used to shorten the identification time of the Hammerstein model system of the motor, and greatly improve the frequency and amplitude of the sequence.
  • the identification efficiency of the motor model further improves the design of the motor system by modeling, and improves the user experience of tactile feedback.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Tests Of Circuit Breakers, Generators, And Electric Motors (AREA)
  • Control Of Electric Motors In General (AREA)

Abstract

本发明涉及触觉感知技术领域,提供一种Chirp信号Hammerstein模型系统辨识方法,该方法包括:设计并输出变幅度Chirp信号;输出所述变幅度Chirp信号激励马达,得到第一输出,并对所述第一输出使用傅里叶变换得到频域响应;根据所述频域响应使用变幅度Chirp信号的逆信号频域解析计算马达系统响应函数H(t);使用所述马达系统响应函数H(t)通过转换矩阵转换得到函数Kernals;由变幅度Chirp信号与所述核函数Kernals得到模型输出y_est;使用所述模型输出y_est与马达实际输出对比得到差异模型误差。通过本发明提供技术方案,缩短了对马达Hammerstein模型系统的辨识时间,较大地提升了马达模型的辨识效率,进一步提升建模对马达系统的设计,提高了用户触觉反馈的体验效果。

Description

Chirp信号Hammerstein模型系统辨识方法 【技术领域】
本发明涉及触觉感知技术领域,尤其涉及一种Chirp信号Hammerstein模型系统辨识方法。
【背景技术】
线性马达作为一种用户体验更好的触觉反馈器件,日渐在手机等移动终端上得到广泛应用。为了实现对线性马达系统更精确的控制,提高马达建模精度十分必要。
现有技术中一般利用恒定幅度Chirp信号Hammerstein模型系统辨识方法。其缺点为,实际非线性系统不同频率最大可承受安全电压(极限电压)不是恒定值。如果使用恒定幅度Chirp信号进行辨识,则只能使用整个辨识频率范围内极限电压最小值以下的电压进行系统辨识。
前期工作已经发现,对于线性马达系统,不同幅度信号具有不同的核函数,某一电压下核函数用于其它电压会引起较大误差。由于起始处f0极限电压小,且大电压时特定频率存在异向打壳现象(打壳现象即马达振子打到马达外壁上,在实际使用中应极力避免这种现象发生),恒定幅度Chirp信号无法避免在达到较大电压发生打壳现象,这使得恒定幅度Chirp信号可用幅度限制在较小的电压范围,这限制了核函数应用的电压范围,也就限制了对Hammerstein模型系统辨识,进而影响到后续的建模对马达的驱动,这样影响了后续马达系统的设计,影响到用户体验触觉反馈的效果差。
【发明内容】
本发明提供一种Chirp信号Hammerstein模型系统辨识方法,实现和step序列信号一样任意设计序列频率与幅度,step序列信号形式为一段一段的单频连续信号,每段之间频率逐渐变化,像step即阶梯一样;而Chirp 信号是频率连续变化的连续信号,利用Chirp信号的优点缩短马达Hammerstein模型系统的辨识时间。
为实现上述目的,本发明提供一种Chirp信号Hammerstein模型系统辨识方法包括:
步骤S10:设计并输出变幅度Chirp信号x(t);
步骤S20:使用所述变幅度Chirp信号x(t)激励马达,得到第一输出,并对所述第一输出使用傅里叶变换得到频域响应;
步骤S30:根据所述频域响应使用变幅度Chirp信号x(t)的逆信号频域解析计算马达系统响应函数H(t);
步骤S40:使用所述马达系统响应函数H(t)通过转换矩阵转换得到核函数Kernals;
步骤S50:由变幅度Chirp信号x(t)与所述核函数Kernals得到模型输出y_est;
步骤S60:将所述模型输出y_est与所述第一输出对比得到差异模型误差。
进一步地,所述变幅度Chirp信号x(t)通过以下公式计算得到:
Figure PCTCN2020102721-appb-000001
Figure PCTCN2020102721-appb-000002
其中,f 1为Chirp的起始频率,a(t)为幅度,t为时间,即a(t)为随时间变化的频率的函数,T为Chirp信号时长,f 2为Chirp信号截止频率。
进一步地,设定第一输出为y,所述频域响应为Y,则所述频域响应Y由所述第一输出y进行傅里中变化的公式表示为:
Y=fft(y)。
进一步地,所述马达系统响应函数H(t)的计算公式为:
H(t)=Y*X_*1/a(t)
其中,Y为频域响应,X_为变幅度Chirp信号的逆信号频域解析,a(t)为幅度。
进一步地,所述核函数Kernals的计算公式为:
k i(t)=AH i(t)
其中,k i(t)为核函数Kernals,A为转换矩阵,H i(t)表示第i次谐波响应的马达系统响应,i为自然数。
进一步地,所述模型输出y_est的计算公式为:
Figure PCTCN2020102721-appb-000003
其中,t为时间,x(t)为输入的变幅度Chirp信号,k i(t)为第i阶核函数Kernals,i为自然数。
进一步地,所述核函数的阶数为5时,转换矩阵A表示为:
Figure PCTCN2020102721-appb-000004
a 0为a(t)的常数幅值,其中,a(t)
与a0之间的公式为:a(t)=a 0·γ(t),为归一化的可变电压曲线
进一步地,设所述差异模型误差为ε t,则其计算公式为:
ε(t)=y(t)-y_est(t)
其中,y为所述第一输出。
本发明提供的Chirp信号Hammerstein模型系统辨识方法,实现和step序列信号一样任意设计序列频率与幅度,利用Chirp信号的优点缩短马达Hammerstein模型系统的辨识时间,较大地提升了马达模型的辨识效率,进一步提升建模对马达系统的设计,提高了用户触觉反馈的体验效果。
【附图说明】
为了更清楚地说明本发明实施例中的技术方案,下面将对实施例描述中所需要使用的附图作简单地介绍,显而易见地,下面描述中的附图仅仅是本发明的一些实施例,对于本领域普通技术人员来讲,在不付出创造性劳动的前提下,还可以根据这些附图获得其它的附图,其中:
图1为本发明一实施例提供的Chirp信号Hammerstein模型系统辨识方法流程示意图;
图2为本发明一实施例提供的线性马达的极限电压测试平台的结构示意图;
图3为本发明一实施例提供的6V变幅度Chirp信号的测试结果示意图;
图4为本发明一实施例提供的2V变幅度Chirp信号的测试结果示意图;
图5为本发明一实施例提供的1V变幅度Chirp信号的测试结果示意图。
【具体实施方式】
下面将结合本发明实施例中的附图,对本发明实施例中的技术方案进行清楚、完整地描述,显然,所描述的实施例仅是本发明的一部分实施例,而不是全部的实施例。基于本发明中的实施例,本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其它实施例,都属于本发明保护的范围。
请参阅图1,本发明提供一种Chirp信号Hammerstein模型系统辨识方法,该方法包括:
步骤S10:设计并输出变幅度Chirp信号x(t);
步骤S20:使用所述变幅度Chirp信号x(t)激励马达,得到第一输出,并对所述第一输出使用傅里叶变换得到频域响应;
步骤S30:根据所述频域响应使用变幅度Chirp信号x(t)的逆信号频域解析计算马达系统响应函数H(t);
步骤S40:使用所述马达系统响应函数H(t)通过转换矩阵转换得到核函数Kernals;
步骤S50:由变幅度Chirp信号x(t)与所述核函数Kernals得到模型输 出y_est;
步骤S60:将所述模型输出y_est与所述第一输出对比得到差异模型误差。
具体地,在等幅Chirp信号系统辨识的基础上通过下列的变幅度Chirp信号x(t)的系统辨识推导:
变幅信号如下:
Figure PCTCN2020102721-appb-000005
其中,
Figure PCTCN2020102721-appb-000006
若a[φ(t)]为变幅度电压,为了保持其与信号在频域的同步,其瞬时频率φ(t)也为:
Figure PCTCN2020102721-appb-000007
Figure PCTCN2020102721-appb-000008
在利用频域法进行辨识时,需要得到x(t)的频域逆滤波器,在文献‘Synchronized Swept-sine:Theory,Application and Implementation’中,等幅信号频域逆滤波器为文献中(43)式为:
Figure PCTCN2020102721-appb-000009
不等幅信号频域逆滤波器推导如下:
对不等幅信号,其对应的解析函数表达式如下:
Figure PCTCN2020102721-appb-000010
若积分中的相位用
Figure PCTCN2020102721-appb-000011
表示,则可以发现,Θ(t)是一个凸函数,其全局最小值在
Figure PCTCN2020102721-appb-000012
时取得,可见这与定义的群延时相同。
Figure PCTCN2020102721-appb-000013
由于a为实函数,由静态相位法,可略去n>2的项,又:
Figure PCTCN2020102721-appb-000014
其中用到,
Figure PCTCN2020102721-appb-000015
得,不等幅信号频域逆滤波器如下
Figure PCTCN2020102721-appb-000016
因此,在本发明提供的Chirp信号Hammerstein模型系统辨识方法的步骤S10中,具体变幅度Chirp信号x(t)的计算公式为:
Figure PCTCN2020102721-appb-000017
Figure PCTCN2020102721-appb-000018
其中,f 1为Chirp的起始频率,a(t)为幅度,t为时间,即a(t)为随时间变化的频率的函数,T为Chirp信号时长,f 2为Chirp信号截止频率。
进一步地,使用所述变幅度Chirp信号x(t)激励马达,得到第一输出,并对所述第一输出使用傅里叶变换得到频域响应;设定第一输出为y,所述频域响应为Y,则所述频域响应Y由所述第一输出y进行傅里中变化的公式表示为:
Y=fft(y)。
根据所述频域响应使用变幅度Chirp信号x(t)的逆信号频域解析计算马达系统响应函数H(t);所述马达系统响应函数H(t)的计算公式为:
H(t)=Y*X_*1/a(t)
其中,Y为频域响应,X_为变幅度Chirp信号的逆信号频域解析,a(t)为幅度。
使用所述马达系统响应函数H(t)通过转换矩阵转换得到核函数Kernals;所述马达系统响应函数H(t)的计算公式为:
H(t)=Y*X_*1/a(t)
其中,Y为频域响应,X_为变幅度Chirp信号的逆信号频域解析,a(t)为幅度。
其中,所述核函数Kernals的计算公式为:
k i(t)=AH i(t)
其中,k i(t)为核函数Kernals,A为转换矩阵,H i(t)表示第i次谐波响应的马达系统响应,i为自然数。
其中a 0为a(t)的常数幅值,两者之间的关系为:a 0为a(t)的常数幅值,其中,a(t)与a 0之间的公式为:a(t)=a 0·γ(t),为归一化的可变电压曲线。在实际中,幅度a(t)大部分为定值a 0,只有在某些马达容易打壳的频点才降低幅度。
所述核函数的阶数为5时,转换矩阵A表示为:
Figure PCTCN2020102721-appb-000019
由变幅度Chirp信号与所述核函数Kernals得到模型输出y_est;所述模型输出y_est的计算公式为:
Figure PCTCN2020102721-appb-000020
其中,t为时间,x(t)为输入的变幅度Chirp信号,k i(t)为第i阶核函数Kernals,i为自然数。
将所述模型输出y_est与所述第一输出对比得到差异模型误差;设所述差异模型误差为ε t,则其计算公式为:
ε(t)=y(t)-y_est(t)
其中,y为所述第一输出。
请参阅图2,具体在本发明提供的一实施例中,使用线性马达的极限电压测试平台对变幅度Chirp信号进行测试,具体地,所述测试平台包括电脑PC10、马达20、工装30、海绵体40、加速度计50、采集卡80、第一放大器60和第二放大器70,其中,所述PC10与所述采集卡80连接,所述采集卡80使用NI-DAQ 4431采集卡;所述马达20粘性贴合安装在所述工装30上,所述工装30放置安装在所述海绵体40上以避免环境对测量结果的影响,所述马达20为线性谐振传动器(Linear resonance driver,LRA);所述加速度计50安装在所述工装30上以用来测量工装30在马达20振动方向上的加速度,所述加速度计50连接所述第一放大器60,所述第一放大器60连接所述采集卡80,第二放大器70连接采集卡80和马达20;具体地:
在PC10上将生成的数字信号送入到采集卡80进行数模转换成模拟信号,并通过第二放大器70进行放大以激励马达20振动,马达20的振动带动工装30振动,并通过加速度计50采集振动信号并发送给第一放大器60,第一放大器60放大所述振动信号,同时,采集卡80同步采集测量振动方向上的加速度y(n)和激励马达的电压x(n)。
使用所述测试平台进行变幅度Chirp信号的Hammerstein模型系统的辨识测试,其中激励信号的参数如表所示:
表1:
Figure PCTCN2020102721-appb-000021
Figure PCTCN2020102721-appb-000022
请一并参阅图3、图4和图5,分别为使用所述测试平台测试的6V变幅度Chirp信号、2V变幅度Chirp信号和1V变幅度Chirp信号测试结果,其中,线A表示马达实际输出的信号、线B表示模型输出y_est、线C表差异模型误差结果。可以看出,变幅度Chirp信号计算的差异模型误差较小,很容易实现对马达Hammerstein模型系统的辨识,大大缩短了马达Hammerstein模型系统的辨识时间,较大地提升了马达模型的辨识效率。
与现有技术相比,本发明提供的Chirp信号Hammerstein模型系统辨识方法,实现和step序列信号一样任意设计序列频率与幅度,利用Chirp信号的优点缩短马达Hammerstein模型系统的辨识时间,较大地提升了马达模型的辨识效率,进一步提升建模对马达系统的设计,提高了用户触觉反馈的体验效果。
以上所述的仅是本发明的实施方式,在此应当指出,对于本领域的普通技术人员来说,在不脱离本发明创造构思的前提下,还可以做出改进,但这些均属于本发明的保护范围。

Claims (8)

  1. 一种Chirp信号Hammerstein模型系统辨识方法,其特征在于,包括:
    步骤S10:设计并输出变幅度Chirp信号x(t);
    步骤S20:使用所述变幅度Chirp信号x(t)激励马达,得到第一输出,并对所述第一输出使用傅里叶变换得到频域响应;
    步骤S30:根据所述频域响应使用变幅度Chirp信号x(t)的逆信号频域解析计算马达系统响应函数H(t);
    步骤S40:使用所述马达系统响应函数H(t)通过转换矩阵转换得到核函数Kernals;
    步骤S50:由变幅度Chirp信号x(t)与所述核函数Kernals得到模型输出y_est;
    步骤S60:将所述模型输出y_est与所述第一输出对比得到差异模型误差。
  2. 根据权利要求1所述的Chirp信号Hammerstein模型系统辨识方法,其特征在于,所述变幅度Chirp信号x(t)通过以下公式计算得到:
    Figure PCTCN2020102721-appb-100001
    Figure PCTCN2020102721-appb-100002
    其中,f 1为Chirp的起始频率,a(t)为幅度,t为时间,即a(t)为随时间变化的频率的函数,T为Chirp信号时长,f 2为Chirp信号截止频率。
  3. 根据权利要求1所述的Chirp信号Hammerstein模型系统辨识方法,其特征在于,设定第一输出为y,所述频域响应为Y,则所述频域响应Y由所述第一输出y进行傅里中变化的公式表示为:
    Y=fft(y)。
  4. 根据权利要求1所述的Chirp信号Hammerstein模型系统辨识方法,其特征在于,所述马达系统响应函数H(t)的计算公式为:
    H(t)=Y*X_*1/a(t)
    其中,Y为频域响应,X_为变幅度Chirp信号的逆信号频域解析,a(t)为幅度。
  5. 根据权利要求4所述的Chirp信号Hammerstein模型系统辨识方法,其特征在于,所述核函数Kernals的计算公式为:
    k i(t)=AH i(t)
    其中,k i(t)为核函数Kernals,A为转换矩阵,H i(t)表示第i次谐波响应的马达系统响应,i为自然数。
  6. 根据权利要求1所述的Chirp信号Hammerstein模型系统辨识方法,其特征在于,所述模型输出y_est的计算公式为:
    Figure PCTCN2020102721-appb-100003
    其中,t为时间,x(t)为输入的变幅度Chirp信号,k i(t)为第i阶核函数Kernals,i为自然数。
  7. 根据权利要求5所述的Chirp信号Hammerstein模型系统辨识方法,其特征在于,所述核函数的阶数为5时,转换矩阵A表示为:
    Figure PCTCN2020102721-appb-100004
    a 0为a(t)的常数幅值,其中,a(t)与a0之间的公式为:a(t)=a 0·γ(t),为归一化的可变电压曲线。
  8. 根据权利要求1所述的Chirp信号Hammerstein模型系统辨识方法,其特征在于,设所述差异模型误差为ε t,则其计算公式为:
    ε(t)=y(t)-y_est(t)
    其中,y为所述第一输出。
PCT/CN2020/102721 2020-07-10 2020-07-17 Chirp信号Hammerstein模型系统辨识方法 WO2022006980A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010664781.X 2020-07-10
CN202010664781.XA CN111880092B (zh) 2020-07-10 2020-07-10 Chirp信号Hammerstein模型系统辨识方法

Publications (1)

Publication Number Publication Date
WO2022006980A1 true WO2022006980A1 (zh) 2022-01-13

Family

ID=73150318

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/102721 WO2022006980A1 (zh) 2020-07-10 2020-07-17 Chirp信号Hammerstein模型系统辨识方法

Country Status (2)

Country Link
CN (1) CN111880092B (zh)
WO (1) WO2022006980A1 (zh)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2008061470A (ja) * 2006-09-04 2008-03-13 Yaskawa Electric Corp 振動検出装置およびそれを備えたモータ制御装置
CN104007317A (zh) * 2014-06-04 2014-08-27 哈尔滨工业大学 一种获取伺服系统频率特性的方法及装置
CN105425587A (zh) * 2015-11-16 2016-03-23 北京理工大学 迟滞非线性电机辨识与控制方法
CN110907827A (zh) * 2019-11-22 2020-03-24 瑞声科技(新加坡)有限公司 一种马达瞬态失真测量方法及系统

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101600144B (zh) * 2008-11-27 2013-05-15 嘉兴中科声学科技有限公司 采用连续对数扫频信号获得电声产品多个参数的方法及系统
CN102118678B (zh) * 2011-04-02 2014-07-23 嘉兴中科声学科技有限公司 一种使用电流传感器测量扬声器参数的方法及系统
CN106533285B (zh) * 2016-09-30 2019-08-06 郑州轻工业学院 基于Kriging模型的永磁直流电机转速控制方法
CN109901066A (zh) * 2018-12-31 2019-06-18 瑞声科技(新加坡)有限公司 马达系统辨识方法
CN111106783B (zh) * 2019-12-18 2024-05-17 瑞声科技(新加坡)有限公司 一种信号制作方法、信号制作装置、振动马达及触屏设备

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2008061470A (ja) * 2006-09-04 2008-03-13 Yaskawa Electric Corp 振動検出装置およびそれを備えたモータ制御装置
CN104007317A (zh) * 2014-06-04 2014-08-27 哈尔滨工业大学 一种获取伺服系统频率特性的方法及装置
CN105425587A (zh) * 2015-11-16 2016-03-23 北京理工大学 迟滞非线性电机辨识与控制方法
CN110907827A (zh) * 2019-11-22 2020-03-24 瑞声科技(新加坡)有限公司 一种马达瞬态失真测量方法及系统

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
BURRASCANO PIETRO; LAURETI STEFANO; SENNI LUCA; SILIPIGNI GIUSEPPE; TOMASELLO RICCARDO; RICCI MARCO: "Chirp design in a pulse compression procedure for the identification of non-linear systems", 2017 14TH INTERNATIONAL CONFERENCE ON SYNTHESIS, MODELING, ANALYSIS AND SIMULATION METHODS AND APPLICATIONS TO CIRCUIT DESIGN (SMACD), IEEE, 12 June 2017 (2017-06-12), pages 1 - 4, XP033121638, DOI: 10.1109/SMACD.2017.7981565 *

Also Published As

Publication number Publication date
CN111880092A (zh) 2020-11-03
CN111880092B (zh) 2023-06-27

Similar Documents

Publication Publication Date Title
US20210025940A1 (en) Method And Apparatus For Testing Nonlinear Parameter of Motor
CN110146810B (zh) 一种线性马达测试参数的确定方法及装置
CN111551848B (zh) 马达体验失真指标的测试方法、电子设备及存储介质
WO2020211105A1 (zh) 一种马达驱动信号设置方法、电子设备及存储介质
CN111106783B (zh) 一种信号制作方法、信号制作装置、振动马达及触屏设备
CN111478630B (zh) 一种马达稳态单频失真补偿方法及装置
CN110907827B (zh) 一种马达瞬态失真测量方法及系统
CN112506341B (zh) 一种振动效果的生成方法、装置、终端设备及存储介质
WO2021000178A1 (zh) 马达激励信号生成方法、装置和计算机设备
WO2022000600A1 (zh) 马达参数测试方法
WO2022006980A1 (zh) Chirp信号Hammerstein模型系统辨识方法
WO2022000649A1 (zh) 马达均衡电信号的修正方法及设备、计算机可读存储介质
CN105222973A (zh) 用于振动信号阶次跟踪的现场校准方法
CN112491323B (zh) 线性马达超行程控制方法、装置、计算机设备及存储介质
CN110031229A (zh) 一种马达带宽的测量方法及装置
CN109062036B (zh) 基于传递函数的振动谐波迭代控制系统
CN111722108A (zh) 马达失真测量方法及设备、计算机可读存储介质
CN116124279A (zh) 一种确定谐振频率的方法、装置、设备及存储介质
WO2022110502A1 (zh) 车载触控装置的线性参数的测试方法及装置
US20140009990A1 (en) Method and apparatus for characterizing power supply impedance for power delivery networks
Benacchio et al. Active control of string instruments using xenomai
WO2022006788A1 (zh) 马达振动位移估测方法、装置及介质
WO2022000656A1 (zh) 马达系统失真的测量方法及设备、计算机可读存储介质
US20210042519A1 (en) Method for evaluating vibrating sensation similarity, apparatus and storage medium
CN111539089A (zh) 马达非线性模型判断方法和系统

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20944247

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20944247

Country of ref document: EP

Kind code of ref document: A1