CN109190207A - A kind of radio-frequency power amplifier temperature performance prediction technique based on ELM - Google Patents
A kind of radio-frequency power amplifier temperature performance prediction technique based on ELM Download PDFInfo
- Publication number
- CN109190207A CN109190207A CN201810935887.1A CN201810935887A CN109190207A CN 109190207 A CN109190207 A CN 109190207A CN 201810935887 A CN201810935887 A CN 201810935887A CN 109190207 A CN109190207 A CN 109190207A
- Authority
- CN
- China
- Prior art keywords
- elm
- parameter
- frequency power
- power amplifier
- radio
- Prior art date
- Legal status (The legal status 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 status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/30—Circuit design
- G06F30/36—Circuit design at the analogue level
Abstract
The radio-frequency power amplifier temperature performance prediction technique based on ELM that the invention discloses a kind of: using simulation software, obtains radio-frequency power amplifier S parameter simulation value；Test experiments platform is built, actual test environment is consistent with simulated environment, obtains radio-frequency power amplifier S parameter measured value；S parameter simulation value and S parameter measured value form modeling sample, it is divided to two groups, for the training data and test data of ELM model, using S parameter simulation value in training data as ELM mode input, S parameter measured value is exported as ELM model in training data, to ELM model training, training reaches deconditioning after setting error range；S parameter simulation value importing in test data has been trained in ELM model, MSE is calculated according to output result；If MSE is less than desired value, ELM model training is completed；Otherwise re -training.The present invention quickly comprehensively accurate can obtain radio-frequency power amplifier performance parameter in wide temperature range.
Description
Technical field
The present invention relates to radio-frequency power amplifier temperature performances to predict field, and more specifically, it relates to one kind to be based on ELM
Radio-frequency power amplifier temperature performance prediction technique.
Background technique
Radio-frequency power amplifier is the active module embedded therein of radio frequency transmitter system front end, and the quality of performance directly determines
Whole system power-performance.Since the device of composition radio-frequency power amplifier is very sensitive to temperature, so that radio-frequency power amplifies
Device performance is very sensitive to temperature change.By literature survey, discovery: existing device temperature model is not perfect enough, causes to set
There are deviations for the temperature performance simulation result for the radio-frequency power amplifier that meter comes out and experiment test.When radio-frequency power amplifier work
When changing as environment temperature, performance can also change, so that the performance of entire transmitter system changes.Cause
This, before application radio-frequency power amplifier, obtaining the accurate temperature performance of radio-frequency power amplifier is very necessary thing.
Tradition obtains there are two ways to radio-frequency power amplifier temperature performance: (1) testing at different ambient temperatures；
(2) actual test temperature is simulated in business software, simulates the temperature performance of radio-frequency power amplifier.Temperature performance test side
The precision of method is high, still, due in test process, needing that data could be recorded after waiting radio-frequency power amplifier thermostabilization, makes
The testing time for obtaining wide temperature range is at high cost, therefore during actual test, only tests individual temperature spot performances and penetrate to characterize
The temperature performance of frequency power amplifier, and then lose the performance of many temperature spots.Method speed based on commercial simulation software is fast,
But since device model limits, lead to low precision.Therefore, the temperature of radio-frequency power amplifier how is quick and precisely obtained comprehensively
The problem of performance is urgent need to resolve.
Summary of the invention
The purpose of the present invention is to solve deficiencies existing for radio-frequency power amplifier temperature performance acquisition methods, provide one
Radio-frequency power amplifier temperature performance prediction technique of the kind based on ELM, based on radio-frequency power amplifier in wide temperature range
Simulation performance parameter and individual temperature spot test data of experiment, utilize the mapping of Extreme Learning Machine (ELM)
Ability and generalization ability establish radio-frequency power amplifier performance parameter (S parameter) emulation testing mapping relations model one by one, can be fast
Radio-frequency power amplifier performance parameter in fast comprehensively accurate acquisition wide temperature range.
The purpose of the present invention is what is be achieved through the following technical solutions.
Radio-frequency power amplifier temperature performance prediction technique based on ELM of the invention, comprising the following steps:
Step 1, using simulation software, the S parameter of radio-frequency power amplifier is emulated in acquisition certain temperature, frequency range
Value；
Step 2 builds test experiments platform, and actual test environment and simulated environment setting are consistent, obtain radio-frequency power and put
The S parameter measured value of big device；
The S parameter measured value composition modeling sample that step 3, the S parameter simulation value obtained according to step 1 and step 2 obtain
Modeling sample is divided into two groups, the respectively training data and test data of ELM model, using the S parameter in training data by product
Input of the simulation value as ELM model, output of the S parameter measured value as ELM model in training data, to ELM model into
Deconditioning after the error range for reaching setting is trained in row training；
Step 4 imports the S parameter simulation value in test data in trained ELM model, according to output result
Calculate test error MSE；
Step 5, if MSE is less than desired value, ELM model training is completed；If MSE is greater than desired value, adjust
ELM model parameter, re -training, until MSE is less than desired value, modeling is completed in end training.
Test error MSE is calculated as follows in step 4:
Wherein, tiIndicate experiment test performance parameter, i.e., the S parameter measured value in test data, oiIndicate ELM model
Reality output is as a result, N indicates sample size.
Compared with prior art, the beneficial effects brought by the technical solution of the present invention are as follows:
ELM is applied in the prediction of radio-frequency power amplifier temperature performance by the present invention for the first time, by business software temperature performance
Simulation result is combined with experimental results, using the data generaliza-tion ability of ELM, can be predicted the property of non-test temperature point
Energy.Radio-frequency power amplifier temperature performance method is obtained compared to tradition, is had fast and accurately based on the prediction model that ELM is established
Advantage.
Detailed description of the invention
Fig. 1 is the modeling procedure schematic diagram based on ELM；
Fig. 2 is the actual measurement of radio-frequency power amplifier performance parameter and simulation result comparison diagram；
Fig. 3 is the actual measurement of radio-frequency power amplifier amplifier performance parameter and prediction result comparison diagram.
Specific embodiment
The invention will be further described with reference to the accompanying drawing.
Radio-frequency power amplifier temperature performance prediction technique based on ELM of the invention, as shown in Figure 1, including following step
It is rapid:
Step 1, using simulation software, the S parameter of radio-frequency power amplifier is emulated in acquisition certain temperature, frequency range
Value；
Step 2 builds test experiments platform, and actual test environment and simulated environment setting are consistent, obtain radio-frequency power and put
The S parameter measured value of big device；
The S parameter measured value composition modeling sample that step 3, the S parameter simulation value obtained according to step 1 and step 2 obtain
Modeling sample is divided into two groups, the respectively training data and test data of ELM model, using the S parameter in training data by product
Input of the simulation value as ELM model, output of the S parameter measured value as ELM model in training data, to ELM model into
Deconditioning after the error range for reaching setting is trained in row training；
Step 4 imports the S parameter simulation value in test data in trained ELM model, according to output result
Calculate test error MSE；
Test error MSE is calculated as follows:
Wherein, tiIndicate experiment test performance parameter, i.e., the S parameter measured value in test data, oiIndicate ELM model
Reality output is as a result, N indicates sample size.
Step 5, if MSE is less than desired value, ELM model training is completed；If MSE is greater than desired value, adjust
ELM model parameter (excitation function G and hidden neuron number N), re -training, until MSE is less than desired value, end is trained,
Complete modeling.
Embodiment:
Carry out the amplification of the radio-frequency power to proposed by the present invention based on ELM now in conjunction with 500-2100MHz radio-frequency power amplifiers
Device temperature performance prediction technique is described in detail.
Step 1 obtains the S parameter simulation value of radio-frequency power amplifier using ADS simulation software.Frequency range is set as
500-2100MHz, frequency interval 20MHz, temperature range cover-40-90 DEG C, 5 DEG C of temperature interval.Simulation result is as shown in Figure 2.
Step 2 builds test experiments platform using vector network analyzer and temperature device.Platform Instrumental is arranged to join
Number, keeps actual test environment consistent with simulated environment.Using vector network analyzer test radio-frequency power amplifier S parameter
Measured value, test results are shown in figure 2.
The S parameter measured value composition modeling sample that step 3, the S parameter simulation value obtained according to step 1 and step 2 obtain
Product amount to 2187 groups of samples.Sample is divided into 2 groups, respectively as the training data and test data of ELM model, wherein training
1134 groups of sample, 1053 groups of test sample.Using input of the S parameter simulation value as ELM model in training data, training number
Output of the S parameter measured value as ELM model in, is trained ELM model, after training reaches the error range of setting
Deconditioning.；
Step 4 imports the S parameter simulation value in test data in trained ELM model, is counted according to formula (1)
MSE is calculated, herein N=1053；
Step 5 compares MSE and desired value (10-2) size, if MSE be less than desired value (10-2), then complete ELM mould
Type training；If MSE is greater than desired value (10-2), then adjusting parameter (excitation function G and neuron number L), re -training, directly
When being less than desired value to MSE, terminating training, (in the present embodiment, excitation function chooses sigmoid, and hidden neuron number is
50, test error MSE=8.945 × 10-3)；
Finally, drawing the radio-frequency power amplifier temperature performance prediction based on ELM, as shown in Figure 3.Fig. 3 can be seen that
The prediction result and experimental results of ELM model are coincide substantially.Compared to simulation result in Fig. 2, it is pre- to show ELM temperature performance
Survey the advantage of precision.Illustrate, which can quick and precisely obtain the temperature performance of radio-frequency power amplifier comprehensively.
In conclusion obtaining wide temperature range temperature characteristic data and a small number of temperature spots using commercial simulation software tests number
According to, and ELM model is applied, mapping relations are established, the performance of radio-frequency power amplifier in wide temperature range can be predicted.This pole
The time that experiment test obtains wide temperature range radio-frequency power amplifier performance is shortened greatly, is the wide temperature of radio-frequency power amplifier
Range applications provide guidance.
Although function and the course of work of the invention are described above in conjunction with attached drawing, the invention is not limited to
Above-mentioned concrete function and the course of work, the above mentioned embodiment is only schematical, rather than restrictive, ability
The those of ordinary skill in domain under the inspiration of the present invention, is not departing from present inventive concept and scope of the claimed protection situation
Under, many forms can be also made, all of these belong to the protection of the present invention.
Claims (2)
1. a kind of radio-frequency power amplifier temperature performance prediction technique based on ELM, which comprises the following steps:
Step 1, using simulation software, obtain certain temperature, in frequency range radio-frequency power amplifier S parameter simulation value；
Step 2 builds test experiments platform, and actual test environment and simulated environment setting are consistent, obtain radio-frequency power amplifier
S parameter measured value；
The S parameter measured value that step 3, the S parameter simulation value obtained according to step 1 and step 2 obtain forms modeling sample,
Modeling sample is divided into two groups, the respectively training data and test data of ELM model, it is imitative using the S parameter in training data
Input of the true value as ELM model, output of the S parameter measured value as ELM model in training data, carries out ELM model
Deconditioning after the error range for reaching setting is trained in training；
Step 4 imports the S parameter simulation value in test data in trained ELM model, is calculated according to output result
Test error MSE；
Step 5, if MSE is less than desired value, ELM model training is completed；If MSE is greater than desired value, ELM mould is adjusted
Shape parameter, re -training, until MSE is less than desired value, modeling is completed in end training.
2. the radio-frequency power amplifier temperature performance prediction technique according to claim 1 based on ELM, which is characterized in that
Test error MSE is calculated as follows in step 4:
Wherein, tiIndicate experiment test performance parameter, i.e., the S parameter measured value in test data, oiIndicate the reality of ELM model
Output is as a result, N indicates sample size.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810935887.1A CN109190207A (en) | 2018-08-16 | 2018-08-16 | A kind of radio-frequency power amplifier temperature performance prediction technique based on ELM |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810935887.1A CN109190207A (en) | 2018-08-16 | 2018-08-16 | A kind of radio-frequency power amplifier temperature performance prediction technique based on ELM |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109190207A true CN109190207A (en) | 2019-01-11 |
Family
ID=64918483
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810935887.1A Pending CN109190207A (en) | 2018-08-16 | 2018-08-16 | A kind of radio-frequency power amplifier temperature performance prediction technique based on ELM |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109190207A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111273110A (en) * | 2020-03-19 | 2020-06-12 | 中国振华集团云科电子有限公司 | Full-temperature automatic testing system and method for S parameter of microwave component |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140067348A1 (en) * | 2012-08-28 | 2014-03-06 | Taiwan Semiconductor Manufacturing Company Limited | Array modeling for one or more analog devices |
| CN104463209A (en) * | 2014-12-08 | 2015-03-25 | 厦门理工学院 | Method for recognizing digital code on PCB based on BP neural network |
| CN104517035A (en) * | 2014-12-22 | 2015-04-15 | 中国舰船研究设计中心 | Planar array antenna active scattering directional diagram predication method |
| CN107609286A (en) * | 2017-09-20 | 2018-01-19 | 北京安怀信科技股份有限公司 | A kind of mathematical modeling tests die worker's tool |
| CN107977535A (en) * | 2017-12-22 | 2018-05-01 | 天津大学 | A kind of RF power amplification temperature characterisitic modeling method based on ELM |
| CN108154259A (en) * | 2017-11-27 | 2018-06-12 | 国网北京市电力公司 | Load forecasting method and device, storage medium, the processor of heat pump |
-
2018
- 2018-08-16 CN CN201810935887.1A patent/CN109190207A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140067348A1 (en) * | 2012-08-28 | 2014-03-06 | Taiwan Semiconductor Manufacturing Company Limited | Array modeling for one or more analog devices |
| CN104463209A (en) * | 2014-12-08 | 2015-03-25 | 厦门理工学院 | Method for recognizing digital code on PCB based on BP neural network |
| CN104517035A (en) * | 2014-12-22 | 2015-04-15 | 中国舰船研究设计中心 | Planar array antenna active scattering directional diagram predication method |
| CN107609286A (en) * | 2017-09-20 | 2018-01-19 | 北京安怀信科技股份有限公司 | A kind of mathematical modeling tests die worker's tool |
| CN108154259A (en) * | 2017-11-27 | 2018-06-12 | 国网北京市电力公司 | Load forecasting method and device, storage medium, the processor of heat pump |
| CN107977535A (en) * | 2017-12-22 | 2018-05-01 | 天津大学 | A kind of RF power amplification temperature characterisitic modeling method based on ELM |
Non-Patent Citations (2)
| Title |
|---|
| CHENG-YU ZHANG: "Extreme Learning Machine for the Behavioral Modeling of RF Power Amplifiers", 《2017 IEEE MTT-S INTERNATIONAL MICROWAVE SYMPOSIUM (IMS)》 * |
| JIXIAO0526: "神经网络做误差校正", 《HTTP://MUCHONG.COM/T-5599651-1-PID-2》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111273110A (en) * | 2020-03-19 | 2020-06-12 | 中国振华集团云科电子有限公司 | Full-temperature automatic testing system and method for S parameter of microwave component |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105206888A (en) | Lithium ion battery internal temperature monitoring method | |
| CN104539250B (en) | Adjustment method, the device and system of power amplifier | |
| CN102970180B (en) | Real-time simulation method of communication delay of wide area measurement system of electric power system | |
| CN102128990A (en) | Method for judging power amplifier saturation in electromagnetic radiation sensitivity automatic test | |
| CN110320453A (en) | The method and system influenced on GIS partial discharge are leaked based on software emulation analysis gas | |
| CN104965941A (en) | Magnetic flux leakage testing defect reconstruction method based on improved artificial bee colony algorithm | |
| CN106679694A (en) | TACAN beacon simulator air-to-air response delayed measurement precision calibration apparatus and method | |
| CN112492497B (en) | T/S parameter measuring method and device for loudspeaker and electronic equipment | |
| CN104122031A (en) | Silicon pressure sensor temperature compensation method based on extreme learning machine | |
| CN103543026B (en) | Method for identifying structural damage based on vibration transmissibility function and support vector machine | |
| CN105548917A (en) | Calibration method of magnetic sensor array in non-shielding environment | |
| CN107977535A (en) | A kind of RF power amplification temperature characterisitic modeling method based on ELM | |
| CN109839627B (en) | Sonar detection distance measurement and calibration system and method under pool condition | |
| CN109190207A (en) | A kind of radio-frequency power amplifier temperature performance prediction technique based on ELM | |
| CN109635519B (en) | Microwave photonic device modeling method based on electromagnetic field and temperature field coupling | |
| CN101487787A (en) | Method for fast confirming hydraulic parameters of unsaturated soil | |
| US7680615B2 (en) | Parallel testing system with shared golden calibration table and method thereof | |
| CN112285435B (en) | Equivalent simulation method of high-power magnetic field radiation source | |
| CN108446451A (en) | A kind of low-noise amplifier temperature characterisitic modeling method based on ELM | |
| CN103308928B (en) | Pseudo-range precision measurement system of satellite navigation signal simulator | |
| CN107621607A (en) | Battery performance test method, test terminal, readable storage medium storing program for executing and test system | |
| CN107566058A (en) | A kind of 5G communication damping matrix calibrating installations and method | |
| CN115308552A (en) | Method and system for calculating wave head and wave tail resistance values of impulse voltage test device | |
| CN103308930B (en) | Pseudo-range precision measurement method of satellite navigation signal simulator | |
| CN110702344A (en) | Closed-loop bridge structure damage diagnosis method based on distributed strain measurement technology |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190111 |
|
| WD01 | Invention patent application deemed withdrawn after publication |