CN112329371B - IALO-BP neural network-based reverse modeling method for Doherty power amplifier - Google Patents
IALO-BP neural network-based reverse modeling method for Doherty power amplifier Download PDFInfo
- Publication number
- CN112329371B CN112329371B CN202011233431.4A CN202011233431A CN112329371B CN 112329371 B CN112329371 B CN 112329371B CN 202011233431 A CN202011233431 A CN 202011233431A CN 112329371 B CN112329371 B CN 112329371B
- Authority
- CN
- China
- Prior art keywords
- ant
- algorithm
- lion
- ialo
- neural network
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 37
- 241001206881 Myrmeleon inconspicuus Species 0.000 claims abstract description 56
- 238000012549 training Methods 0.000 claims abstract description 16
- 238000011156 evaluation Methods 0.000 claims abstract description 9
- 238000012360 testing method Methods 0.000 claims abstract description 6
- 230000008569 process Effects 0.000 claims abstract description 5
- 241000257303 Hymenoptera Species 0.000 claims description 20
- 230000006872 improvement Effects 0.000 claims description 12
- 238000011161 development Methods 0.000 claims description 8
- 238000004088 simulation Methods 0.000 claims description 8
- 230000008602 contraction Effects 0.000 claims description 6
- 238000013507 mapping Methods 0.000 claims description 6
- 241000258923 Neuroptera Species 0.000 claims description 4
- 238000004458 analytical method Methods 0.000 claims description 3
- 230000008859 change Effects 0.000 claims description 3
- 230000003321 amplification Effects 0.000 claims description 2
- 230000000739 chaotic effect Effects 0.000 claims description 2
- 238000003199 nucleic acid amplification method Methods 0.000 claims description 2
- 238000005457 optimization Methods 0.000 claims description 2
- 230000001105 regulatory effect Effects 0.000 claims description 2
- 230000006870 function Effects 0.000 description 13
- 238000011160 research Methods 0.000 description 5
- 238000013461 design Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- BTCSSZJGUNDROE-UHFFFAOYSA-N gamma-aminobutyric acid Chemical compound NCCCC(O)=O BTCSSZJGUNDROE-UHFFFAOYSA-N 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- 230000001419 dependent effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
Classifications
-
- 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
- G06F30/367—Design verification, e.g. using simulation, simulation program with integrated circuit emphasis [SPICE], direct methods or relaxation methods
-
- 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
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
Landscapes
- Engineering & Computer Science (AREA)
- Computer Hardware Design (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Microelectronics & Electronic Packaging (AREA)
- Amplifiers (AREA)
Abstract
The invention discloses a reverse modeling method of a Doherty power amplifier based on IALO-BP neural network, which mainly comprises the following steps: obtaining experimental data, and dividing the experimental data into a training set and a testing set; the traditional ant lion algorithm is improved; training a BP forward model by using an improved ant-lion algorithm, and storing the optimized weight; the weight is kept unchanged, and electric parameters are input into the established IALO-BP model, so that an inverse solving process is realized; calculating an evaluation function between the output parameter and the target parameter; and updating the input parameters by using a reverse iterative algorithm. According to the invention, a IALO-BP neural network reverse modeling method is applied to the Doherty power amplifier, corresponding frequency f and output power are solved according to return loss S11 and efficiency of an output end, and better structural parameters are obtained.
Description
Technical Field
The invention belongs to a neural network reverse modeling research method of a radio frequency microwave device, and particularly relates to a Doherty power amplifier reverse modeling method based on IALO-BP neural network.
Background
In the twentieth century of high-speed development, wireless communication has played an increasingly important role in people's life. Wireless communication systems have evolved rapidly in recent years and have no tendency to slow down at all. The power amplifier plays an important role in amplifying signals in a transmitting system and is a core device in a transmitter module, so that the performance of the transmitter is basically dependent on the performance of the power amplifier.
Power amplifiers often need to operate over a relatively wide frequency band and as the complexity of the modulated signal increases, the power amplifier needs to maintain a high efficiency over a large output power back-off range. In recent years, the Doherty power amplifier has been a research hotspot of students because the Doherty power amplifier can maintain high efficiency at the peak-to-average ratio of OFDM signals. The common power amplifier design and analysis method adopts the auxiliary design of ADS simulation software. In order to obtain an optimal structure, repeated simulation is needed, manpower and time are wasted, and accuracy cannot be guaranteed, so that a neural network reverse modeling research is provided, and the method is an effective alternative method for modeling a microwave device. Once the model is built, it can be reused, thus avoiding the need for a complete electromagnetic structure to re-simulate the repeated electromagnetic simulation process with a small change in the physical dimensions of the device. The artificial neural network has the advantages of high speed and high precision, and is widely applied to the field of radio frequency microwaves.
Disclosure of Invention
In order to solve the technical problems, the technical problem to be solved by the invention is to provide a Doherty power amplifier reverse modeling method based on IALO-BP neural network, optimize the weight of a self-adaptive BP neural network forward model through IALO algorithm, save the optimal weight, establish a IALO-BP neural network model, apply the reverse neural network to the Doherty power amplifier, and solve the corresponding frequency f and output power according to the return loss S11 and efficiency of the output end to obtain better structural parameters.
The invention is realized by the following technical scheme: the invention relates to a reverse modeling method of a Doherty power amplifier based on IALO-BP neural network, which mainly comprises the following steps: selecting a sample object; setting basic parameters; extracting training set data; improving an ant lion algorithm; training a BP forward model by using an improved ant lion algorithm, optimizing weights and thresholds, and storing; the threshold value of the weight is kept unchanged, return loss S11 and efficiency are input into the established IALO-BP reverse model, and corresponding output parameter frequency f and output power are obtained through operation; calculating an evaluation function F between the output parameter and the target parameter; and updating the input parameters by using a reverse iterative algorithm. Thus, a IALO-BP neural network inverse model is established.
In the invention, selecting a sample object means selecting a Doherty power amplifier, carrying out modeling simulation analysis on radio frequency characteristics of the power amplifier by using ADS simulation software, extracting data of output power and efficiency and return loss S11 and f from the carrier power amplifier, extracting 3000 groups of data from each parameter, respectively selecting 1000 groups of data from the data as training data, and respectively taking 150 groups of data as test data.
In the present invention, the setting of the basic parameters is: the BP neural network input layer is provided with 1 node, the hidden layer is provided with 27 nodes, and the output layer is provided with 1 node. The hidden layer node transfer function selects logsig functions and the output layer node selects purelin functions. Parameter setting of ALO algorithm: the maximum iteration number is 500, the population scale is 50, and the root mean square error of the network training is taken as the fitness value.
In the present invention, the improvement of the ant lion algorithm comprises: the method comprises the steps of initializing population improvement, initializing population by adopting cube mapping, generating a series of initial solutions, and preferentially selecting the initial population, wherein the characteristics of ergodic property, randomness and regularity are provided, so that the diversity of algorithms and the ergodic property of searching are effectively improved; the boundary contraction factor I which is rapidly and continuously increased along with iterative evolution of the algorithm is provided, so that a solving space is more comprehensively searched, and the convergence rate of the algorithm is improved; the dynamic weight coefficient based on the iteration number is introduced into ant position update, so that ants are developed in the neighborhood of the optimal area, and the balance capability of global exploration and local development of the algorithm is improved.
Further, the step of improving the ant lion optimizing IALO-BP weight comprises the following steps:
a. initializing ant and ant lion populations by adopting cube mapping;
b. And taking the optimal ant-lion population obtained by IALO algorithm as the initial weight and threshold of the BP neural network.
C. Taking the structural parameters as input, taking the electrical parameters as output, carrying out forward modeling on the constructed IALO-BP neural network, and taking the root mean square error as an fitness function;
d. The weights and threshold parameters of the established positive model are saved and kept unchanged. And selecting a group of new data, taking the electrical parameters as input and the structural parameters as output, and performing reverse modeling of IALO-BP neural network.
E. calculating an evaluation function between the output parameter and the target parameter, i.e. the sum of squares errorAnd updating input parameters/>, by using reverse iterative algorithmN=0, 1,2, …, where x n+1,xn is the input parameter and η is the learning rate.
And finally, using the established ALO-LMBP neural network inverse model in the research of the amplifier with the Doherty power, testing the model by using 150 groups of test data, and comparing the result of the reverse modeling method with the results of the direct reverse modeling method and the BP reverse modeling method.
Compared with a direct reverse modeling method and a BP reverse modeling method, the method replaces a continuous search program optimization method through an iteration process, so that the problem of multiple solutions of the parameters can be solved, the convergence speed and accuracy are improved, and the modeling operation time is shortened.
The foregoing description is only an overview of the present invention, and is intended to be implemented in accordance with the teachings of the present invention, as well as to provide further clarity and understanding of the above and other objects, features and advantages of the present invention, as described in the following detailed description of the preferred embodiments, taken in conjunction with the accompanying drawings.
Drawings
In order to more clearly illustrate the technical solution of the embodiments of the present invention, the drawings of the embodiments will be briefly described below.
FIG. 1 is a reverse modeling flow chart of IALO-BP neural network of the Doherty power amplifier of the invention;
Fig. 2 is an iteration graph of the ant lion algorithm;
FIG. 3 is an iteration graph of the improved ant lion algorithm;
FIG. 4 is a graph of output power versus efficiency for a Doherty power amplifier of the present invention;
FIG. 5 is a graph of frequency versus return loss for a Doherty power amplifier of the invention;
FIG. 6 is a graph showing a comparison of the output power values corresponding to the efficiency and the three modeling methods according to the present invention;
Fig. 7 is a graph of a comparison of frequency values corresponding to return loss for three modeling methods of the present invention.
Detailed Description
The following detailed description of the invention, taken in conjunction with the accompanying drawings, illustrates the principles of the invention by way of example and by way of a further explanation of the principles of the invention, and its features and advantages will be apparent from the detailed description. In the drawings to which reference is made, the same or similar components in different drawings are denoted by the same reference numerals.
As shown in FIG. 1, the invention provides a reverse modeling method of a Doherty power amplifier based on IALO-BP neural network, which mainly comprises the following steps: selecting a sample object; setting basic parameters; extracting training set data; improving an ant lion algorithm; training a BP forward model by using an improved ant lion algorithm, optimizing weights and thresholds, and storing; the threshold value of the weight is kept unchanged, the return loss S11 and the efficiency are input into the established BP model, and the corresponding output parameter frequency f and output power are obtained through operation; calculating an evaluation function F between the output parameter and the target parameter; and updating the input parameters by using a reverse iterative algorithm. Thus, the reverse model of IALO-BP neural network is built.
1. Sample object selection
And constructing a circuit of the Doherty power amplifier by using ADS simulation software, and carrying out modeling simulation on the radio frequency characteristics of the circuit. Fig. 3 and 4 show the relationship between the output power and efficiency of the carrier amplification of the Doherty power amplifier, the output matching frequency f and the return loss S11. As can be seen from fig. 3, the efficiency is up to 65.96% at a power of 44.92 dBm. Fig. 4 shows that at a frequency of 3.5GHz, S11 is-45.20 dB, achieving a good matching effect. Meanwhile, as can be seen from the figure, the same efficiency corresponds to a plurality of output power values, and the same S11 corresponds to a plurality of f values, respectively. And extracting data of a radio frequency characteristic curve of the BP neural network, and respectively introducing the output power and f serving as input sample data into inverse models of the BP neural network and the IALO-BP neural network.
2. Setting of basic parameters
The BP neural network input layer is provided with 1 node, the hidden layer is provided with 27 nodes, and the output layer is provided with 1 node. The output layer has 1 node. The hidden layer node transfer function selects logsig functions and the output layer node selects purelin functions. Parameter setting of IALO algorithm: the maximum iteration number is 500, the population scale is 50, and the root mean square error of the network training is taken as the fitness value.
3. Extracting training set data
Data of output power, efficiency and return loss S11 and f are extracted from the carrier power amplifier, 3000 groups of data are extracted from each parameter, 1000 groups of data are selected as training data respectively, and 150 groups of data are selected as test data respectively.
4. Improvements to ant lion algorithm
The ant lion algorithm is improved in three aspects: the method comprises the steps of initializing population improvement, initializing population by adopting cube mapping, generating a series of initial solutions, and preferentially selecting the initial population, wherein the characteristics of ergodic property, randomness and regularity are provided, so that the diversity of algorithms and the ergodic property of searching are effectively improved; the boundary contraction factor I which is rapidly and continuously increased along with iterative evolution of the algorithm is provided, so that a solving space is more comprehensively searched, and the convergence rate of the algorithm is improved; the dynamic weight coefficient based on the iteration number is introduced into ant position update, so that ants are developed in the neighborhood of the optimal area, and the balance capability of global exploration and local development of the algorithm is improved. Fig. 2 is an iteration graph of the ant-lion algorithm, and fig. 3 is an iteration graph of the improved ant-lion algorithm, and it can be seen from the two graphs that the improved ant-lion algorithm is significantly superior to the basic ant-lion algorithm in convergence speed and ergodic performance.
The improvement of the ant lion algorithm comprises:
(1) Improvements to initialisation populations
In the traditional ant lion optimization algorithm, an initial ant lion ant population is randomly selected, if a plurality of individuals with poor fitness in the initial ant lion population appear in a certain generation in the iteration process, ants which are selected to walk around the ant lion can sink into local extremum to a great extent. Thus, using a cube map to initialize a population, a random initial sequence x ranging between [0,1] is first generated and substituted into the cube map expressionIn the method, a chaotic sequence is generated, wherein x i,j can be the position Ant i,j of ants or the position Antlion i,j of Ant lions, the generated sequence is effective as long as the initial value is not zero, and then the sequence is expressed by/>The mapping returns to the solution space, where,/>Values given to ants or ant-lion individuals after returning to the solving space are constants, ub and lb are upper and lower limits of the solving space, and therefore the diversity of algorithms and the search ergodic performance are effectively improved.
(2) Edge shrink factor improvement
Boundary contraction factor during the phase of ant wander around trap in ALO algorithmW is a speed regulating factor of ants, the change of I shows a discontinuous increasing trend, and the intermittent increase easily leads to the missing of partial areas of the ants, so that the algorithm easily misses the optimal value, the search boundary is uneven and slowly decays, and the convergence speed is inferior to other algorithms;
Aiming at the problems, in order to more comprehensively search the solving space and improve the convergence rate of the algorithm, a boundary contraction factor which increases rapidly and continuously along with iterative evolution of the algorithm is provided, and the ant walking boundary updating mode is defined as: wherein: gamma is the shrinkage adjustment coefficient; lambda is a scale factor; through multiple experimental simulations, the result is optimal when gamma=400 and lambda=20 are selected.
(3) Development mode improvement for ant lion
During the elite phase, ants select ant lions and elite ant lion behavior to update position based on roulette. The known probabilities of selection of ant lions by roulette are: wherein f (x i) is an individual fitness value. Because the elite ant lion has the optimal fitness value, the elite ant lion has high probability of being selected as the roulette to select the elite ant lion, so that ants only walk around the elite ant lion, and the global exploration capacity of the algorithm is reduced, such as/>, for example
For the above problems, the position of ants can be updatedIn the method, the dynamic weight coefficient of the iteration number is introduced, namely
In formula (1), roulette selects ant lionThe weight coefficient l 1 of the ant is larger in the early stage of iteration, so that ants explore more optimal areas in the search space; in the later period, the elite ant lion is adjacent to the optimal area, and the weight coefficient l 2 of the elite ant lion is gradually increased, so that ants are developed in the neighborhood of the optimal area, and the balance capability of the global exploration and the local development of the algorithm is improved.
5. Training IALO-BP forward model, optimizing weight and threshold value and storing
Initializing ant and ant lion populations by adopting cube mapping, and taking the optimal ant lion populations obtained by IALO algorithm as initial weight and threshold of BP neural network. And taking the structural parameter as input, the electrical parameter as output, carrying out forward modeling on the constructed IALO-BP neural network, and taking the root mean square error as a fitness function.
6. Keeping the weight threshold unchanged, and performing reverse modeling
And saving the weight and threshold parameters of the established IALO-BP neural network positive model, and keeping the weight and threshold parameters unchanged. And selecting a group of new data, taking the electrical parameters as input and the structural parameters as output, and performing reverse modeling of IALO-BP neural network.
7. Calculating an evaluation function F, and updating input parameters
Calculating an evaluation function between the output parameter and the target parameterReuse of reverse iterative algorithm/>N=0, 1,2, … updates the input parameters.
The direct reverse modeling method, the BP reverse modeling method and the IALO-BP reverse modeling method are respectively utilized to respectively model and output corresponding parameter frequency f and output power at known return loss S11 and efficiency; the fitting effect of the two modeling methods and the actual output of the ADS design are compared, as well as their mean square error and run time. And simultaneously carrying out neural network simulation on the MATLAB platform to obtain a prediction result. Fig. 6 is a graph of the two modeling methods and the output power corresponding to the efficiency, and fig. 7 is a graph of the two modeling methods and the frequency corresponding to the return loss, and it can be seen from the comparison of fig. 6 and fig. 7 that the difference between the value obtained by direct reverse modeling and the value in ADS is the largest, and the optimal value cannot be given, and cannot be used to describe the characteristics of the amplifier. Since direct reverse modeling is generally used in a one-to-one correspondence of input parameters and output parameters, and the output power and efficiency and S11 and f correspondence are one-to-many, as in fig. 4 and 5, it cannot solve such a complex one-to-many relationship, so the modeling method has a large error. The fitting effect of the value obtained by the IALO-BP reverse modeling method and the value in ADS is much better than that of the direct reverse modeling method, and the method can meet the actual design requirement in terms of precision.
Here, the output power is synthesized under the condition of known efficiency, the corresponding frequency f is synthesized under the condition of known S11, and the modeling time and the mean square error of the two reverse modeling methods are compared, as shown in table 1, by using the direct reverse modeling method and the IALO-BP reverse modeling method, respectively. Compared with the BP reverse modeling method, the running time obtained by the IALO-BP reverse modeling method is respectively reduced by 37.81 percent and 29.69 percent, and the mean square error is respectively reduced by 95.92 percent and 95.55 percent. Therefore, the IALO-BP neural network reverse modeling method not only can improve modeling accuracy, but also can shorten running time of the network.
Table 1 comparison of the performance of two modeling methods
The invention optimizes the weight and the threshold of the BP neural network by utilizing an improved ant lion algorithm, updates the input parameters by utilizing a reverse iterative algorithm, and constructs a research method for reverse modeling of the Doherty power amplifier based on IALO-BP neural network. The method not only can solve the problem of multiple solutions of the required parameters, but also improves the convergence speed and accuracy and shortens the modeling running time.
While the invention has been described with respect to the preferred embodiments, it will be understood that the invention is not limited thereto, but is capable of modification and variation without departing from the spirit of the invention, as will be apparent to those skilled in the art.
Claims (3)
1. A reverse modeling method of a Doherty power amplifier based on IALO-BP neural network is characterized by comprising the following steps:
(1) Training data are extracted, and data of output power and efficiency, return loss S11 and frequency f of the training data are extracted from a Doherty power amplifier circuit in ADS simulation software to serve as input samples of a model;
(2) The traditional ant lion algorithm is improved, and the three directions are respectively improved: improving the initialized population; improving the boundary shrinkage factor; improving the development mode of ant lion;
(3) Training a BP forward model by using an improved ant-lion algorithm, and storing the optimized weight and threshold;
(4) The threshold value of the weight is kept unchanged, return loss S11 and efficiency are input into the established IALO-BP model, and corresponding frequency f and output power are solved;
(5) Calculating an evaluation function F between the output parameter and the target parameter;
(6) Updating the input parameters by using a reverse iterative algorithm;
the improvement of the ant lion algorithm comprises:
(1) Improvements to initialisation populations
In the traditional ant lion optimization algorithm, an initial ant lion ant population is randomly selected, if a plurality of individuals with poor fitness in the initial ant lion population appear in a certain generation in the iteration process, ants which are selected to walk around the ant lion can sink into local extremum to a great extent; thus, using a cube map to initialize a population, a random initial sequence x ranging between [0,1] is first generated and substituted into the cube map expressionIn the method, a chaotic sequence is generated, wherein x i,j is the Ant position Ant i,j or the Ant lion position Antlion i,j, the initial value is not zero, the generated sequence is effective, and then the sequence is expressed by/>The mapping returns to the solution space, where,/>Values given to ants or ant-lion individuals after returning to the solving space, wherein ub and lb are constants and are the upper limit and the lower limit of the solving space, so that the diversity of the algorithm and the search ergodic property are effectively improved;
(2) Edge shrink factor improvement
Boundary contraction factor during the phase of ant wander around trap in ALO algorithmW is a speed regulating factor of ants, the change of I shows a discontinuous increasing trend, and the intermittent increase easily leads to the missing of partial areas of the ants, so that the algorithm easily misses the optimal value, the search boundary is uneven and slowly decays, and the convergence speed is inferior to other algorithms;
in order to more comprehensively search the solving space and improve the convergence rate of the algorithm, a boundary contraction factor which increases rapidly and continuously along with iterative evolution of the algorithm is provided, and an ant walking boundary updating mode is defined as follows: Wherein: gamma is the shrinkage adjustment coefficient; lambda is a scale factor; through multiple experimental simulations, selecting gamma=400 and lambda=20;
(3) Development mode improvement for ant lion
In elite stage, ants select ant lion and elite ant lion behavior to update position according to roulette; the known probabilities of selection of ant lions by roulette are: Wherein f (x i) is an individual fitness value; because the elite ant lion has the optimal fitness value, the elite ant lion has high probability of being selected as the roulette to select the elite ant lion, so that ants only walk around the elite ant lion, and the global exploration capacity of the algorithm is reduced, such as/>, for example
Updated at ant positionIn the method, the dynamic weight coefficient of the iteration number is introduced, namely
In formula (1), roulette selects ant lionThe weight coefficient l 1 of the ant is larger in the early stage of iteration, so that ants explore more optimal areas in the search space; in the later period, the elite ant lion is adjacent to the optimal area, and the weight coefficient l 2 of the elite ant lion is gradually increased, so that ants are developed in the neighborhood of the optimal area, and the balance capability of the global exploration and the local development of the algorithm is improved.
2. The reverse modeling method of the Doherty power amplifier based on IALO-BP neural network as claimed in claim 1, wherein in the step (1), modeling simulation analysis of radio frequency characteristics is carried out on the Doherty power amplifier by using ADS simulation software, when the resonance of a power amplification circuit is at a central frequency, the circuit meets the parameter indexes of return loss S11 < -10dB, output power P of the whole circuit is less than or equal to 10W, power addition efficiency PAE is more than or equal to 30%, f takes a value of 3.0-3.7 GHz, data of output power and efficiency and return loss S11 and f of an output matching end are extracted from a carrier power amplifier, 3000 groups of data are extracted from each parameter, 1000 groups of data are selected as training data respectively, and 150 groups of data are taken as test data respectively.
3. The reverse modeling method of Doherty power amplifier based on IALO-BP neural network as defined in claim 1, wherein the evaluation function between the output parameter and the target parameter in the step (5)Wherein Y is the output of the LMBP neural network forward model, and T is a known target electrical parameter; the evaluation function is to find the square sum error, and E p represents the square sum error obtained by the p-th group of data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011233431.4A CN112329371B (en) | 2020-11-06 | 2020-11-06 | IALO-BP neural network-based reverse modeling method for Doherty power amplifier |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011233431.4A CN112329371B (en) | 2020-11-06 | 2020-11-06 | IALO-BP neural network-based reverse modeling method for Doherty power amplifier |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112329371A CN112329371A (en) | 2021-02-05 |
| CN112329371B true CN112329371B (en) | 2024-06-14 |
Family
ID=74315720
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011233431.4A Active CN112329371B (en) | 2020-11-06 | 2020-11-06 | IALO-BP neural network-based reverse modeling method for Doherty power amplifier |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112329371B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114329933B (en) * | 2021-12-20 | 2022-08-23 | 北京力通通信有限公司 | Broadband power amplifier input-output characteristic fitting method |
| CN117991349A (en) * | 2024-04-07 | 2024-05-07 | 吉林大学 | Microseism positioning method based on improved ant lion optimization algorithm |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107798199B (en) * | 2017-11-09 | 2020-07-07 | 华中科技大学 | Hydroelectric generating set parameter closed-loop identification method |
| CN109212465B (en) * | 2018-09-01 | 2024-01-30 | 哈尔滨工程大学 | Special array dynamic direction finding method based on cultural ant lion mechanism |
| CN111638055B (en) * | 2020-05-18 | 2021-11-16 | 东南大学 | Gearbox fault diagnosis method based on resonance sparse decomposition improved algorithm |
-
2020
- 2020-11-06 CN CN202011233431.4A patent/CN112329371B/en active Active
Non-Patent Citations (2)
| Title |
|---|
| 基于改进型蚁狮算法的主动配电网优化重构;温东翰;徐敏;;电气应用;20200715(第07期);全文 * |
| 改进蚁群算法的BRBP神经网络功放逆向建模方法;南敬昌;臧净;高明明;;激光与光电子学进展;20190729(第01期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112329371A (en) | 2021-02-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112329371B (en) | IALO-BP neural network-based reverse modeling method for Doherty power amplifier | |
| CN103379608A (en) | Method for configuring multilevel transmitting power in wireless sensor network | |
| CN111199126B (en) | Quick optimization design method for chip antenna | |
| CN111291582B (en) | RFID antenna optimal deployment method based on jellyfish group algorithm | |
| CN111432368A (en) | Ranging and positioning method suitable for sparse anchor node WSN | |
| CN103648139A (en) | Cultural ant colony algorithm-based wireless sensor network node deployment design method | |
| CN111757266B (en) | UAV data acquisition trajectory algorithm based on solar power supply type agricultural Internet of things | |
| CN111935724A (en) | Wireless sensor network topology optimization method based on asynchronous deep reinforcement learning | |
| CN114615730A (en) | Content coverage oriented power distribution method for backhaul limited dense wireless network | |
| CN105451268B (en) | A kind of isomery cell accessing method of high energy efficiency | |
| CN114662638A (en) | Mobile robot path planning method based on improved artificial bee colony algorithm | |
| CN1889486A (en) | Dynamic energy managing method and apparatus in wireless sensor network | |
| CN116566777B (en) | Frequency hopping signal modulation identification method based on graph convolution neural network | |
| Guney et al. | Input resistance calculation for circular microstrip antennas using adaptive neuro-fuzzy inference system | |
| CN110505293B (en) | Cooperative caching method based on improved drosophila optimization algorithm in fog wireless access network | |
| CN110766072A (en) | Automatic generation method of computational graph evolution AI model based on structural similarity | |
| CN116187383A (en) | Method for determining 3D NoC optimal test plan based on improved whale optimization algorithm | |
| Ginde et al. | Game theoretic analysis of joint link adaptation and distributed power control in GPRS | |
| CN116112934A (en) | End-to-end network slice resource allocation method based on machine learning | |
| CN114912398A (en) | Method for selecting analog integrated circuit parameters by combining quantum genetic algorithm with neural network | |
| Zhang et al. | K-means-based multigroup differential evolution optimization framework for design of MIMO antenna with decoupling elements | |
| CN112149360A (en) | ALO-LMBP neural network-based double-notch characteristic ultra-wideband antenna reverse modeling research method | |
| CN114900871A (en) | Wireless sensor network clustering method based on affinity propagation and chaos lion group | |
| Maddala | Improved selective mapping technique for reduction of PAPR in MIMO-OFDM wireless communication | |
| CN115865220B (en) | AUV auxiliary underwater sound network data transmission method based on intelligent super surface |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |