CN115796095B - Pulse interpolation timing parameter optimization method based on genetic algorithm - Google Patents
Pulse interpolation timing parameter optimization method based on genetic algorithm Download PDFInfo
- Publication number
- CN115796095B CN115796095B CN202310073705.5A CN202310073705A CN115796095B CN 115796095 B CN115796095 B CN 115796095B CN 202310073705 A CN202310073705 A CN 202310073705A CN 115796095 B CN115796095 B CN 115796095B
- Authority
- CN
- China
- Prior art keywords
- pulse
- candidate solution
- amplitude
- fitness
- iteration
- 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
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Complex Calculations (AREA)
- Nuclear Medicine (AREA)
Abstract
The invention discloses a pulse interpolation timing parameter optimization method based on a genetic algorithm, which randomly generates a certain number of candidate solutionsConstructing an initial candidate solution set { s }; randomly generating a set of pulse signals { V [ n ]]The pulse signal is a digital signal acquired at a fixed sampling rate, and the characteristic parameters comprise pulse amplitude V p And rise time t r The method comprises the steps of carrying out a first treatment on the surface of the Based on the generated pulse signal { V [ n ]]Sequentially calculating fitness of each candidate solution in the candidate solution set { s }, and updating the candidate solution set; judging whether the set iteration times are reached, if so, selecting an optimal solution from a candidate solution set after iteration is completed, and achieving the purpose of parameter optimization; otherwise, continuing to update the candidate solution iteration. The method can carry out optimal design on parameters in the pulse interpolation timing method, realizes high time measurement precision under a relatively low sampling rate, and has good robustness on the shape and amplitude change of the pulse.
Description
Technical Field
The invention relates to the technical field of high-precision time measurement, in particular to a pulse interpolation timing parameter optimization method based on a genetic algorithm.
Background
The high-precision time measurement has wide application in the fields of nuclear physics and particle physics experiments, nuclear medicine imaging and the like. For example, in Time-of-flight (ToF) positron emission tomography (Positron Emission Tomography, PET), extremely high pulse arrival Time measurement accuracy is required to achieve high spatial imaging resolution. The waveform digitizing method collects analog pulses output by the detector through a high-speed sampling technology, extracts time information carried by the pulses through a digital signal processing method, and can obtain information in the full frequency band of the pulses compared with the traditional method of screening and combining time digital conversion, and can obtain other interesting information such as amplitude, shape and the like of the pulses besides extracting the time information based on the same digital signals.
The current waveform digitizing method is widely studied internationally, but in order to achieve high enough time measurement accuracy, a common technical scheme adopts an extremely high sampling rate (> 5 GSps), and the high sampling rate can bring problems such as high cost, high power consumption, high data transmission and processing complexity and the like. Thus, a time extraction method at a relatively low sampling rate needs to be studied. Under the lower sampling rate, in order to reduce the time measurement error, interpolation is generally required to be performed on the acquired digital signals, and then digital threshold screening and other methods are applied to obtain the arrival time of the pulse, and the interpolation process can be realized by an interpolation filter. In the practical application process, a plurality of parameters can influence the time measurement precision, including interpolation multiple, filter order, filter cut-off frequency, window function shape parameters, threshold constant ratio and the like.
The selection of the parameters in the prior art can adopt a simple empirical method, but the parameters are very easy to fall into a local optimal solution, and the time precision is difficult to reach; or use too high a filter order, which results in more computational resources being consumed.
Disclosure of Invention
The invention aims to provide a pulse interpolation timing parameter optimization method based on a genetic algorithm, which can optimally design parameters in the pulse interpolation timing method, realizes high time measurement precision under a relatively low sampling rate, and has good robustness to the shape and amplitude variation of pulses.
The invention aims at realizing the following technical scheme:
a method for optimizing pulse interpolation timing parameters based on a genetic algorithm, the method comprising:
Each candidate solution comprises one-time selection of each parameter to be optimized, and the selected probability meets the uniform distribution in a certain range;
parameters to be optimized comprise interpolation filter parameters and digital threshold screening parameters; plug-inThe filter parameters include interpolation multiple M, filter order N, and filter cut-off frequency f c Window function shape parameter β; the digital threshold screening parameter comprises a threshold constant ratio c f The method comprises the steps of carrying out a first treatment on the surface of the Wherein M and N are discrete variables and f c Beta and c f Is a continuous variable;
According to the technical scheme provided by the invention, the parameters in the pulse interpolation timing method can be optimally designed, so that high time measurement accuracy under a relatively low sampling rate is realized, and meanwhile, the method has good robustness to pulse shape and amplitude variation, has good effectiveness, universality and practicability, and is easy for hardware integration.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the description of the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a schematic flow chart of a pulse interpolation timing parameter optimization method based on a genetic algorithm according to an embodiment of the present invention.
Fig. 2 is a reference PMT pulse waveform used in an embodiment of the present invention.
Fig. 3 is a time precision simulation verification diagram based on genetic algorithm parameter optimization provided by the embodiment of the invention.
Detailed Description
The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments of the present invention, and this is not limiting to the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to fall within the scope of the invention.
The method is suitable for pulse signals output by detectors such as PMT/SiPM, waveform digitization is completed through a sampling circuit, interpolation filtering is performed in a digital domain, and extraction of pulse arrival time is completed through a digital threshold screening method; wherein the parameters of the interpolation filter and the digital threshold screening are obtained by genetic algorithm simulation optimization. Fig. 1 is a schematic flow chart of a pulse interpolation timing parameter optimization method based on a genetic algorithm according to an embodiment of the present invention, where the method includes:
Each candidate solution comprises one-time selection of each parameter to be optimized, and the selected probability meets the uniform distribution in a certain range;
parameters to be optimized comprise interpolation filter parameters and digital threshold screening parameters; the interpolation filter parameters comprise interpolation multiple M, filter order N and filter cut-off frequency f c Window function shape parameter β; the digital threshold screening parameter comprises a threshold constant ratio c f The method comprises the steps of carrying out a first treatment on the surface of the Wherein M and N are discrete variables and f c Beta and c f Is a continuous variable;
in addition, the number of candidate solution individuals to be optimized is not only one, but is optimized as a whole of one candidate solution set { s }.
For example, the number of candidate solutions in the candidate solution set { s } is set to 400; selection of interpolation multiple MTaking the range of 2, 4 or 8; the filter order N is set to be proportional to M, the coefficient is chosen to be 6 to 10 times, for example, in the case of 8 times interpolation, the order nminium is 48 and the maximum is 80; filter cut-off frequency f c Set inversely proportional to M, the coefficient is chosen to be between 0.05 and 1, for example in the case of 8-fold interpolation, the cut-off frequency f c The representation in the digital domain is at a minimum
Maximum->
The method comprises the steps of carrying out a first treatment on the surface of the The window function uses a Kaiser window, i.e. is expressed by the following formula:
wherein beta is a window function shape parameter, and the selection range is set to be between 1 and 20; i 0 Zero order first class bessel function;
threshold constant ratio c f For a digital constant ratio, between 0.3 and 0.9.
In this step, the randomness of the pulse signal is reflected in several aspects: sampling start point t 0 Random of pulse amplitude V p Random of (a), rise time t r Random of (d) and noise amplitude V n [n]Random of (a);
pulse amplitude V p And rise time t r The range of (2) is set according to the actual detector type, the noise amplitude V n [n]Referring to the noise setting of the actual front-end electronics hardware circuit;
for the analog pulse waveform V (t), the expression for generating the random digital pulse signal V [ n ] is as follows:
wherein T is a sampling period; v (V) p0 Is the reference pulse amplitude; t is t r0 Is the reference rise time; t is t 0 Is a sampling starting point; v (V) n [n]The value of n is a natural number for noise amplitude.
For example, as shown in fig. 2, the reference PMT pulse waveform used in the embodiment of the present invention has a leading edge of about 4ns, the PMT pulse is digitized by selecting a sampling rate of 500 MSps, and then the time information carried in the digitized pulse is obtained by a time extraction algorithm, where the amplitude and time of the reference pulse are the amplitude and time of the pulse waveform shown in fig. 2; the sampling period T is 2ns; range of variation of amplitude
Between 0.2 and 1; variation range of rise time->
Between 0.9 and 1.8.
In addition, to ensure the optimized result for pulse amplitude V p And rise time t r Is a random set of pulse signals generated at the beginning of each iteration.
in this step, the fitness is determined by the particular candidate solution s (M, N, f) in the set of candidate solutions { s } c , β, c f ) The corresponding time precision is determined, namely, the higher the precision is, the higher the individual fitness is;
the specific calculation process of the fitness comprises the following steps:
for the original pulse signal { V [ n ]]Each pulse in the sequence of V at the same amplitude p And rise time t r At the sampling start point t 0 Plus a fixed delay T d Selecting a random noise amplitude V n ’[n]Regenerating a set of delayed pulse signals { V ] d [n]}；
Statistics of the original pulsePunching signal { V [ n ]]Sum of the delayed pulse signals { V } d [n]Standard deviation RMS of the difference between the arrival times of the two pulse waveforms, fitness being the reciprocal of the RMS value; the pulse waveform arrival time is obtained by adopting a digital constant ratio timing method;
selecting partial candidate solution individuals from the current candidate solution set { s } with a certain probability to become parents;
generating offspring from the parents through a certain crossing and mutation method, and sequentially calculating the fitness of offspring individuals;
and selecting a candidate solution individual with high fitness from the parent and the offspring as a next generation candidate solution set { s }, thereby updating the candidate solution set.
For example, the number of pulse signals { V [ n ] } is 1000, the parent is selected with equal probability, and each candidate solution is selected with a probability of 80%.
The crossing method is as follows: for each continuous variable, i.e. f c Beta and c f If the value of parent 1 is x 1 The value of parent 2 is x 2 The values of the filings generated by the crossover are respectively
And->
The method comprises the steps of carrying out a first treatment on the surface of the For discrete variables, i.e., M and N, the child inherits the value of one of the parents.
The mutation method is as follows: for each progeny of crossover generation, the variation occurred with a 40% probability as a single point variation, i.e., at (M, N, f c , β, c f ) A value is selected again randomly from the set range.
In this step, the number of iterations may be set according to the requirement, for example, the number of iterations is set to 100.
In specific implementation, after iteration is completed, the time precision under the reference pulse waveform can be further verified, and the specific process is as follows:
dividing an input reference pulse waveform into two paths by adopting a line delay method, wherein fixed delay is added to one path, and digitizing is carried out respectively; calculating pulse arrival time according to the digitized pulse waveform by adopting a digital constant ratio timing method, and counting the RMS of the difference between the arrival times of two paths of pulse waveforms; the accuracy of a single channel is the delay RMS, taking into account the uncorrelated measurement of the two channels
The method comprises the steps of carrying out a first treatment on the surface of the Single channel accuracy was calculated separately at different delay amounts to verify time accuracy performance.
As shown in FIG. 3, which is a time precision simulation verification diagram based on genetic algorithm parameter optimization provided by the embodiment of the invention, it can be seen from FIG. 3 that the method of the invention can achieve higher time measurement precision at a relatively lower sampling rate.
It is noted that what is not described in detail in the embodiments of the present invention belongs to the prior art known to those skilled in the art.
In summary, the method provided by the embodiment of the invention has the following advantages:
(1) The invention realizes high time measurement precision under relatively low sampling rate, and has good robustness to the shape and amplitude variation of the pulse;
(2) The method can optimize the order of the filter in the parameter iteration process, can reduce the resource consumption in hardware implementation, and has good effectiveness, universality and practicability.
The foregoing is only a preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions easily contemplated by those skilled in the art within the scope of the present invention should be included in the scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims. The information disclosed in the background section herein is only for enhancement of understanding of the general background of the invention and is not to be taken as an admission or any form of suggestion that this information forms the prior art already known to those of ordinary skill in the art.
Claims (3)
1. A method for optimizing pulse interpolation timing parameters based on a genetic algorithm, the method comprising:
step 1, randomly generating a certain number of candidate solution individuals to construct a candidate solution set, wherein the candidate solution set is expressed as { s (M, N, f) c , β, c f ) -a }; each candidate solution comprises one-time selection of each parameter to be optimized, and the selected probability meets the uniform distribution in a certain range; parameters to be optimized comprise interpolation filter parameters and digital threshold screening parameters; the interpolation filter parameters comprise interpolation multiple M, filter order N and filter cut-off frequency f c Window function shape parameter β; the digital threshold screening parameter comprises a threshold constant ratio c f The method comprises the steps of carrying out a first treatment on the surface of the Wherein M and N are discrete variables and f c Beta and c f Is a continuous variable;
step 2, randomly generating a group of pulse signals { V [ n ]]The pulse signal is a digital signal acquired at a fixed sampling rate, and the characteristic parameters comprise pulse amplitude V p And rise time t r ；
Step 3, based on the pulse signal { V [ n ] } generated in the step 2, sequentially calculating the fitness of each candidate solution in the candidate solution set { s }, and updating the candidate solution set;
the fitness is determined by the time precision corresponding to a specific candidate solution in the candidate solution set { s }, namely, the higher the precision is, the higher the individual fitness is; the specific calculation process of the fitness comprises the following steps:
for the original pulse signal { V [ n ]]Each pulse in the sequence of V at the same amplitude p And rise time t r At the sampling start point t 0 Plus a fixed delay T d Selecting a random noise amplitude, denoted as V n ’[n]Regenerating a set of delayed pulse signals { V ] d [n]}；
Counting primary pulse signal { V [ n ]]Sum of the delayed pulse signals { V } d [n]Standard deviation RMS of the difference between the arrival times of the two pulse waveforms, fitness being the reciprocal of the RMS value; the pulse waveform arrival time is obtained by adopting a digital constant ratio timing method;
selecting partial candidate solution individuals from the candidate solution set { s } with a certain probability to become parents;
generating offspring from the parent through a crossover and mutation method, and sequentially calculating the fitness of offspring individuals;
selecting candidate solution individuals with high fitness from the parent and the offspring as a next generation candidate solution set, thereby updating the candidate solution set;
step 4, judging whether the set iteration times are reached, if so, selecting an optimal solution from a candidate solution set after iteration is completed, and achieving the purpose of parameter optimization; otherwise, continuing to update the candidate solution iteration.
2. The method for optimizing pulse interpolation timing parameters based on genetic algorithm according to claim 1, wherein in step 2, randomness of the pulse signal is represented in the following aspects: sampling start point t 0 Random of pulse amplitude V p Random of (a), rise time t r Random of (d) and noise amplitude V n [n]Random of (a);
pulse amplitude V p And rise time t r The range of (2) is set according to the actual detector type, the noise amplitude V n [n]Referring to the noise setting of the actual front-end electronics hardware circuit;
for the analog pulse waveform V (t), the expression for generating the random digital pulse signal V [ n ] is as follows:
3. The method of optimizing pulse interpolation timing parameters based on genetic algorithm according to claim 1, wherein in step 2, the pulse amplitude V is determined for the optimization result p And rise time t r Is a random set of pulse signals generated at the beginning of each iteration.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310073705.5A CN115796095B (en) | 2023-02-07 | 2023-02-07 | Pulse interpolation timing parameter optimization method based on genetic algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310073705.5A CN115796095B (en) | 2023-02-07 | 2023-02-07 | Pulse interpolation timing parameter optimization method based on genetic algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115796095A CN115796095A (en) | 2023-03-14 |
| CN115796095B true CN115796095B (en) | 2023-04-21 |
Family
ID=85430208
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310073705.5A Active CN115796095B (en) | 2023-02-07 | 2023-02-07 | Pulse interpolation timing parameter optimization method based on genetic algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115796095B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3529628A1 (en) * | 2016-10-24 | 2019-08-28 | Xia LLC | Ratio-reference measurement of the arrival time and/or amplitude of a digitized electronic pulse |
Family Cites Families (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10671923B2 (en) * | 2013-12-20 | 2020-06-02 | Gemtrex Inc. | Genetic method for the tracking of time varying signals |
| CN106680778B (en) * | 2017-01-03 | 2019-04-26 | 中国科学技术大学 | 3 D stereo random antenna array structure method |
| US10419785B2 (en) * | 2017-07-21 | 2019-09-17 | Cisco Technology, Inc. | Distributed learning video encoder/decoder |
| US12001931B2 (en) * | 2018-10-31 | 2024-06-04 | Allstate Insurance Company | Simultaneous hyper parameter and feature selection optimization using evolutionary boosting machines |
| US11785340B2 (en) * | 2020-03-16 | 2023-10-10 | Sony Semiconductor Solutions Corporation | Signal processing device, signal processing method, and parameter search method |
| US11501044B1 (en) * | 2020-12-23 | 2022-11-15 | Cadence Design Systems, Inc. | System and method for decoupling capacitor selection and placement using genetic optimization |
| CN113010319A (en) * | 2021-03-31 | 2021-06-22 | 华南理工大学 | Dynamic workflow scheduling optimization method based on hybrid heuristic rule and genetic algorithm |
| CN113204743B (en) * | 2021-05-19 | 2022-09-02 | 成都大学 | Neutron-gamma discrimination method based on genetic algorithm |
| CN114184870A (en) * | 2021-12-14 | 2022-03-15 | 河北科技大学 | Non-invasive load identification method and equipment |
| CN114841106A (en) * | 2022-04-06 | 2022-08-02 | 山东大学 | Integrated circuit optimization method and system based on rule-guided genetic algorithm |
-
2023
- 2023-02-07 CN CN202310073705.5A patent/CN115796095B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3529628A1 (en) * | 2016-10-24 | 2019-08-28 | Xia LLC | Ratio-reference measurement of the arrival time and/or amplitude of a digitized electronic pulse |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115796095A (en) | 2023-03-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Nakhostin | Recursive algorithms for real-time digital ${\rm CR}\!-\!({\rm RC})^{\rm n} $ pulse shaping | |
| Back et al. | Centrality dependence of the charged particle multiplicity near midrapidity in Au+ Au collisions at s NN= 130 and 200 GeV | |
| Eldar et al. | Robust recovery of signals from a structured union of subspaces | |
| US8325076B2 (en) | Time-to-digital converter | |
| CN106156852B (en) | A kind of Gauss overlap kernel impulse response estimation method | |
| CN109274369B (en) | Method and device for digitizing scintillation pulse | |
| WO2016127357A1 (en) | Fpga-based time-to-digital converter | |
| Saxena et al. | Investigation of FPGA-based real-time adaptive digital pulse shaping for high-count-rate applications | |
| CN105989410B (en) | A kind of overlap kernel impulse decomposition method | |
| JP6251393B2 (en) | Method and device for obtaining when a low light pulse passes a threshold | |
| CN109086555A (en) | Using the simulation spectrum curve emulation mode of Monte Carlo method | |
| CN115796095B (en) | Pulse interpolation timing parameter optimization method based on genetic algorithm | |
| CN102916683A (en) | Parameter-adjustable nuclear pulse simulation method | |
| WO2006117191A2 (en) | Use of digital filters in a detector and method for obtaining energy spectra of nuclear radiation | |
| Rehman et al. | Classification and Denoising of Cosmic-Ray Radio Signals using Deep Learning | |
| CN108333912B (en) | Time correction method for switched capacitor array chip | |
| CN110061733A (en) | Method of counting, method for reconstructing and the chip of pulse signal | |
| CN104242877B (en) | Core pulse signal digital Gaussian manufacturing process based on simulation CR RC circuits | |
| CN116432703B (en) | Pulse height estimation method, system and terminal based on composite neural network model | |
| CN105718240B (en) | A kind of bit stream adder of any road input signal based on summation delta modulation | |
| CN107450886B (en) | Method and device for generating Gaussian random signal simulating Gaussian white noise | |
| CN114252899B (en) | Cascade impulse convolution forming method and device for kernel signal | |
| CN110334471B (en) | PRI sub-period extraction method for ragged sequence and pulse group ragged sequence | |
| CN112134545A (en) | Trapezoidal forming method, system, terminal and medium based on optimal filter | |
| McLean et al. | Limitations of decision based pile-up correction algorithms |
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 |