WO2020252673A1 - 一种提升可穿戴可拉伸电化学传感器检测性能的设计方法 - Google Patents

一种提升可穿戴可拉伸电化学传感器检测性能的设计方法 Download PDF

Info

Publication number
WO2020252673A1
WO2020252673A1 PCT/CN2019/091799 CN2019091799W WO2020252673A1 WO 2020252673 A1 WO2020252673 A1 WO 2020252673A1 CN 2019091799 W CN2019091799 W CN 2019091799W WO 2020252673 A1 WO2020252673 A1 WO 2020252673A1
Authority
WO
WIPO (PCT)
Prior art keywords
design
performance
wses
performance index
electrochemical sensor
Prior art date
Application number
PCT/CN2019/091799
Other languages
English (en)
French (fr)
Inventor
纪震
Original Assignee
纪震
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 纪震 filed Critical 纪震
Priority to PCT/CN2019/091799 priority Critical patent/WO2020252673A1/zh
Priority to CN201980000860.6A priority patent/CN110799976A/zh
Publication of WO2020252673A1 publication Critical patent/WO2020252673A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10Complex mathematical operations
    • G06F17/18Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01NINVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N27/00Investigating or analysing materials by the use of electric, electrochemical, or magnetic means
    • G01N27/26Investigating or analysing materials by the use of electric, electrochemical, or magnetic means by investigating electrochemical variables; by using electrolysis or electrophoresis
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/24Classification techniques
    • G06F18/241Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2411Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines

Definitions

  • the invention relates to the field of electrochemical sensors, in particular to a design method for improving the detection performance of a wearable stretchable electrochemical sensor.
  • ISM ion selective membrane
  • the new sweat sensor design is further developed into a wearable and stretchable electrochemical sensor (WSES).
  • WSES wearable and stretchable electrochemical sensor
  • the purpose of the present invention is to provide a design method for improving the detection performance of a wearable stretchable electrochemical sensor, aiming to solve the problem of insufficient robustness of the existing WSES design method to tensile strength .
  • a design method for improving the detection performance of a wearable stretchable electrochemical sensor which includes the following steps:
  • the wearable stretchable electrochemical sensor is optimized.
  • the design method for improving the detection performance of a wearable stretchable electrochemical sensor wherein the orthogonal test design is used to obtain a design parameter matrix interpolation combination with a variable degree of dispersion and a stronger generalization ability.
  • the steps to obtain the performance indicators of WSES under different tensile strengths include:
  • L is the maximum tensile strength, and generate different tensile strength performance index matrices through WSES experiment among them y km,l represents the mth performance index obtained under the condition of tensile strength l using the kth group of design parameters.
  • the performance indicators include detection range, linearity, stability and electrical impedance spectrum.
  • the step of establishing a nonlinear regression model corresponding to each performance index according to the performance index includes:
  • the step of using a multi-objective optimization method to simultaneously search for the Pareto non-dominant optimal solution of the performance index, and obtaining the WSES design result includes:
  • Update g g+1, if g ⁇ G, return to calculation Corresponding performance index steps; if g ⁇ G, the update ends, and each individual in the P ND becomes the optimized design on the Pareto front end.
  • rand(a,b) means returning a random number that obeys uniform distribution in the range of [a,b], max(x n ) and min(x n ) respectively represent the maximum and minimum values of the nth design parameter, L max Indicates the maximum tensile strength that WSES can withstand.
  • the design method for improving the detection performance of a wearable stretchable electrochemical sensor wherein the calculation
  • the corresponding performance indicators include steps:
  • Update m m+1, if m ⁇ M, re-apply the regression model R m to estimate the current input Output when Until m ⁇ M, use non-dominated sorting genetic algorithm II to update the evolutionary population ps.
  • the design method for improving the detection performance of a wearable stretchable electrochemical sensor provided by the present invention participates in the entire multi-objective optimization process by taking the tensile strength as the N+1 dimensional variable of the evolutionary individual, which can take into account other designs
  • the relationship between parameters and tensile strength enables the optimal individuals in a non-dominant position to fully reflect the impact of different tensile strengths on the performance of WSES, ensuring the optimal performance under different tensile strengths, which solves the problem There is a problem of insufficient robustness of the WSES design method to tensile strength.
  • Figure 1 is a schematic diagram of the structure of a WSES of the present invention.
  • FIG. 2 is a schematic diagram of the graphene oxide-carbon nanotube structure of the three-dimensional hollow structure of the present invention.
  • FIG. 3 is a flowchart of a preferred embodiment of a design method for improving the detection performance of a wearable stretchable electrochemical sensor according to the present invention.
  • Fig. 4 is a flow chart of establishing multiple performance index models under different tensile strengths according to the present invention.
  • Fig. 5 is a schematic diagram of the design parameter matrix X (N ⁇ K dimension) of the present invention.
  • Fig. 6 is a schematic diagram of the performance index matrix Y (M ⁇ K dimension) of the present invention.
  • Figure 7 is a matrix of multiple performance indicators for different tensile strengths of the present invention (M ⁇ K ⁇ L dimension) schematic diagram.
  • Figure 8 is the regression training data pair of the mth performance index of the present invention (Under different tensile strength conditions) Schematic diagram of construction.
  • Figure 9 is a flow chart of the multi-objective optimization design of the present invention.
  • the present invention provides a design method for improving the detection performance of a wearable stretchable electrochemical sensor.
  • the present invention will be described in further detail below. It should be understood that the specific embodiments described here are only used to explain the present invention, but not to limit the present invention.
  • Figure 1 and Figure 2 are a schematic diagram of a typical WSES structure and a three-dimensional hollow structure graphene oxide-carbon nanotube structure diagram, where REM is the reference electrode membrane (Reference Electrode Membrane), PU is Polyurethane (Polyurethane), PDMS is Polyurethane Poly-dimethylsiloxane (Poly-dimethylsiloxane), PU and PDMS are very soft after curing, suitable for flexible materials.
  • Nonactin is an ionophore (ionophore), a chemical substance that reversibly binds ions, and can transmit target ions through the dielectric membrane. Nonactin is specifically used for the transmission of ammonia ions.
  • Polyurethane, dioctyl sebacate (Bis-ethylhexyl , Sebacate, DOS) and tetrahydrofuran (Tetra Hydofuran, THF) are used as the matrix polymer of ion selective membranes, plasticizers and solvents.
  • the WSES involves the setting of multiple parameters in the design process, as shown in Table 1.
  • the main performance indicators of WSES include: 1. Detection range: The normal range of human sweat ammonia ion concentration is 10 -4 -10 -3 M, and the detection range is generally required to be 10 -6 -1 M. 2. Sensitivity: Take the logarithm based on 10 for the change of ammonia ion concentration from 10 -6 M to 1M. The higher the corresponding voltage, the greater the sensitivity. The general slope is 59mv/log[NH 4 +]. 3. Linearity: Under different tensile strengths (0%-40%), the sensor maintains a good linear response when the ammonia ion concentration is from 10 -4 M to 1M; at the same time, the electrode potential changes with the ion activity.
  • the existing WSES design technology represented by 3D rGO-CNT has the following disadvantages:
  • the adjustment process of the most important design parameters in the design process relies too much on the experience of the design engineer. If the engineer's ability is insufficient, the final performance indicators will inevitably decline; even if the design engineer has rich experience, the "adjust-verify" process It needs to be repeated many times, and the design engineer needs to participate in the whole process until the performance indicators are met. This is very time-consuming and labor-intensive, and the final result cannot be guaranteed to be the theoretical optimal value.
  • the key and difficult design is the introduction of different tensile strengths.
  • An optimal set of design parameters is obtained under a certain tensile strength.
  • the performance indicators of WSES may be greatly interfered or even deteriorated sharply. Therefore, it is difficult for the same set of design parameters to take into account all the tensile strengths.
  • Tensile strength is likely to cause poor overall performance of WSES.
  • the changes of different tensile strengths and performance indicators show a non-linear relationship.
  • the use of a simple linear regression model cannot guarantee the accuracy of performance prediction and simulation, resulting in high analysis difficulty and insufficient generalization performance.
  • an embodiment of the present invention provides a design method for improving the detection performance of a wearable stretchable electrochemical sensor, wherein, as shown in FIG. 3, it includes the steps:
  • This embodiment takes tensile strength as the N+1 dimensional variable of the evolutionary individual to participate in the entire multi-objective optimization process, which can take into account the relationship between other design parameters and tensile strength, so that the optimal individuals in a non-dominated position It can fully reflect the influence of different tensile strengths on the performance of WSES, and ensure the optimal performance under different tensile strengths. This solves the problem of insufficient robustness of the existing WSES design methods to tensile strength.
  • the orthogonalization experiment design is used to obtain a design parameter matrix interpolation combination with variable dispersion and stronger generalization ability, and the WSES at different tensile strengths is obtained through electrochemical experiments.
  • the following performance indicators and the steps of establishing a nonlinear regression model corresponding to each performance indicator according to the performance indicators specifically include:
  • the design parameters can be selected from Table 1, As an example, the design parameters include electrodeposition voltage, electrodeposition time, the amount of REM, the amount of PVB, the amount of methanol sodium chloride, the amount of PU, the amount of PDMS, and the amount of ISM, but are not limited thereto;
  • the performance indexes include detection range, linearity, stability, resistance Specific values such as anti-spectrum and sensitivity, but not limited to this;
  • the model can simulate different tensile strength changes with a distance of 1% or less, avoiding setting L to 40, which significantly reduces the number of experiments; at the same time, the model is The simulation between different design parameters of K groups far exceeds the interpolation parameters of K groups. These two kinds of simulation performance improvements brought by regression models significantly reduce the cost of experiments and manpower equipment. Efficient nonlinear regression model ensures the accuracy of performance estimation and effectively reduces or even avoids design deviation.
  • the step S30 using a multi-objective optimization method to simultaneously search for the Pareto non-dominant optimal solution of the performance index, to obtain the WSES design result, specifically includes:
  • rand(a,b) means returning a random number that obeys uniform distribution in the range of [a,b], max(x n ) and min(x n ) respectively represent the maximum and minimum values of the nth design parameter, L max Indicates the maximum tensile strength that WSES can withstand;
  • the method provided in this embodiment can realize the formation of multiple Pareto non-dominated set optimal schemes through independent operation, and their performance biases are different. If the WSES design requirements are changed, there is no need to re-experiment and model calculations, just select another set of design parameters that meet the requirements from the non-dominated optimal set.
  • the method provided in this embodiment is suitable for a sensor for detecting the ammonia ion concentration of human sweat.
  • the material of the sensor is changed, it is used to detect the blood glucose, lactate, sodium (Na+) ion or potassium ion (K+) of human sweat.
  • this method is also suitable for the optimal design of the sensor, and only needs to modify the multi-performance index matrix and the multi-performance index matrix of different tensile strengths.
  • the design method for improving the detection performance of a wearable stretchable electrochemical sensor takes the tensile strength as the N+1 dimensional variable of the evolutionary individual to participate in the entire multi-objective optimization process, which can take into account other aspects.
  • the relationship between design parameters and tensile strength allows the optimal individuals in a non-dominant position to fully reflect the impact of different tensile strengths on the performance of WSES, ensuring the optimal performance under different tensile strengths, which solves
  • the existing WSES design method has insufficient robustness to tensile strength.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Pure & Applied Mathematics (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Mathematical Physics (AREA)
  • General Engineering & Computer Science (AREA)
  • Computational Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Mathematical Optimization (AREA)
  • Chemical & Material Sciences (AREA)
  • Mathematical Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Operations Research (AREA)
  • Biochemistry (AREA)
  • General Health & Medical Sciences (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Electrochemistry (AREA)
  • Chemical Kinetics & Catalysis (AREA)
  • Molecular Biology (AREA)
  • Analytical Chemistry (AREA)
  • Probability & Statistics with Applications (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Algebra (AREA)
  • Artificial Intelligence (AREA)
  • Databases & Information Systems (AREA)
  • Software Systems (AREA)
  • Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
  • Investigating Strength Of Materials By Application Of Mechanical Stress (AREA)

Abstract

一种提升可穿戴可拉伸电化学传感器检测性能的设计方法,方法包括步骤:使用正交化试验设计取得设计参数矩阵内插组合,通过电化学实验得到WSES在不同拉伸强度下的性能指标(S10);建立与每个性能指标对应的非线性回归模型(S20);使用多目标优化方法同时搜索性能指标的Pareto非支配最优解,获得WSES设计结果(S30);根据WSES设计结果对可穿戴可拉伸电化学传感器进行优化设计(S40)。本设计方法能够兼顾其它设计参数和拉伸强度之间的关系,使得处于非支配地位的最优个体们能够充分反映不同拉伸强度对WSES性能的影响,保证了在不同拉伸强度下的性能最优,这解决了现有WSES设计方法对拉伸强度的鲁棒性不足问题。

Description

一种提升可穿戴可拉伸电化学传感器检测性能的设计方法 技术领域
本发明涉及电化学传感器领域,尤其涉及一种提升可穿戴可拉伸电化学传感器检测性能的设计方法。
背景技术
近年来,可穿戴电化学传感器(Wearable Electrochemical Sensor,WES)备受关注并且取得长足发展。包含了丰富临床相关生物标志物的汗液是最适合于连续监测的生物体液之一,传统的丝网印刷电极(Screen-Printed Electrode,SPE)汗液传感器基于纺织物衬底,限制了皮肤可检测的区域,对纺织物材料种类的选取也有着较为苛刻的要求,这两种因素会极大地影响汗液检测性能,限制了传感器的灵活性。耐用性和稳定性是另外两种挑战,因而基于聚合物(Polymeric)衬底成为WES新的研究。最新的技术之一是使用三维中空结构的氧化石墨烯-碳纳米管(3D ReducedGraphene Oxide-CarbonNanotube,3DrGO-CNT),它的优点在于:1、由于它具有较大的表面积,所以能增强传感器的电荷传输能力,进而加快传感器的反应速度,增强传感器的整体性能。2、它的疏水特性(Hydrophobicity)进一步增强了传感器的长期稳定性。3、使用特定的离子载体(Ionophore)作为离子选择性膜(Ion Selective Membrane,ISM)能让只有目标离子通过同时隔绝其它干扰离子,可以进一步提高传感器的灵敏度、检测范围和可靠性。
由于采用新的衬底可以承受更大范围的拉伸程度,新的汗液传感器设计进一步发展为可穿戴可拉伸电化学传感器(Wearable and Stretchable Electrochemical Sensor,WSES)。整体设计难度大为增加,由于不同拉伸程度的引入,即使固定了传感器的各设计参数,也会导致传感器性能的较大波动,因而设计工程师很难发现最优或者接近最优的设计方案。
因此,现有技术还有待于改进和发展。
发明内容
鉴于上述现有技术的不足,本发明的目的在于提供一种提升可穿戴可拉伸电 化学传感器检测性能的设计方法,旨在解决现有WSES设计方法对拉伸强度的鲁棒性不足的问题。
本发明的技术方案如下:
一种提升可穿戴可拉伸电化学传感器检测性能的设计方法,其中,包括步骤:
使用正交化试验设计取得设计参数矩阵内插组合,通过电化学实验得到WSES在不同拉伸强度下的性能指标;
根据所述性能指标建立与每个性能指标对应的非线性回归模型;
使用多目标优化方法同时搜索性能指标的Pareto非支配最优解,获得WSES设计结果;
根据所述WSES设计结果对所述可穿戴可拉伸电化学传感器进行优化设计。
所述提升可穿戴可拉伸电化学传感器检测性能的设计方法,其中,所述使用正交化试验设计取得离散程度可变、泛化能力更强的设计参数矩阵内插组合,通过电化学实验得到WSES在不同拉伸强度下的性能指标的步骤包括:
设定WSES的设计参数,构成N维设计参数矢量X=[x 1,x 2,…,x n,…,x N],x n∈X,设定M维性能指标矢量Y=[y 1,y 2,…,y m,…,y M],y m∈Y;
根据ODE设计,计算设计参数的内插组合,构成设计参数矩阵X={X 1,X 2,…,X k,…,X K};
采用WSES进行实验,获得任意设计参数组合X k∈X所对应的性能指标矢量Y k,生成性能指标矩阵Y={Y 1,Y 2,…,Y k,…,Y K};
对于任意设计参数矢量X k,使用不同拉伸强度l,l∈[1,L],L为最大拉伸强度,通过WSES实验生成不同拉伸强度性能指标矩阵
Figure PCTCN2019091799-appb-000001
其中
Figure PCTCN2019091799-appb-000002
y km,l表示采用第k组设计参数,在拉伸强度l条件下获得的第m个性能指标。
所述提升可穿戴可拉伸电化学传感器检测性能的设计方法,其中,所述性能指标包括检测范围、线性度、稳定性和电阻抗光谱。
所述提升可穿戴可拉伸电化学传感器检测性能的设计方法,其中,根据所述性能指标建立与每个性能指标对应的非线性回归模型的步骤包括:
对于第m个性能指标,设计参数矢量X k和不同拉伸强度性能指标矢量
Figure PCTCN2019091799-appb-000003
构成回归数据对
Figure PCTCN2019091799-appb-000004
综合k从1到K的所有设计参数组合,得到第m个性能指标的回归训练数据集Φ m={Φ 1m2m,…,Φ km,…Φ Km},重复回归训练过程,直到形成所有M个性能指标的回归训练数据集Φ={Φ 12,…,Φ m,…Φ M};
对于任意Φ m,使用非线性回归模型中的支持向量机或者极限学习机估计出回归模型R m,重复这过程直到估计出所有M个性能指标的回归模型;
集成所有性能指标的回归模型构成多模型集合R={R 1,R 2,…,R m,…R M}。
所述提升可穿戴可拉伸电化学传感器检测性能的设计方法,其中,所述使用多目标优化方法同时搜索性能指标的Pareto非支配最优解,获得WSES设计结果的步骤包括:
初始化迭代计数器g=0,设最大迭代次数G,初始化Pareto非支配集合P ND为空集;
初始化用于多目标优化的进化种群ps,其中每个个体为N+1维矢量
Figure PCTCN2019091799-appb-000005
|ps|是进化种群的个体数量;
计算
Figure PCTCN2019091799-appb-000006
对应的性能指标;
使用非支配排序遗传算法II更新进化种群ps;
计算ps及P ND中所有个体的位置关系以及Pareto前端,选择其中处于非支配地位的个体,更新为新的P ND
更新g=g+1,若g<G,则返回计算
Figure PCTCN2019091799-appb-000007
对应的性能指标的步骤;若g≥G,则结束更新,此时P ND中每个个体都成为Pareto前端上的优化设计。
所述提升可穿戴可拉伸电化学传感器检测性能的设计方法,其中,所述
Figure PCTCN2019091799-appb-000008
每 维的取值为
Figure PCTCN2019091799-appb-000009
其中rand(a,b)表示返回[a,b]范围内服从均匀分布的随机数,max(x n)和min(x n)分别表示第n个设计参数的最大值和最小值,L max表示WSES能够承受的最大拉伸强度。
所述提升可穿戴可拉伸电化学传感器检测性能的设计方法,其中,所述计算
Figure PCTCN2019091799-appb-000010
对应的性能指标包括步骤:
初始化计数器m=0,初始化其性能指标的M维矢量
Figure PCTCN2019091799-appb-000011
为空;
应用回归模型R m,估计出当前输入
Figure PCTCN2019091799-appb-000012
时的输出
Figure PCTCN2019091799-appb-000013
更新m=m+1,若m<M,则重新应用回归模型R m,估计出当前输入
Figure PCTCN2019091799-appb-000014
时的输出
Figure PCTCN2019091799-appb-000015
直至m≥M时,使用非支配排序遗传算法II更新进化种群ps。
有益效果:本发明提供的提升可穿戴可拉伸电化学传感器检测性能的设计方法,通过以拉伸强度作为进化个体的第N+1维变量参与了整个多目标优化过程,这能够兼顾其它设计参数和拉伸强度之间的关系,使得处于非支配地位的最优个体们能够充分反映不同拉伸强度对WSES性能的影响,保证了在不同拉伸强度下的性能最优,这解决了现有WSES设计方法对拉伸强度的鲁棒性不足问题。
附图说明
图1为本发明一种WSES的结构示意图。
图2为本发明三维中空结构的氧化石墨烯-碳纳米管结构示意图。
图3为本发明一种提升可穿戴可拉伸电化学传感器检测性能的设计方法较佳实施例的流程图。
图4为本发明建立不同拉伸强度下多性能指标模型流程图。
图5为本发明设计参数矩阵X(N×K维)示意图。
图6为本发明性能指标矩阵Y(M×K维)示意图。
图7为本发明不同拉伸强度多性能指标矩阵
Figure PCTCN2019091799-appb-000016
(M×K×L维)示意图。
图8为本发明第m个性能指标的回归训练数据对
Figure PCTCN2019091799-appb-000017
(不同拉伸强度条件下)构建示意图。
图9为本发明多目标优化设计流程图。
具体实施方式
本发明提供一种提升可穿戴可拉伸电化学传感器检测性能的设计方法,为使本发明的目的、技术方案及效果更加清楚、明确,以下对本发明进一步详细说明。应当理解,此处所描述的具体实施例仅仅用以解释本发明,并不用于限定本发明。
本技术领域技术人员可以理解,除非特意声明,这里使用的单数形式“一”、“一个”、“所述”和“该”也可包括复数形式。应该进一步理解的是,本发明的说明书中使用的措辞“包括”是指存在所述特征、整数、步骤、操作、元件和/或组件,但是并不排除存在或添加一个或多个其他特征、整数、步骤、操作、元件、组件和/或它们的组。应该理解,当我们称元件被“连接”或“耦接”到另一元件时,它可以直接连接或耦接到其他元件,或者也可以存在中间元件。此外,这里使用的“连接”或“耦接”可以包括无线连接或无线耦接。这里使用的措辞“和/或”包括一个或更多个相关联的列出项的全部或任一单元和全部组合。
本技术领域技术人员可以理解,除非另外定义,这里使用的所有术语(包括技术术语和科学术语),具有与本发明所属领域中的普通技术人员的一般理解相同的意义。还应该理解的是,诸如通用字典中定义的那些术语,应该被理解为具有与现有技术的上下文中的意义一致的意义,并且除非像这里一样被特定定义,否则不会用理想化或过于正式的含义来解释。
图1和图2分别为典型的WSES结构示意图和三维中空结构的氧化石墨烯-碳纳米管结构示意图,其中REM为参比电极膜(Reference Electrode Membrane),PU为聚胺酯(Polyurethane),PDMS为聚二甲基硅氧烷(Poly-dimethylsiloxane),PU和PDMS固化后材质很软,适合作柔性材料。Nonactin是一种离子载体(ionophore),可逆地结合离子的化学物质,可将目标离子传输穿过介质膜,Nonactin专门用于氨离子的传输,聚胺酯,二辛基癸二酸酯(Bis-ethylhexyl,Sebacate,DOS)和四氢呋喃(Tetra Hydofuran,THF)是作为离子选择性膜的基体聚合物,增塑剂和溶剂使用。所述WSES在设计过程涉及多个 参数的设置,主要如表1所示。
表1主要设计参数
Figure PCTCN2019091799-appb-000018
在本实施例中,WSES的主要性能指标包括:1、检测范围:人体汗液氨离子浓度正常范围为10 -4-10 -3M,检测范围一般会要求在10 -6-1M。2、灵敏度:对氨离子浓度从10 -6M到1M的变化取以10为底的对数,对应产生的电压越高表示灵敏度越大,一般斜率为59mv/log[NH 4+]。3、线性度:在不同拉伸强度下(0%-40%),传感器保持氨离子浓度从10 -4M到1M都能得到较好的线性响应;同时电极电位随离子活度的变化符合具有线性特性的Nerstian方程。4、稳定度:当氨离子浓度从10 -4M到0.1M变化多个周期时,电压输出能保持稳定的相应周期性变化;稳定程度还受皮肤温度的影响,在运动的不同阶段,人体皮肤表面温度会从20℃左右到37℃之间变化,这会较大地影响氨离子浓度,从而体现在电压的不一致。5、电阻抗光谱。
现有以3D rGO-CNT为代表的WSES设计技术存在以下缺点:
第一、设计过程最重要的设计参数调整过程过分依赖设计工程师的经验,若工程师能力不足,则最终形成的各项性能指标将不可避免下降;即使设计工程师拥有丰富经验,“调整-验证”过程需要反复多次进行,需要设计工程师全程参与, 直到各项性能指标得到满足,这非常耗时耗力,最终结果也无法保证为理论最优值。
第二、设计关键和难度在于引入了不同的拉伸强度。在某种拉伸强度下获得了最优的一组设计参数,当拉伸强度发生变化时WSES的各项性能指标可能会发生较大干扰甚至急剧恶化,所以同一组设计参数很难兼顾所有拉伸强度,容易造成WSES整体性能不佳。同时,对于不同拉伸强度,需要增加相应实验次数,导致成本和时间都急剧上升。不同拉伸强度与性能指标的变化呈非线性关系,采用简单的线性回归模型无法保证性能预测与仿真的准确性,导致分析难度高,泛化性能不足。
第三、采用传统的计算智能优化方法来优化WSES的设计,可以避免设计工程师的全程参与,部分实现自动化设计,但会遇到下列问题:1、计算智能基于进化种群的反复迭代,每次迭代均需要种群所有进化个体的适应度函数值,该适应度函数值的计算离不开WSES的性能指标值,必须通过图1的设备经过一定时间的电化学实验完成。由于主要设计参数如何影响性能指标的机理尚不明确,目前这样的实验尚无软件仿真实现。如何主要设计参数的细小变化,都会导致新的实验和重新测量新的性能指标。在计算智能的反复迭代可能重复进行逾数万次,完成这个数量级的电化学实验是完全不现实的。2、传统计算智能方法大多以单目标优化为主,而WSES具有多个性能指标,并且这些性能指标往往互相冲突。若仅用单目标优化,难以同时改进多个性能指标,导致整体性能不佳;若使用多目标优化,计算复杂度将显著增加,并且导致完全不合理的实验次数。
基于现有WSES在设计上所存在的上述问题,本发明实施例提供了一种提升可穿戴可拉伸电化学传感器检测性能的设计方法,其中,如图3所示,包括步骤:
S10、使用正交化试验设计取得设计参数矩阵内插组合,通过电化学实验得到WSES在不同拉伸强度下的性能指标;
S20、根据所述性能指标建立与每个性能指标对应的非线性回归模型;
S30、使用多目标优化方法同时搜索性能指标的Pareto非支配最优解,获得WSES设计结果;
S40、根据所述WSES设计结果对所述可穿戴可拉伸电化学传感器进行优化 设计。
本实施例以拉伸强度作为进化个体的第N+1维变量参与了整个多目标优化过程,这能够兼顾其它设计参数和拉伸强度之间的关系,使得处于非支配地位的最优个体们能够充分反映不同拉伸强度对WSES性能的影响,保证了在不同拉伸强度下的性能最优,这解决了现有WSES设计方法对拉伸强度的鲁棒性不足问题。
在一些实施方式中,如图4所示,所述使用正交化试验设计取得离散程度可变、泛化能力更强的设计参数矩阵内插组合,通过电化学实验得到WSES在不同拉伸强度下的性能指标,以及根据所述性能指标建立与每个性能指标对应的非线性回归模型的步骤具体包括:
设定WSES的设计参数,构成N维设计参数矢量X=[x 1,x 2,…,x n,…,x N],x n∈X,所述设计参数可从上述表1中选取,作为举例,所述设计参数包括电沉积电压、电沉积时间、REM的用量、PVB的用量、甲醇氯化钠的用量、PU的用量、PDMS的用量、以及ISM的用量等,但不限于此;
设定M维性能指标矢量Y=[y 1,y 2,…,y m,…,y M],y m∈Y,作为举例,所述性能指标包括检测范围、线性度、稳定性、电阻抗光谱、灵敏度等具体数值,但不限于此;
根据ODE设计,计算设计参数的内插组合,构成设计参数矩阵,X={X 1,X 2,…,X k,…,X K},如图5所示,对于图5矩阵中的第n个参数,从x 1n到x kn,参数的离散程度需要根据经验进行合理设定。考虑到实验次数应该控制在合理范围内,K取值不应过大,一般控制在K<10;
使用图1所示的设备,采用WSES进行实验,获得任意设计参数组合X k∈X所对应的性能指标矢量Y k,生成性能指标矩阵Y={Y 1,Y 2,…,Y k,…,Y K},如图6所示。
对于任意设计参数矢量X k,使用不同拉伸强度l,l∈[l,L],L为最大拉伸强度,通过WSES实验生成不同拉伸强度性能指标矩阵
Figure PCTCN2019091799-appb-000019
如图7所示,其 中
Figure PCTCN2019091799-appb-000020
y km,l表示采用第k组设计参数,在拉伸强度l条件下获得的第m个性能指标;
对于第m个性能指标,设计参数矢量X k和不同拉伸强度性能指标矢量
Figure PCTCN2019091799-appb-000021
构成回归数据对
Figure PCTCN2019091799-appb-000022
如图8所示,综合k从1到K的所有设计参数组合,得到第m个性能指标的回归训练数据集Φ m={Φ 1m2m,…,Φ km,…Φ Km},重复回归训练过程,直到形成所有M个性能指标的回归训练数据集Φ={Φ 12,…,Φ m,…Φ M};对于任意Φ m,使用非线性回归模型中的支持向量机或者极限学习机估计出回归模型R m,重复这过程直到估计出所有M个性能指标的回归模型;集成所有性能指标的回归模型构成多模型集合R={R 1,R 2,…,R m,…R M}。
在本实施例中,除了最初的L种拉伸强度、K组N维设计参数(总计N×K×L个)需要手工设定和对应的电化学实验,整个设计过程已经交给计算机自动实现。实验次数完全可以控制在合理范围内,一般情况下,拉伸强度可选0%、10%、20%、30%和40%共5种,即L=5;设计参数选择8种,即N=8;不同设计参数取5组,即K=5,此时实验次数仅为200次。因为采用非线性回归模型估算出WSES的多性能指标,使得该模型可以仿真不同拉伸强度变化间距为1%或者更小,避免将L设定为40,显著降低了实验次数;同时该模型在K组不同设计参数之间仿真出远超过K组的内插参数,这两种由回归模型带来的仿真性能提升显著减少了实验以及人力设备成本。高效的非线性回归模型保证了性能估计的准确性,有效减少甚至避免了设计偏差。
在一些实施方式中,如图9所示,所述步骤S30、使用多目标优化方法同时搜索性能指标的Pareto非支配最优解,获得WSES设计结果,具体包括:
S31、初始化迭代计数器g=0,设最大迭代次数G,初始化Pareto非支配集合P ND为空集;
S32、初始化用于多目标优化的进化种群ps,其中每个个体为N+1维矢量
Figure PCTCN2019091799-appb-000023
|ps|是进化种群的个体数量,其中,所述
Figure PCTCN2019091799-appb-000024
每维的取值为
Figure PCTCN2019091799-appb-000025
其中rand(a,b)表示返回[a,b]范围内服从均匀分布的随机数,max(x n)和min(x n)分别表示第n个设计参数的最大值和最小值,L max表示WSES能够承受的最大拉伸强度;
S33、计算
Figure PCTCN2019091799-appb-000026
对应的性能指标,其包括步骤:初始化计数器m=0,初始化其性能指标的M维矢量
Figure PCTCN2019091799-appb-000027
为空;应用回归模型R m,估计出当前输入
Figure PCTCN2019091799-appb-000028
时的输出
Figure PCTCN2019091799-appb-000029
更新m=m+1,若m<M,则重新应用回归模型R m,估计出当前输入
Figure PCTCN2019091799-appb-000030
时的输出
Figure PCTCN2019091799-appb-000031
直至m≥M时,使用非支配排序遗传算法II更新进化种群ps。
S34、使用非支配排序遗传算法II更新进化种群ps;
S35、计算ps及P ND中所有个体的位置关系以及Pareto前端,选择其中处于非支配地位的个体,更新为新的P ND
S36、更新g=g+1,若g<G,则返回步骤S33;若g≥G,则结束更新,此时P ND中每个个体都成为Pareto前端上的优化设计,根据所述WSES设计结果对所述可穿戴可拉伸电化学传感器进行优化设计。
本实施例提供的方法能够实现通过单独运行形成多个Pareto非支配集最优方案,其性能偏重各有不同。若WSES设计要求有所改变,则不必重新实验和模型计算,只需在非支配最优集合中另外选择符合要求的一组设计参数即可。
本实施例提供的方法以适用于检测人体汗液氨离子浓度的传感器,当传感器的材料改变用于检测人体汗液的血糖(glucose)、乳酸(lactate)、钠(Na+)离子或者钾离子(K+)等含量时,本方法同样适用于传感器的优化设计,只需要修改多性能指标矩阵以及不同拉伸强度多性能指标矩阵。
综上所述,本发明提供的提升可穿戴可拉伸电化学传感器检测性能的设计方法,以拉伸强度作为进化个体的第N+1维变量参与了整个多目标优化过程,这 能够兼顾其它设计参数和拉伸强度之间的关系,使得处于非支配地位的最优个体们能够充分反映不同拉伸强度对WSES性能的影响,保证了在不同拉伸强度下的性能最优,这解决了现有WSES设计方法对拉伸强度的鲁棒性不足问题。
应当理解的是,本发明的应用不限于上述的举例,对本领域普通技术人员来说,可以根据上述说明加以改进或变换,所有这些改进和变换都应属于本发明所附权利要求的保护范围。

Claims (7)

  1. 一种提升可穿戴可拉伸电化学传感器检测性能的设计方法,其特征在于,包括步骤:
    使用正交化试验设计取得设计参数矩阵内插组合,通过电化学实验得到WSES在不同拉伸强度下的性能指标;
    根据所述性能指标建立与每个性能指标对应的非线性回归模型;
    使用多目标优化方法同时搜索性能指标的Pareto非支配最优解,获得WSES设计结果;
    根据所述WSES设计结果对所述可穿戴可拉伸电化学传感器进行优化设计。
  2. 根据权利要求1所述提升可穿戴可拉伸电化学传感器检测性能的设计方法,其特征在于,所述使用正交化试验设计取得离散程度可变、泛化能力更强的设计参数矩阵内插组合,通过电化学实验得到WSES在不同拉伸强度下的性能指标的步骤包括:
    设定WSES的设计参数,构成N维设计参数矢量X=[x 1,x 2,…,x n,…,x N],x n∈X,设定M维性能指标矢量Y=[y 1,y 2,…,y m,…,y M],y m∈Y;
    根据ODE设计,计算设计参数的内插组合,构成设计参数矩阵X={X 1,X 2,…,X k,…,X K};
    采用WSES进行实验,获得任意设计参数组合X k∈X所对应的性能指标矢量Y k,生成性能指标矩阵Y={Y 1,Y 2,…,Y k,…,Y K};
    对于任意设计参数矢量X k,使用不同拉伸强度l,l∈[1,L],L为最大拉伸强度,通过WSES实验生成不同拉伸强度性能指标矩阵
    Figure PCTCN2019091799-appb-100001
    其中
    Figure PCTCN2019091799-appb-100002
    y km,l表示采用第k组设计参数,在拉伸强度l条件下获得的第m个性能指标。
  3. 根据权利要求2所述提升可穿戴可拉伸电化学传感器检测性能的设计方法,其特征在于,所述性能指标包括检测范围、线性度、稳定性和电阻抗光谱。
  4. 根据权利要求2所述提升可穿戴可拉伸电化学传感器检测性能的设计方 法,其特征在于,根据所述性能指标建立与每个性能指标对应的非线性回归模型的步骤包括:
    对于第m个性能指标,设计参数矢量X k和不同拉伸强度性能指标矢量
    Figure PCTCN2019091799-appb-100003
    构成回归数据对
    Figure PCTCN2019091799-appb-100004
    综合k从1到K的所有设计参数组合,得到第m个性能指标的回归训练数据集Φ m={Φ 1m2m,…,Φ km,…Φ Km},重复回归训练过程,直到形成所有M个性能指标的回归训练数据集Φ={Φ 12,…,Φ m,…Φ M};
    对于任意Φ m,使用非线性回归模型中的支持向量机或者极限学习机估计出回归模型R m,重复这过程直到估计出所有M个性能指标的回归模型;
    集成所有性能指标的回归模型构成多模型集合R={R 1,R 2,…,R m,…R M}。
  5. 根据权利要求1所述提升可穿戴可拉伸电化学传感器检测性能的设计方法,其特征在于,所述使用多目标优化方法同时搜索性能指标的Pareto非支配最优解,获得WSES设计结果的步骤包括:
    初始化迭代计数器g=0,设最大迭代次数G,初始化Pareto非支配集合P ND为空集;
    初始化用于多目标优化的进化种群ps,其中每个个体为N+1维矢量
    Figure PCTCN2019091799-appb-100005
    |ps|是进化种群的个体数量;
    计算
    Figure PCTCN2019091799-appb-100006
    对应的性能指标;
    使用非支配排序遗传算法II更新进化种群ps;
    计算ps及P ND中所有个体的位置关系以及Pareto前端,选择其中处于非支配地位的个体,更新为新的P ND
    更新g=g+1,若g<G,则返回计算
    Figure PCTCN2019091799-appb-100007
    对应的性能指标的步骤;若g≥G,则结束更新,此时P ND中每个个体都成为Pareto前端上的优化设计。
  6. 根据权利要求5所述提升可穿戴可拉伸电化学传感器检测性能的设计方法,其特征在于,所述
    Figure PCTCN2019091799-appb-100008
    每维的取值为
    Figure PCTCN2019091799-appb-100009
    其中rand(a,b)表示返回[a,b]范围内服从均匀分布的随机数,max(x n)和min(x n)分别表示第n个设计参数的最大值和最小值,L max表示WSES能够承受的最大拉伸强度。
  7. 根据权利要求5所述提升可穿戴可拉伸电化学传感器检测性能的设计方法,其特征在于,所述计算
    Figure PCTCN2019091799-appb-100010
    对应的性能指标包括步骤:
    初始化计数器m=0,初始化其性能指标的M维矢量
    Figure PCTCN2019091799-appb-100011
    为空;
    应用回归模型R m,估计出当前输入
    Figure PCTCN2019091799-appb-100012
    时的输出
    Figure PCTCN2019091799-appb-100013
    更新m=m+1,若m<M,则重新应用回归模型R m,估计出当前输入
    Figure PCTCN2019091799-appb-100014
    时的输出
    Figure PCTCN2019091799-appb-100015
    直至m≥M时,使用非支配排序遗传算法II更新进化种群ps。
PCT/CN2019/091799 2019-06-19 2019-06-19 一种提升可穿戴可拉伸电化学传感器检测性能的设计方法 WO2020252673A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
PCT/CN2019/091799 WO2020252673A1 (zh) 2019-06-19 2019-06-19 一种提升可穿戴可拉伸电化学传感器检测性能的设计方法
CN201980000860.6A CN110799976A (zh) 2019-06-19 2019-06-19 一种提升可穿戴可拉伸电化学传感器检测性能的设计方法

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2019/091799 WO2020252673A1 (zh) 2019-06-19 2019-06-19 一种提升可穿戴可拉伸电化学传感器检测性能的设计方法

Publications (1)

Publication Number Publication Date
WO2020252673A1 true WO2020252673A1 (zh) 2020-12-24

Family

ID=69448486

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/091799 WO2020252673A1 (zh) 2019-06-19 2019-06-19 一种提升可穿戴可拉伸电化学传感器检测性能的设计方法

Country Status (2)

Country Link
CN (1) CN110799976A (zh)
WO (1) WO2020252673A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117530504A (zh) * 2023-12-05 2024-02-09 东莞市佳铠精密金属制品有限公司 一种电子烟外壳的制作方法及系统

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114518864B (zh) * 2022-02-17 2023-04-07 中国工程物理研究院电子工程研究所 一种柔性可拉伸光学随机数生成器、制备及使用方法

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8332342B1 (en) * 2009-11-19 2012-12-11 The United States of America as represented by the Administrator of the National Aeronautics & Space Administration (NASA) Model-based prognostics for batteries which estimates useful life and uses a probability density function
CN103246757A (zh) * 2013-01-15 2013-08-14 河海大学常州校区 一种压缩弹簧的优化设计方法
CN105548318A (zh) * 2016-01-25 2016-05-04 深圳大学 一种可穿戴电化学传感器电极及可穿戴电化学传感器
CN105740212A (zh) * 2016-02-02 2016-07-06 天津大学 一种基于正则化向量自回归模型的传感器异常检测方法

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103076031B (zh) * 2013-01-04 2015-04-22 青岛大学 一种柔性可拉伸应变式传感器的制备方法
CN104239634B (zh) * 2014-09-12 2017-06-20 深圳市欧克蓝科技有限公司 一种天线设计方法

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8332342B1 (en) * 2009-11-19 2012-12-11 The United States of America as represented by the Administrator of the National Aeronautics & Space Administration (NASA) Model-based prognostics for batteries which estimates useful life and uses a probability density function
CN103246757A (zh) * 2013-01-15 2013-08-14 河海大学常州校区 一种压缩弹簧的优化设计方法
CN105548318A (zh) * 2016-01-25 2016-05-04 深圳大学 一种可穿戴电化学传感器电极及可穿戴电化学传感器
CN105740212A (zh) * 2016-02-02 2016-07-06 天津大学 一种基于正则化向量自回归模型的传感器异常检测方法

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
TANG, WEI: "Sensor dynamic modeling based on genetic optimization LS-SVM algorithm", ZIDONGHUA-YIBIAO = PROCESS AUTOMATION INSTRUMENTATION, ZHONGGUO YIQI YIBIAO XUEHUI GUOCHENG JIANCE KONGZHI YIBIAO XUEHUI [ET.AL.], SHANGHAI, vol. 32, no. 3, 20 March 2011 (2011-03-20), Shanghai, XP055767321, ISSN: 1000-0380 *
X U KAI, WUL YIN, HUAXIANG WANG: "EMT Sensor Optimization Based on Analysis of Sensitivity Matrix", JOURNAL OF TEST AND MEASUREMENT TECHNOLOGY, vol. 25, no. 1, 15 January 2011 (2011-01-15), XP055767324 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117530504A (zh) * 2023-12-05 2024-02-09 东莞市佳铠精密金属制品有限公司 一种电子烟外壳的制作方法及系统

Also Published As

Publication number Publication date
CN110799976A (zh) 2020-02-14

Similar Documents

Publication Publication Date Title
WO2020252673A1 (zh) 一种提升可穿戴可拉伸电化学传感器检测性能的设计方法
CN108095716A (zh) 一种基于置信规则库和深度神经网络的心电信号检测方法
CN105259331B (zh) 一种节理岩体单轴强度预测方法
Ozer et al. Low-cost Internet of Things (IoT)-enabled a wireless wearable device for detecting potassium ions at the point of care
Zhong et al. Deep extended hazard models for survival analysis
CN112508063A (zh) 基于增量学习的医学图像分类方法
Tian et al. Pushing OECTs toward wearable: development of a miniaturized analytical control unit for wireless device characterization
CN113315593A (zh) 一种基于flom协方差矩阵和lstm神经网络的频谱感知算法
Mayrink et al. Sparse latent factor models with interactions: Analysis of gene expression data
Liu et al. Machine-learning assisted handwriting recognition using graphene oxide-based hydrogel
Zec et al. Specialized federated learning using a mixture of experts
Kharoubi et al. The cluster correlation-network support vector machine for high-dimensional binary classification
CN101853426A (zh) 普适计算环境中基于神经网络的上下文融合方法
Basterrech et al. Evolutionary Echo State Network: A neuroevolutionary framework for time series prediction
Zhang et al. CeSpGRN: Inferring cell-specific gene regulatory networks from single cell multi-omics and spatial data
Gao et al. High-dimensional variable selection in regression and classification with missing data
Ackleh et al. Competitive exclusion and coexistence for a quasilinear size-structured population model
CN110866403A (zh) 基于卷积循环实体网络的端对端对话状态跟踪方法及系统
Xu et al. Reinforced Knowledge Distillation for Time Series Regression
CN112965595B (zh) 一种神经电刺激模拟触觉的建模和预测方法
Park et al. Regional quantile regression for multiple responses
CN116319378A (zh) 基于深度学习的网络流量矩阵估计、模型训练方法和系统
Klaassen et al. Uniform inference in high-dimensional gaussian graphical models
Monleón-Getino et al. Shannon entropy ratio, a bayesian biodiversity index used in the uncertainty mixtures of metagenomic populations
Li et al. sEMG pattern recognition based on GRNN and Adaboost

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: 19934250

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: 19934250

Country of ref document: EP

Kind code of ref document: A1

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 23.06.2022)

122 Ep: pct application non-entry in european phase

Ref document number: 19934250

Country of ref document: EP

Kind code of ref document: A1