Novel electronic tongue based on voltammetry and Internet of Things

Abstract

The invention relates to the technical field of detection equipment, and discloses a novel electronic tongue based on voltammetry and Internet of Things, which comprises an electronic tongue body, an Internet of Things platform, a database and a mobile phone APP, the electronic tongue body comprises an integrated top cover, the bottom of the integrated top cover is detachably connected with an electrolytic tank, a circuit module is arranged in the integrated top cover, an auxiliary electrode, a reference electrode and a working electrode are arranged in the electrolytic tank, and an opening for connecting the auxiliary electrode, the reference electrode and the working electrode with the circuit module is formed in the bottom of the integrated top cover. The system has the advantages of low cost, high portability, high detection speed, high accuracy, convenience and easiness in operation and the like, can meet related requirements of various fields and various crowds on substance detection and characteristic analysis, and can also meet various requirements of real-time detection and remote monitoring.

Description

Technical Field

[0001] This invention relates to the field of detection equipment technology, specifically a novel electronic tongue based on the voltammetry method and the Internet of Things. Background Technology

[0002] The concept of electronic tongues originated in the 1980s. It's a sensor technology that simulates the human taste system, primarily used for the chemical composition analysis and identification of liquid samples. It uses multiple sensor arrays to detect chemical components in liquids and combines this with pattern recognition algorithms such as principal component analysis and artificial neural networks to process the collected data, thereby achieving the differentiation and identification of different liquid components. With the rapid development of sensor technology and artificial intelligence, electronic tongue technology has gradually moved from the laboratory to practical applications. Early electronic tongues were mainly based on potentiometric methods and ion-selective electrodes, but later evolved into multi-sensor fusion technology, combining electrochemical, optical, and mass spectrometry detection methods, showing broad application prospects in fields such as the food industry, environmental monitoring, and pharmaceutical research and development.

[0003] In the practical implementation of electronic tongue technology, the design of the sensor array and the data processing algorithm are crucial. In recent years, there has been considerable research and improvement in these areas. The following are the technical conditions closely related to the improvements in this invention:

[0004] Traditional electronic tongue sensor arrays typically employ a single type of sensor (such as a potential sensor), which limits their ability to analyze complex samples. In recent years, researchers have proposed the concept of multimodal sensor arrays, integrating multiple types of sensors (such as combining electrochemical and optical sensors) within the same system to improve the comprehensiveness and accuracy of detection. Data processing in electronic tongues usually relies on pattern recognition algorithms, such as principal component analysis (PCA), linear discriminant analysis (LDA), and artificial neural networks (ANNs). In recent years, the introduction of deep learning algorithms has significantly enhanced the data analysis capabilities of electronic tongues. By incorporating deep learning, electronic tongue algorithms can more accurately identify the taste characteristics of complex samples.

[0005] Although electronic tongues have application potential in fields such as the food industry and environmental monitoring, the existing technology still has the following shortcomings that urgently need to be addressed:

[0006] On the one hand, multi-source sensor data fusion efficiency is low. Existing electronic tongue systems employ a discrete data processing architecture, and the varying response speeds of sensors can cause significant timing synchronization errors, resulting in reduced data fusion accuracy. This makes traditional principal component analysis (PCA) insufficient for feature extraction from heterogeneous sensor data, leading to poor data synergy among multimodal sensors, especially when more than three types of sensors are used. Secondly, due to the different types of sensors, electronic tongues developed to improve detection speed and sensitivity often use specifically modified and unique bioelectrochemical sensors. This limits the scalable application areas of such electronic tongues, allowing them to perform well only in specific fields. On the other hand, existing electronic tongue devices lack intelligent IoT capabilities. Currently, mainstream electronic tongue systems, both domestically and internationally, only support USB wired data transmission, requiring the device to be deployed close to a host computer. Furthermore, the lack of a unified database for management makes them unsuitable for distributed industrial environments and fails to meet the real-time monitoring needs of industrial IoT scenarios. In addition, high R&D and maintenance costs further limit their adoption by small and medium-sized enterprises and ordinary consumers, resulting in a narrow market application scope and hindering widespread civilian adoption to meet the needs of diverse scenarios. Summary of the Invention

[0007] The technical problem that this invention aims to solve is that existing electronic tongue source sensors have low data fusion efficiency, few scalable application areas, and cannot meet the real-time monitoring needs in industrial IoT scenarios.

[0008] To solve the above problems, the technical solution of the present invention is: a novel electronic tongue based on the voltammetry method and the Internet of Things, comprising an electronic tongue body, an Internet of Things platform and database, and a mobile APP;

[0009] The electronic tongue body includes an integrated top cover, the bottom of which is detachably connected to an electrolytic cell, which contains a built-in circuit module. The electrolytic cell is equipped with an auxiliary electrode, a reference electrode, and a working electrode. The bottom of the integrated top cover has an opening for connecting the auxiliary electrode, the reference electrode, and the working electrode to the circuit module.

[0010] The IoT platform uses the open-source EMQX framework combined with the ESP8266 module. Communication between the mobile APP and the STM32 microcontroller is achieved by subscribing to fixed topics. The ESP8266 sends the collected data using a circular buffer method and listens for EMQX communication information through a server. The collected data is read and written to the database using the Flask framework as the backend program.

[0011] The database contains four entities: user, solution, configuration, and collected data, with each entity having its id as the primary key.

[0012] The mobile app's front-end is developed using the H5-based uniapp framework, while the back-end uses the Java technology stack. It supports running a single codebase across multiple platforms and utilizes the uView UI framework. Global state management is achieved through Vuex, ensuring compatibility with multiple platforms including iOS, Android, H5, and mini-programs. This guarantees clear and controllable data flow while ensuring the stability and scalability of the back-end services.

[0013] Furthermore, the integrated top cover is equipped with a Type-C power interface and an indicator light, which are used for power supply and operation status display of the device, respectively. The Type-C power interface adopts a package model of Type-C-SMD_Type-C-6P_5, and the indicator light is an LED light with a package size of LED-TH_BD3.9-P2.54-RD.

[0014] Furthermore, the circuit module includes a power supply circuit, a potentiostat circuit based on the LMP91000 chip, and an excitation and acquisition circuit based on the STM32 microcontroller and ESP8266 module.

[0015] Furthermore, the power supply circuit uses a Type-C interface with a package model of Type-C-SMD_Type-C-6P_5. The WE, RE, and CE pins of the LMP91000 are connected to the BAT WIRELESS RF connector used by the external electrolytic cell to eliminate environmental interference to the micro-current signal and ensure the accuracy of the acquisition.

[0016] The RXD and TXD pins of the ESP8266 module are connected to the PA9 and PA10 pins of the STM32 microcontroller, respectively, for data communication via serial port. The LMP91000 communicates with the STM32 microcontroller via I2C protocol to control the read and write operations of the LMP91000 chip. The SCL, SDA, and MEMB pins of the LMP91000 chip are connected to the PA7, PA6, and PA5 of the microcontroller, respectively.

[0017] The LMP91000 has two VREF input methods: one is to obtain a stable 2.5V voltage input through the LM4120AIM5X-2.5 voltage regulator module, and the other is to generate a specific excitation waveform through the PA15 pin of the STM32, and then input it after passing through the voltage follower LMV358 and low-pass filter.

[0018] A resistor R4 is placed between the C1 and C2 pins of the LMP91000 chip as an adjustable external amplification resistor. A capacitor C4 is set to further filter the excitation current. Finally, C2 is connected to the PA0 pin of the STM32 through a voltage follower LMV358 to use an ADC for data acquisition.

[0019] Furthermore, the attributes of each entity in the database are as follows:

[0020] User: User ID, Username, Password, Priority, Creation Time, where the "Priority" attribute serves as an identifier for different user types;

[0021] Solution: Solution ID, Category, Solution Name, Solution Concentration;

[0022] Configuration: Configuration ID, electrochemical method, scan frequency, electrode type. Among them, the "electrochemical method" attribute reflects the type of electrochemical analysis method used, which is essentially the category of the excitation waveform used.

[0023] Data collected: group ID, collection time, feature 1, feature 2... feature N, where the number of features N is related to the number of data collected per collection session set by the above-mentioned collection device.

[0024] Furthermore, the relationships between the entities are as follows:

[0025] Each user can use this product to identify multiple solutions, and each solution can also be identified by different users using this product. Therefore, the relationship between user entities and solution entities is many-to-many.

[0026] Different users can use various electrochemical analysis methods (i.e. different excitation signals) when using this product to identify solutions. Therefore, the relationships between user entities and configuration entities, and between configuration entities and solution entities, are many-to-many.

[0027] Each solution can be determined by multiple sets of collected data, and each set of collected data corresponds to only one type of solution. Therefore, the solution entity and the collected data entity have a one-to-many relationship.

[0028] Furthermore, principal component analysis was performed on the N-feature dataset, the results were saved and grouped, and n of the data groups were used as the training set.

[0029] The training set is used as training samples to input into the random forest algorithm model. GridSearchCV and RandomizedSearchCV are used to select the parameter combination with the highest accuracy from the predetermined parameter combination based on the training results.

[0030] The trained model is exported and deployed as the pattern recognition algorithm for this electronic tongue.

[0031] Furthermore, in the mobile app:

[0032] In terms of routing management, uniapp's built-in routing system, combined with custom route interception, maintains a consistent interface style and meets the needs of page navigation and access control. The network request module is uniformly encapsulated and integrates interceptors for easy unified processing of requests and responses.

[0033] In terms of backend architecture design, Spring Boot 2.x was adopted as the basic framework, combined with Spring Security and JWT to implement security authentication and ensure system security. The data persistence layer used MyBatis Plus, which simplified database operations, while Redis caching improved system performance.

[0034] In terms of page design, we followed design principles such as consistency, responsiveness, accessibility, and performance optimization. UI components were developed in strict accordance with specifications to ensure a consistent style across platforms while adapting to devices of different sizes.

[0035] In terms of page loading performance, user experience has been improved through code optimization, and component development follows a unified coding standard to ensure code readability and maintainability.

[0036] The design of the control logic mainly focuses on state management and request encapsulation. Vuex is used for global state management, and data such as user information and login status are stored centrally for easy sharing and access by various components. The request module is uniformly encapsulated, integrating request interception and response processing logic, simplifying the complexity of interface calls, and adding authentication information to the request header to ensure the security of interface calls.

[0037] The advantages of this invention compared to existing technologies are:

[0038] (1) By combining the LMP91000 chip with independent circuit design, the sensor signal processing is optimized, the signal-to-noise ratio is improved, the hardware complexity is reduced, and the system stability and consistency are enhanced.

[0039] (2) By utilizing the ESP8266 module, EMQX IoT messaging middleware, and a mobile app, wireless transmission of sensor data, cloud management, device control, and data visualization functions were achieved, supporting remote monitoring and interaction via the mobile app. This reduced the complexity of local devices, broke through the local usage limitations of traditional electronic tongue devices, and made the device more compact and convenient compared to traditional electronic tongues. It also lowered the user's usage and operation threshold, expanded the application scenarios of this product's electronic tongue, and enhanced the scalability and update convenience of the device.

[0040] (3) Based on the above structure, the electronic tongue of this product has greatly reduced the development cost, processing cost and operation difficulty compared with the traditional electronic tongue. The total cost of a single product hardware device and network service fee (excluding matching electrodes) is less than 350 yuan.

[0041] (4) When training and testing the model, the accuracy of the identification and prediction of the test substance reached 97.5%, and the model weighted loss was as low as 0.069. After clicking the "Start Detection" button on the mobile APP, the time required for signal stabilization, data acquisition, data preprocessing and identification model discrimination, and visualization display on the user's mobile APP varies from 60s to 180s depending on the order of magnitude of the collected features. In general, the product has a high recognition accuracy and fast detection speed.

[0042] This invention has the advantages of low cost, high portability, fast detection speed, high accuracy, and easy operation. It can meet the needs of various fields and groups of people for material detection and characteristic analysis, as well as the needs of various real-time detection and remote monitoring. Attached Figure Description

[0043] Figure 1 This is a structural diagram of the present invention.

[0044] Figure 2 This is the circuit schematic diagram of the present invention.

[0045] Figure 3 This is the ER diagram of the database conceptual design of this invention.

[0046] Figure 4 This is a schematic diagram illustrating the operation logic and functions of the mobile APP of this invention.

[0047] Figure 5 This is a flowchart of the operation process of the present invention.

[0048] Figure 6 This is the excitation waveform used in the experimental testing of this invention.

[0049] Figure 7 This is a graph showing the original collected data of the four test solutions of this invention.

[0050] Figure 8 This is a graph showing the change in OOB error during the training of the random forest algorithm model in this invention.

[0051] As shown in the figure:

[0052] Figure 1 In the diagram, 1a is the main view of the electronic tongue, 1b is a schematic diagram of the top interface of the integrated top cover, and 1c is a schematic diagram of the bottom opening of the integrated top cover.

[0053] Figure 2 In the diagram, 2a is the circuit diagram of LMP91000, 2b is the circuit diagram of STM32 microcontroller, 2c is the circuit diagram of three-electrode connection, 2d is the circuit diagram of ESP8266, 2e is the circuit diagram of power supply and ground, and 2f is the output circuit diagram. Detailed Implementation

[0054] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0055] As shown in the figure, a new type of electronic tongue based on the voltammetry method and the Internet of Things includes an electronic tongue body, an Internet of Things platform and database, and a mobile APP.

[0056] The electronic tongue body includes an integrated top cover, the bottom of which is detachably connected to an electrolytic cell, which contains a built-in circuit module. The electrolytic cell is equipped with an auxiliary electrode, a reference electrode, and a working electrode. The bottom of the integrated top cover has an opening for connecting the auxiliary electrode, the reference electrode, and the working electrode to the circuit module.

[0057] The IoT platform uses the open-source EMQX framework combined with the ESP8266 module. Communication between the mobile APP and the STM32 microcontroller is achieved by subscribing to fixed topics. The ESP8266 sends the collected data using a circular buffer method and listens for EMQX communication information through a server. The collected data is read and written to the database using the Flask framework as the backend program.

[0058] The database contains four entities: user, solution, configuration, and collected data, with each entity having its id as the primary key.

[0059] The mobile app's front-end is developed using the H5-based uniapp framework, while the back-end uses the Java technology stack. It supports running a single codebase across multiple platforms and utilizes the uView UI framework. Global state management is achieved through Vuex, ensuring compatibility with multiple platforms including iOS, Android, H5, and mini-programs. This guarantees clear and controllable data flow while ensuring the stability and scalability of the back-end services.

[0060] The workflow of an electronic tongue is as follows:

[0061] S1. Preparations:

[0062] 1) Clean the electrolytic cell and all electrodes, and dry them by drying or air-drying;

[0063] 2) Add the liquid to be tested into the device;

[0064] 3) Connect the physical objects according to the product connection relationships described above;

[0065] 4) Click the start button on the mobile app to forward and receive messages via the ESP8266 module + EMQX, and control the product to start working;

[0066] S2, Data Acquisition:

[0067] 1) Excitation signal generation: The excitation waveform (such as linear scanning waveform of the volt-ampere method, multi-frequency pulse waveform, etc.) is generated by programmably adjusting the combination of the PWM frequency and duty cycle of the STM32 microcontroller. The fitting formula for the analog signal voltage value is as follows: V avg =D×V CC ;

[0068] 2) Micro-current signal acquisition: The micro-current signal generated by the three-electrode electrolytic cell is amplified by the built-in gain resistor of LMP91000 and the external adjustable resistor, and then acquired by the built-in ADC of STM32.

[0069] S3, Data Upload:

[0070] 1) Collect data in groups of a fixed number N and send them to the ESP8266 module via serial port;

[0071] 2) The ESP8266 module forwards data via EMQX, and the collected data is uploaded to the database through server monitoring and backend program.

[0072] 3) Once the number of samples collected reaches the upper limit of a set of feature values, the device will begin the next collection.

[0073] S4. Model Training:

[0074] 1) Perform principal component analysis on the collected experimental data, select the optimal number of principal components based on the cumulative explanatory power of the principal components, and set the number of principal components to 3. Save the transformed data.

[0075] 2) The converted data is divided into training set and test set. The random forest algorithm is used to train the model on the training set to obtain the trained model. Then the test set is used as the input of the trained model to classify the sample solution and predict its concentration.

[0076] S5. Results Visualization:

[0077] The results processed by the algorithm recognition model are forwarded to the mobile device, and the recognition results are visualized through a mobile APP.

[0078] Use this product to identify four common pesticide diluent implementation examples.

[0079] First, sample solutions of four different mass fractions of pesticide emulsifiable concentrates were prepared as follows: malathion at 0.1 mg / L, 0.05 mg / L, and 0.025 mg / L; phoxim at 0.1 mg / L, 0.05 mg / L, and 0.025 mg / L; deltamethrin at 0.1 mg / L, 0.05 mg / L, and 0.025 mg / L; and cypermethrin at 0.1 mg / L, 0.05 mg / L, and 0.025 mg / L.

[0080] When preparing the solution, dilute it with 95% anhydrous ethanol. Use a pipette to transfer the diluted solution to a volumetric flask and make up to volume. The specific amount of solution and volumetric flask should be calculated using m = C * V (where m is the mass of the solution, C is the volume concentration of the solution, and V is the volume of the solution).

[0081] After the solution is prepared, the electrolytic cell and equipment need to be prepared: the electronic tongue uses a platinum disk electrode as the working electrode, a platinum column electrode as the auxiliary electrode, and an Ag / AgCl electrode as the reference electrode. Before use, the electrodes must be rinsed with distilled water (do not use surfactants to clean the electrodes) and then allowed to air dry in a ventilated place. The reference electrode must be cleaned and activated before use to remove impurities and contaminants. The electrolytic cell also needs to be cleaned, and it is essential to maintain the cleanliness of the electrolytic cell. The container must be dried before pouring in the test solution.

[0082] After completing the preparations, connect the electronic tongue, pour the solution to be tested into the beaker of the electronic tongue, ensure that the three electrodes are completely immersed in the solution, and connect the power supply through the Type-C port.

[0083] The selection is made via a mobile app, using a 10Hz positive linear scanning waveform as the excitation waveform (e.g., Figure 6 Each solution was tested multiple times, with three sets of data collected each time. Some of the raw data results are shown below. Figure 7 As shown. The cleaning steps in the preparation process need to be repeated after each data acquisition to avoid affecting subsequent experiments. During acquisition, the electronic tongue needs to wait approximately 10-20 seconds to ensure the excitation signal stabilizes. The ESP8622 acquires 10 bits of data each time, with a total of 4000 data points collected per group. This means that the database contains 4000 features describing each solution.

[0084] Principal component analysis was performed on the collected experimental data. The optimal number of principal components (3) was selected based on their cumulative explanatory power. The transformed data was then saved. The transformed data was divided into training and testing sets. A random forest algorithm was used to train the model on the training set. The test set was then used as input to the trained model, and the visualization results were returned via a mobile app.

[0085] The results were compared with those of other electronic tongues and similar benchmark methods on the market to verify the effectiveness of this method.

[0086] In terms of hardware design, the LMP91000 chip is combined with proprietary circuit design to optimize sensor signal processing, improve the signal-to-noise ratio, and solve the problems of signal drift and noise interference in traditional electronic tongue sensors. This also reduces hardware complexity, improves system stability and consistency, and extends device battery life. Secondly, regarding IoT and cloud collaboration, the ESP8266 module and EMQX IoT messaging middleware are used to achieve wireless transmission and cloud management of sensor data, supporting remote monitoring and interaction via a mobile app. This breaks through the local usage limitations of traditional electronic tongue devices and expands application scenarios. Finally, in terms of user interaction and scalability, the mobile app enables device control and data visualization, lowering the operational threshold. The modular design and standard communication protocols facilitate subsequent expansion of the sensor array and algorithm upgrades, enhancing the device's scalability and ease of updates.

[0087] Based on the voltammetric electronic tongue of this project, its non-specificity and low selectivity make it widely applicable in various fields such as food quality analysis, water quality monitoring, and medical diagnosis. Therefore, this invention conducts multi-level and multi-dimensional testing on complex liquid samples. Compared with other chemical analysis instruments and electronic tongue systems in the prior art, this invention preprocesses and extracts features from the electrochemical signals generated by different liquid samples under the excitation of a fixed waveform signal. Data acquisition is completed using an STM32 microcontroller, and the data is transmitted to a cloud database through an ESP8266 module and the EMQX IoT platform. Principal component analysis and an improved BP neural network algorithm are then used to model and classify the data, extracting key feature information of the liquid samples. Wireless communication is achieved through the ESP8266 module, and remote control and real-time monitoring can be realized with a mobile APP, ensuring that the system is more portable and intelligent, the sample recognition speed is faster, the results are more accurate and objective, and the production cost is lower.

[0088] The method of this invention can be extended to the classification and concentration prediction of more liquids (including other substances whose properties can be characterized by being made into liquids), taste evaluation and quality detection in various food and pharmaceutical industries. Although there are theoretical differences in the application prospects of different fields, which may lead to differences in the selected excitation waveform and some parameters in the recognition algorithm, and the waveform of the original data itself may have various differences in shape, or the number of features extracted from the database may change, the basic structure, basic design and actual operation of the hardware, database, recognition algorithm and mobile APP remain unchanged.

[0089] The parts not disclosed in this invention are all prior art, and their specific structures and working principles will not be described in detail.

[0090] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0091] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

[0092] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. A novel electronic tongue based on the voltammetry method and the Internet of Things, characterized in that: This includes the electronic tongue itself, an IoT platform and database, and a mobile app; The electronic tongue body includes an integrated top cover, the bottom of which is detachably connected to an electrolytic cell, which contains a built-in circuit module. The electrolytic cell is equipped with an auxiliary electrode, a reference electrode, and a working electrode. The bottom of the integrated top cover has an opening for connecting the auxiliary electrode, the reference electrode, and the working electrode to the circuit module. The IoT platform uses the open-source EMQX framework combined with the ESP8266 module. Communication between the mobile APP and the STM32 microcontroller is achieved by subscribing to fixed topics. The ESP8266 sends the collected data using a circular buffer method and listens for EMQX communication information through a server. The collected data is read and written to the database using the Flask framework as the backend program. The database contains four entities: user, solution, configuration, and collected data, with each entity having its id as the primary key. The mobile app's front-end is developed using the H5-based uniapp framework, while the back-end uses the Java technology stack. It supports running a single codebase across multiple platforms and utilizes the uView UI framework. Global state management is achieved through Vuex, ensuring compatibility with multiple platforms including iOS, Android, H5, and mini-programs. This guarantees clear and controllable data flow while ensuring the stability and scalability of the back-end services.

2. The novel electronic tongue based on the voltammetry and the Internet of Things as described in claim 1, characterized in that: The integrated top cover is equipped with a Type-C power interface and an indicator light, which are used for power supply and operation status display of the device, respectively. The Type-C power interface adopts a package model of Type-C-SMD_Type-C-6P_5, and the indicator light is an LED light with a package size of LED-TH_BD3.9-P2.54-RD.

3. The novel electronic tongue based on the voltammetry method and the Internet of Things as described in claim 1, characterized in that: The circuit module includes a power supply circuit, a potentiostat circuit based on the LMP91000 chip, and an excitation and acquisition circuit based on the STM32 microcontroller and ESP8266 module.

4. The novel electronic tongue based on the voltammetry and the Internet of Things as described in claim 3, characterized in that: The power supply circuit uses a Type-C interface with a package model of Type-C-SMD_Type-C-6P_5. The WE, RE, and CE pins of the LMP91000 are connected to the BAT WIRELESS RF connector used by the external electrolytic cell to eliminate environmental interference to the micro-current signal and ensure the accuracy of the acquisition. The RXD and TXD pins of the ESP8266 module are connected to the PA9 and PA10 pins of the STM32 microcontroller, respectively, for data communication via serial port. The LMP91000 communicates with the STM32 microcontroller via I2C protocol to control the read and write operations of the LMP91000 chip. The SCL, SDA, and MEMB pins of the LMP91000 chip are connected to the PA7, PA6, and PA5 of the microcontroller, respectively. The LMP91000 has two VREF input methods: one is to obtain a stable 2.5V voltage input through the LM4120AIM5X-2.5 voltage regulator module, and the other is to generate a specific excitation waveform through the PA15 pin of the STM32, and then input it after passing through the voltage follower LMV358 and low-pass filter. A resistor R4 is placed between the C1 and C2 pins of the LMP91000 chip as an adjustable external amplification resistor. A capacitor C4 is set to further filter the excitation current. Finally, C2 is connected to the PA0 pin of the STM32 through a voltage follower LMV358 to use an ADC for data acquisition.

5. A novel electronic tongue based on the voltammetry method and the Internet of Things according to claim 1, characterized in that, The attributes of each entity in the database are: User: User ID, Username, Password, Priority, Creation Time, where the "Priority" attribute serves as an identifier for different user types; Solution: Solution ID, Category, Solution Name, Solution Concentration; Configuration: Configure ID, electrochemical method, scan frequency, and electrode type. The "electrochemical method" attribute reflects the type of electrochemical analysis method used. Data collected: group ID, collection time, feature 1, feature 2... feature N, where the number of features N is related to the number of data collected per collection session set by the above-mentioned collection device.

6. A novel electronic tongue based on the voltammetry method and the Internet of Things as described in claim 5, characterized in that, The relationships between the entities are as follows: The relationship between the user entity and the solution entity is many-to-many; The relationships between the user entity and the configuration entity, and between the configuration entity and the solution entity, are many-to-many. The solution entity and the collected data entity have a one-to-many relationship.

7. A novel electronic tongue based on the voltammetry method and the Internet of Things as described in claim 6, characterized in that: Perform principal component analysis on the N-feature dataset, save the results and group them, and use n of the data groups as the training set. The training set is used as training samples to input into the random forest algorithm model. GridSearchCV and RandomizedSearchCV are used to select the parameter combination with the highest accuracy from the predetermined parameter combination based on the training results. The trained model is exported and deployed as the pattern recognition algorithm for this electronic tongue.

8. A novel electronic tongue based on the voltammetry method and the Internet of Things according to claims 1-7, characterized in that, The electronic tongue operates as follows: S1. Preparations: 1) Clean the electrolytic cell and all electrodes, and dry them by drying or air-drying; 2) Add the liquid to be tested into the device; 3) Connect the physical objects according to the product connection relationships described above; 4) Click the start button on the mobile app to forward and receive messages via the ESP8266 module + EMQX, and control the product to start working; S2, Data Acquisition: 1) Excitation signal generation: The excitation waveform is generated by programmably adjusting the combination of the PWM frequency and duty cycle of the STM32 microcontroller. The fitting formula for the analog signal voltage value is as follows: V avg =D×V CC ; 2) Micro-current signal acquisition: The micro-current signal generated by the three-electrode electrolytic cell is amplified by the built-in gain resistor of LMP91000 and the external adjustable resistor, and then acquired by the built-in ADC of STM32. S3, Data Upload: 1) Collect data in groups of a fixed number N and send them to the ESP8266 module via serial port; 2) The ESP8266 module forwards data via EMQX, and the collected data is uploaded to the database through server monitoring and backend program. 3) Once the number of samples collected reaches the upper limit of a set of feature values, the device will begin the next collection. S4. Model Training: 1) Perform principal component analysis on the collected experimental data, select the optimal number of principal components based on the cumulative explanatory power of the principal components, and set the number of principal components to 3. Save the transformed data. 2) The converted data is divided into training set and test set. The random forest algorithm is used to train the model on the training set to obtain the trained model. Then the test set is used as the input of the trained model to classify the sample solution and predict its concentration. S5. Results Visualization: The results processed by the algorithm recognition model are forwarded to the mobile device, and the recognition results are visualized through a mobile APP.

Images