High-precision measurement system and correction method
By introducing machine learning models and lookup tables into electronic measurement systems, the problem of difficulty in dealing with nonlinear relationships in the prior art is solved, and the high-precision measurement system correction effect is achieved.
Patent Information
- Application Number
- CN202311804856.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-12-26
- Publication Date
- 2025-06-27
AI Technical Summary
When the existing electronic measurement system corrects the actual output power and the set output power, it is difficult to effectively deal with the nonlinear relationship, resulting in poor correction effect.
Using a high-precision measurement system including data collection circuit, machine learning circuit and output circuit, a lookup table is established based on multiple set data, output data and correction parameters through the machine learning model, and the corresponding correction parameters are calculated to correct the output data.
Effective correction of nonlinear relationships is achieved, the accuracy of the measurement system is improved, and the actual output power is more consistent with the set output power.
Smart Images
Figure CN120218269A_ABST
Abstract
Description
Technical Field
[0001] This case relates to a high-precision measurement system and a calibration method, and particularly to a high-precision measurement system and a calibration method that perform calibration through a machine learning model. Background Art
[0002] Before an electronic device or electronic component leaves the factory, an electronic measurement system must be used to perform electrical measurements on the electronic device or electronic component. That is, the electronic measurement system will provide the electrical energy required by the electronic device or electronic component, and verify whether the function of the electronic device or electronic component is normal according to the feedback signal of the electronic device or electronic component.
[0003] However, due to factors such as hardware circuits, environmental temperature, and environmental humidity, there is an error between the actual output electrical energy and the set output electrical energy of the electronic measurement system. Therefore, it is necessary to calibrate the actual output electrical energy of the electronic measurement system by measuring the feedback signal so that the actual output electrical energy of the electronic measurement system is consistent with the set output electrical energy.
[0004] The conventional calibration method is to use the point-slope form to calibrate the actual output electrical energy when the actual output electrical energy and the set output electrical energy are linearly related. However, in practice, the relationship between the actual output electrical energy and the set output electrical energy is not linear. Therefore, if the point-slope form is used for calibration, the optimal calibration cannot be achieved. Summary of the Invention
[0005] The Summary of the Invention aims to provide a simplified summary of the present disclosure so that readers can have a basic understanding of the present disclosure. This Summary of the Invention is not a complete overview of the present disclosure, and its purpose is not to point out the important / critical elements of the embodiments of this case or to define the scope of this case.
[0006] One technical aspect of the content of this case discloses a high-precision measurement system. This high-precision measurement system includes a data collection circuit, a machine learning circuit, and an output circuit. The data collection circuit is used to obtain a plurality of first output data corresponding to a plurality of first set data. The machine learning circuit is coupled to the data collection circuit and is used to establish a machine learning model based on the plurality of first set data, the plurality of first output data, and a plurality of first calibration parameters between the plurality of first set data and the plurality of first output data, and is used to generate a second calibration parameter corresponding to the second set data according to the machine learning model. The output circuit is coupled to the machine learning circuit and is used to calibrate the second output data corresponding to the second set data according to the second calibration parameter to generate calibrated output data.
[0007] In some embodiments, it is characterized in that the plurality of first calibration parameters include a plurality of gain values and a plurality of offset values.
[0008] In some embodiments, it is characterized in that the machine learning model further includes a lookup table, and the lookup table is established based on the multiple first setting data, the multiple first output data, and the multiple first calibration parameters.
[0009] In some embodiments, it is characterized in that the machine learning circuit is further configured to obtain at least two of the multiple first calibration parameters from the lookup table according to the second setting data, and obtain the second calibration parameter by an interpolation method.
[0010] In some embodiments, it is characterized in that the data collection circuit is further configured to obtain multiple environmental data, and the machine learning circuit is further configured to establish the machine learning model based on the multiple first setting data, the multiple first output data, the multiple first calibration parameters, and the multiple environmental data.
[0011] Another technical aspect of the present case relates to a calibration method. This calibration method is applicable to a high-precision measurement system. The calibration method includes the following steps: obtaining multiple first setting data, multiple first output data, and multiple first calibration parameters between the multiple first setting data and the multiple first output data; establishing a machine learning model based on the multiple first setting data, the multiple first output data, and the multiple first calibration parameters; generating a second calibration parameter corresponding to the second setting data according to the machine learning model; and calibrating the second output data corresponding to the second setting data according to the second calibration parameter to generate a calibrated output data.
[0012] In some embodiments, it is characterized in that the multiple first calibration parameters include multiple gain values and multiple offset values.
[0013] In some embodiments, it is characterized in that it further includes: establishing a lookup table of the machine learning model based on the multiple first setting data, the multiple first output data, and the multiple first calibration parameters.
[0014] In some embodiments, it is characterized in that it further includes: obtaining at least two of the multiple first calibration parameters from the lookup table according to the second setting data, and obtaining the second calibration parameter by an interpolation method.
[0015] In some embodiments, it is characterized in that it further includes: obtaining multiple environmental data to establish the machine learning model based on the multiple first setting data, the multiple first output data, the multiple first calibration parameters, and the multiple environmental data. Description of the Drawings
[0016] To make the above and other purposes, features, advantages, and embodiments of the present case more obvious and understandable, the descriptions of the accompanying drawings are as follows:
[0017] Figure 1 It is a schematic diagram of a high-precision measurement system illustrated according to some embodiments of the present case;
[0018] Figure 2 It is a flowchart of a calibration method illustrated according to some embodiments of the present case;
[0019] Figure 3 It is a schematic diagram of a calibration method illustrated according to some embodiments of the present case;
[0020] Figure 4 It is a schematic diagram of a machine learning circuit illustrated according to some embodiments of the present case.
[0021] According to the usual operation mode, various features and components in the figure are not drawn to scale, and the drawing method is to present the specific features and components related to the present case in the best way. In addition, between different drawings, the same or similar component symbols are used to refer to similar components / parts.
[0022]
Symbol Explanation
[0023] 10: High-precision measurement system
[0024] 110: Data collection circuit
[0025] 130: Machine learning circuit
[0026] 150: Output circuit
[0027] VT: Output voltage
[0028] 900: Device under test
[0029] 200: Calibration method
[0030] S210, S230, S250, S270: Steps
[0031] G(I1, V2), G(I2, V2), G(I1, V1), G(I2, V1): Gain values
[0032] G(I*, V*), G1, G2: Gain values
[0033] I1, I2, I3, I4, I*: Current values
[0034] V1, V2, V3, V4, V*: Voltage values
[0035] 131: Analog-to-digital conversion circuit
[0036] 135: Lookup table processing circuit
[0037] 137: Calibration circuit Detailed Implementation Modes
[0038] To make the description of the present disclosure more detailed and complete, the following provides an illustrative description of the implementation aspects and specific embodiments of this case; however, this is not the only form of implementing or applying the specific embodiments of this case. The implementation modes cover the features of multiple specific embodiments and the method steps and their sequences for constructing and operating these specific embodiments. However, other specific embodiments can also be used to achieve the same or equivalent functions and step sequences.
[0039] Unless otherwise defined in this specification, the meanings of the scientific and technical terms used herein are the same as those understood and commonly used by those with ordinary knowledge in the technical field to which this case belongs. In addition, without conflict with the context, the singular nouns used in this specification cover the plural forms of the nouns; and when the plural nouns are used, they also cover the singular forms of the nouns.
[0040] In addition, regarding the "coupling" used herein, it can refer to two or more elements making direct physical or electrical contact with each other, or making indirect physical or electrical contact with each other, and can also refer to two or more elements operating or acting on each other.
[0041] Figure 1 is a schematic diagram of a high-precision measurement system 10 illustrated according to some embodiments of this case. As shown in the figure, the high-precision measurement system 10 includes a data collection circuit 110, a machine learning circuit 130, and an output circuit 150. In terms of the connection relationship, the machine learning circuit 130 is coupled to the data collection circuit 110, and the output circuit 150 is coupled to the machine learning circuit 130. The operation mode of the high-precision measurement system 10 will be described below with reference to Figure 2 be described.
[0042] It should be noted that the implementation modes of this case are not limited to Figure 1 the structure and operation shown, which is only used to illustratively show one of the implementation ways of this case to make the technology of this case easy to understand. The patent scope of this case shall be subject to the scope of the invention patent application. Those skilled in the art, without departing from the spirit of this case, the modifications and refinements made to the embodiments of this case still fall within the scope of the invention patent application of this case.
[0043] Figure 2 is a flowchart of a calibration method 200 illustrated according to some embodiments of this case. To make the calibration method 200 easy to understand, please refer to Figure 1 together, and the detailed process of the calibration method 200 will be introduced later.
[0044] In step S210, a plurality of first set data, a plurality of first output data, and a plurality of first calibration parameters between the plurality of first set data and the plurality of first output data are obtained. In some embodiments, step S210 is performed byFigure 1 is executed by the machine learning circuit 130 in Figure 1 .
[0045] Please refer to Figure 1 as well. In one embodiment, the output circuit 150 of the high-precision measurement system 10 generates and outputs an output voltage VT to the device under test 900 according to the first set data. For example, in one embodiment, a user operates an operation interface (not shown) of the high-precision measurement system 10 to set the first set data to 10 volts, expecting the output circuit 150 of the high-precision measurement system 10 to output an output voltage VT of 10 volts to the device under test 900. However, the actually output output voltage VT of the output circuit 150 may be different from the set data of 10 volts due to internal circuits or factors such as environmental temperature and humidity, that is, the actually output voltage is not 10 volts. The above output voltage VT is the output data generated and output corresponding to the set data.
[0046] For example, in one embodiment, the output voltage VT generated and output by the output circuit 150 according to the set data of 10 volts is 9.7 volts. The output voltage VT generated and output by the output circuit 150 according to the set data of 20 volts is 20.3 volts. The data collection circuit 110 collects the above set data and its corresponding output data (output voltage VT).
[0047] In some embodiments, after the high-precision measurement system 10 generates and outputs output voltages VT with multiple different voltage values according to multiple different first set data, the voltage values of the output voltages VT obtained by the data collection circuit 110 are obtained through an external circuit (not shown) and then fed back to the data collection circuit 110 via the external circuit.
[0048] Next, after the machine learning circuit 130 obtains the above multiple different set data and their corresponding multiple first output data, it calculates multiple first correction parameters between the set data with multiple different voltage values and their corresponding multiple first output data in the way of point-slope correction. In some embodiments, the multiple first correction parameters include multiple gain values and multiple offset values.
[0049] In some embodiments, after being corrected by the correction parameters, the generated corrected output data is the same as the set data. That is, when the set data is 10 volts, the corrected output data is also 10 volts.
[0050] In some embodiments, the calculation formula of point-slope correction is as follows in formula (1).
[0051] Formula (1): Y = AX + B.
[0052] In the above formula (1), Y is the corrected output data, X is the first output data, A is the gain value, and B is the offset value. Through point-slope correction, the output data can be corrected to the corrected output data.
[0053] The above formula (1) is only for illustrative purposes. In some other embodiments, the first correction parameter may include multiple correction formulas, multiple gain values, and multiple offset values.
[0054] In step S230, a machine learning model is established based on multiple first setting data, multiple first output data, and multiple first correction parameters. In some embodiments, step S230 is performed by Figure 1 the machine learning circuit 130 therein.
[0055] In some embodiments, the machine learning model includes a look-up table stored in a memory (not shown).
[0056] In one embodiment, the first correction parameters (including gain values and offset values) and the corrected output data calculated by the machine learning circuit 130 based on multiple different first setting data and their corresponding first output data are as shown in the following Look-up Table 1.
[0057] Look-up Table 1.
[0058]
[0059] The first output data in the above Look-up Table 1 are the multiple first setting data obtained by the data collection circuit 110 of the high-precision measurement system 10 and the first output data corresponding to the first setting data. After correction according to the above formula (1) through the gain value and the offset value, the first output data is corrected to the corrected output data so that the corrected output data is the same as the first setting data, as shown in the above Look-up Table 1.
[0060] In some embodiments, since the error between the first setting data and the first output data is non-linear, the first correction parameters corresponding to different first setting data and first output data are also different.
[0061] In another embodiment, the above Look-up Table 1 is the first setting data, the first output data, and their corresponding first correction parameters obtained when the output current of the high-precision measurement system 10 is the first current value (for example, 5 amperes). In another embodiment, the machine learning circuit 130 can establish another look-up table based on the first setting data, the first output data, and their corresponding first correction parameters obtained when the output current of the high-precision measurement system 10 is the second current value (for example, 10 amperes). That is to say, the setting data and their corresponding output data can be voltage values / current values / power values, etc. The above Look-up Table 1 is only illustrated by taking voltage values as an example.
[0062] In some embodiments, when establishing a machine learning model, the machine learning circuit 130 is further configured to generate a look-up table based on a plurality of environmental data obtained by the data collection circuit 110, a first calibration parameter (including a gain value and an offset value) calculated based on a plurality of different first setting data and their corresponding first output data, and the environmental data, so as to establish a machine learning model.
[0063] For example, it is assumed that the above-mentioned look-up table one is established under the conditions of an environmental temperature of 27 degrees Celsius and an environmental humidity of 70%. In the case of another environmental temperature and environmental humidity, the machine learning circuit 130 calculates a calibration parameter based on the first setting data and their corresponding first output data obtained under the conditions of the other environmental temperature and environmental humidity, and then establishes another look-up table.
[0064] In this way, by establishing a plurality of look-up tables, the machine learning circuit 130 establishes a machine learning model. In addition, the above-mentioned look-up table one is a look-up table of the output voltage VT. In some other embodiments, the look-up table established by the machine learning circuit 130 may be a look-up table established based on the setting data and output data of the output current of the high-precision measurement system 10, may be based on the setting data and output data of both the output current and output voltage of the high-precision measurement system 10, may be a look-up table established based on the setting data and output data of the input current of the high-precision measurement system 10, may be a look-up table established based on the setting data and output data of the input voltage of the high-precision measurement system 10, or may be a look-up table established based on the setting data and output data of both the input current and input voltage of the high-precision measurement system 10. The implementation manners of this case are not limited to the above.
[0065] In step S250, a second calibration parameter corresponding to the second setting data is generated according to the machine learning model. In some embodiments, step S250 is performed by Figure 1 the machine learning circuit 130 therein.
[0066] In some embodiments, in step S250, after establishing the machine learning model, the machine learning circuit 130 inputs the second setting data into the machine learning model to obtain a second calibration parameter corresponding to the second setting data. In some embodiments, the second setting data is data set and input into the high-precision measurement system 10 by the user through an operation interface (not shown).
[0067] In some embodiments, the machine learning circuit 130 obtains at least two of the first calibration parameters from the look-up table of the machine learning model according to the second setting data, and obtains the second calibration parameter by interpolation.
[0068] In one embodiment, the machine learning circuit 130 obtains at least two first setting data closest to the second setting data from the look-up table of the machine learning model according to the second setting data, and obtains the second calibration parameter in the manner of interpolation according to the first calibration parameters corresponding to the at least two first setting data.
[0069] For example, please refer to Look-up Table 1. If in step S250, the second setting data is 15 volts, then the first setting data obtained by the machine learning circuit 130 that is closest to the second setting data is 10 volts and 20 volts. Then, the machine learning circuit 130 can obtain the gain value corresponding to the second setting data of 15 volts by interpolation as And the offset value corresponding to the second setting data of 15 volts is
[0070] For another example, please also refer to Figure 3 . Figure 3 is a schematic diagram of a calibration method illustrated according to some embodiments of the present case. Figure 3 In it, G(I1, V2) is the gain value when the current value in the first setting data is I1 and the voltage value is V2, G(I2, V2) is the gain value when the current value in the first setting data is I2 and the voltage value is V2, G(I1, V1) is the gain value when the current value in the first setting data is I1 and the voltage value is V1, and G(I2, V1) is the gain value when the current value in the first setting data is I2 and the voltage value is V1.
[0071] The above gain values G(I1, V2), G(I2, V2), G(I1, V1), and G(I2, V1) are the gain values obtained by the machine learning circuit 130 according to the look-up table in the machine learning model, and the voltage values and current values in the above gain values G(I1, V2), G(I2, V2), G(I1, V1), and G(I2, V1) and the voltage value V* and current value I* in the gain value G(I*, V*) to be calculated are at least two closest ones in the look-up table. That is, the current values I1 and I2 are the current values closest to the current value I* in the look-up table, and the voltage values V1 and V2 are the voltage values closest to the voltage value V* in the look-up table.
[0072] Next, the machine learning circuit 130 obtains the gain value G(I * , V * )) of the second setting data (i.e., the current value I* and the voltage value V*) according to the above gain values G(I1, V2), G(I2, V2), G(I1, V1), and G(I2, V1) and the following formulas (2) to (5).
[0073] Formula (2):
[0074] Formula (3):
[0075] Formula (4):
[0076] Formula (5):
[0077] In step S270, the second output data corresponding to the second set data is corrected according to the second correction parameter to generate corrected output data. In some embodiments, step S270 is performed by Figure 1 the output circuit 150 in
[0078] In some embodiments, after the machine learning circuit 130 obtains the second correction data corresponding to the second set data, the output circuit 150 corrects the second output data according to the second correction data to generate corrected output data.
[0079] In some embodiments, the machine learning circuit 130 corrects the second output data in the point-slope formula (such as the above formula (1)).
[0080] For example, if the gain value corresponding to the second set data of 15 volts obtained by the machine learning circuit 130 in step S250 is and the offset value corresponding to the second set data of 15 volts is The machine learning circuit 130 calculates the corrected output data according to formula (1) as the following formula:
[0081] where Y2 is the corrected output data, and X2 is the second output data generated by the output circuit 150 corresponding to the second set data of 15 volts before correction. is the gain value, is the offset value. Through point-slope correction, the second output data can be corrected so that the corrected output data is the same as the set data.
[0082] In some embodiments, the output circuit 150 adjusts the second output data in a PWM (pulse width modulation) manner according to the corrected output data to generate corrected output data. All ways of adjusting the second output data are within the implementation manners of this case, and the implementation manners of this case are not limited to PWM.
[0083] It should be noted that this case is not limited to Figure 2The process shown is limited to merely illustratively depicting one implementation manner of this case, so as to make the technology of this case easy to understand. The patent scope of this case shall be subject to the scope of the invention patent application. Those skilled in the art, without departing from the spirit of this case, the modifications and refinements made to the embodiments of this case still fall within the scope of the invention patent application of this case.
[0084] Please refer to Figure 4 together. Figure 4 is a schematic diagram of a machine learning circuit 130 illustrated according to some embodiments of this case. As Figure 4 shown, the machine learning circuit 130 includes an analog-to-digital conversion circuit 131, a look-up table processing circuit 135, and a calibration circuit 137. The look-up table processing circuit 135 is coupled to the analog-to-digital conversion circuit 131, and the calibration circuit 137 is coupled to the look-up table processing circuit 135.
[0085] In some embodiments, the analog-to-digital conversion circuit 131 is used to convert the output data obtained by the data acquisition circuit 110 from analog data to digital data. The look-up table processing circuit 135 is used to input the output data converted into digital data into the machine learning model to obtain calibration parameters. In some embodiments, the look-up table processing circuit 135 performs interpolation calculation by the moving average method to obtain calibration parameters.
[0086] The calibration circuit 137 is used to generate a PWM control signal according to the calibration parameters, and transmit the PWM control signal to an output circuit 150 as Figure 1 shown, so that the output circuit 150 generates a calibrated output signal according to the PWM control signal.
[0087] In some embodiments, the high-precision measurement system 10 further includes a memory (not shown), which is used to store the machine learning model and the look-up table for access by the machine learning circuit 130.
[0088] From the above implementation manners of this case, it can be seen that applying this case has the following advantages. The high-precision measurement system and calibration method shown in the embodiments of this case calculate calibration parameters in the way of a machine learning model in response to changes in different environmental factors and circuit factors, so as to calibrate the output data into calibrated output data, making the actually output data of the high-precision measurement system the same as the set data. In addition, the implementation manner of this case can continuously learn, update, and correct calibration parameters in the way of a machine learning model. The above method can quickly obtain calibration parameters and effectively reduce calibration errors.
[0089] In some embodiments, the machine learning circuit 130 as described above can be integrated into an electronic device or an electronic system. In some embodiments, the server can be implemented as a cloud server and includes a central processing unit. In some embodiments, the server can be, but is not limited to, a single processor or a collection of multiple microprocessors, and includes a memory and an input / output circuit.
[0090] In some embodiments, the machine learning circuit 130 can be a central processing unit (CPU), a microcontroller unit (MCU), a digital signal processor (DSP), a field programmable gate array (FPGA), a server, or other computing circuits, processing circuits, or components having functions such as data access, data calculation, data storage, data transmission and reception, or similar functions.
[0091] In some embodiments, the machine learning circuit 130 includes a processor and an input / output circuit. The processor can be a circuit or component having functions such as data access, data calculation, data storage, or similar functions. The input / output circuit can be a circuit or component having functions such as data transmission and reception or similar functions.
[0092] In some embodiments, the data collection circuit 110 can be a current detection circuit, a voltage detection circuit, a temperature detection circuit, a humidity detection circuit, or other circuits or components having functions such as data detection, data transmission and reception, or similar functions. In some embodiments, the data collection circuit 110 obtains input data and environmental data via an external current detection circuit, voltage detection circuit, temperature detection circuit, humidity detection circuit, or other circuits or components having similar functions.
[0093] In some embodiments, the high-precision measurement system 10 further includes a display circuit (not shown) for displaying set data, output data, corrected output data, and the like.
[0094] In some embodiments, the output circuit 150 can be a PWM drive circuit or other circuits or components having current output / voltage output / power output or other identical or similar functions.
[0095] Although specific embodiments of the present case are disclosed in the above embodiments, they are not intended to limit the present case. Those of ordinary skill in the art to which the present case pertains can make various changes and modifications without departing from the principles and spirit of the present case. Therefore, the protection scope of the present case shall be subject to that defined by the appended patent application scope.
Claims
1. A high-precision measurement system, characterized in that, Comprising: A data collection circuit for obtaining a plurality of first output data corresponding to a plurality of first setting data; A machine learning circuit coupled to the data collection circuit for establishing a machine learning model based on the plurality of first setting data, the plurality of first output data, and a plurality of first calibration parameters between the plurality of first setting data and the plurality of first output data, and for generating a second calibration parameter corresponding to a second setting data based on the machine learning model; and An output circuit coupled to the machine learning circuit for correcting a second output data corresponding to the second setting data based on the second calibration parameter to generate a corrected output data.
2. The high-precision measurement system according to claim 1, wherein Wherein the plurality of first calibration parameters include a plurality of gain values and a plurality of offset values.
3. The high-precision measurement system according to claim 1, wherein Wherein the machine learning model further includes a look-up table, and the look-up table is established based on the plurality of first setting data, the plurality of first output data, and the plurality of first calibration parameters.
4. The high-precision measurement system according to claim 3, wherein Wherein the machine learning circuit is further configured to obtain at least two of the plurality of first calibration parameters from the look-up table based on the second setting data, and obtain the second calibration parameter by an interpolation method.
5. The high-precision measurement system according to claim 1, characterized in that, Wherein the data collection circuit is further configured to obtain a plurality of environmental data, and the machine learning circuit is further configured to establish the machine learning model based on the plurality of first setting data, the plurality of first output data, the plurality of first calibration parameters, and the plurality of environmental data.
6. A calibration method, applicable to a high-precision measurement system, characterized in that, Wherein the calibration method includes: Obtaining a plurality of first setting data, a plurality of first output data, and a plurality of first calibration parameters between the plurality of first setting data and the plurality of first output data; Establishing a machine learning model based on the plurality of first setting data, the plurality of first output data, and the plurality of first calibration parameters; Generating a second calibration parameter corresponding to a second setting data based on the machine learning model; and Correcting a second output data corresponding to the second setting data based on the second calibration parameter to generate a corrected output data.
7. The calibration method according to claim 6, wherein Wherein the plurality of first calibration parameters include a plurality of gain values and a plurality of offset values.
8. The calibration method according to claim 6, wherein Further comprising: Establishing a look-up table of the machine learning model based on the plurality of first setting data, the plurality of first output data, and the plurality of first calibration parameters.
9. The calibration method according to claim 8, wherein Further comprising: Obtaining at least two of the plurality of first calibration parameters from the look-up table based on the second setting data, and obtaining the second calibration parameter by an interpolation method.
10. The calibration method according to claim 6, wherein Further comprising: Obtaining a plurality of environmental data to establish the machine learning model based on the plurality of first setting data, the plurality of first output data, the plurality of first calibration parameters, and the plurality of environmental data.