A multi-sensor fusion intelligent filing monitoring handle and evaluation method

Abstract

A multi-sensor fusion intelligent file monitoring handle and evaluation method belong to the technical field of bench work processing and manufacturing, temperature sensing unit, humidity sensing unit, pressure sensing unit, pose sensing unit, display unit, vibration unit, wireless transmission unit, power supply unit and micro control unit are integrated in the file handle. The micro control unit transmits the data collected by each sensor unit to the upper computer, the multi-dimensional signal processing fusion algorithm is embedded in the upper computer, the comprehensive quality evaluation of the file processing is made, the threshold is triggered when the score is too low, the instruction is fed back to the intelligent handle, the vibration unit on the handle produces vibration to provide tactile warning feedback, and the display unit displays the adjustment suggestion given by the upper computer. The present application can solve the problem that the existing file monitoring function is single and does not have dynamic feedback and quality evaluation, realize the closed loop of detection, analysis and feedback in the file processing, introduce the multi-dimensional data fusion algorithm based on the file operation, and realize the scientific evaluation of the file quality.

Description

Technical Field

[0001] This invention belongs to the field of fitter's machining and manufacturing technology, and relates to an intelligent filing monitoring handle with multi-sensor fusion and an evaluation method. Background Technology

[0002] With the rapid development of aerospace and high-precision equipment, fitters remain the primary choice and play an irreplaceable role in the machining of many ultra-precision parts in high-end manufacturing. However, there is a significant shortage of highly skilled fitters. This is because training a high-level fitter is not only time-consuming and labor-intensive but also heavily reliant on the quality of the instructors. Currently, fitter training in schools largely relies on on-site demonstrations or experience transfer. This data-driven training model is inherently uncertain and hinders the identification and resolution of problems during the learning process.

[0003] Filing is one of the most important basic skills in fitter work, and the ability to master filing skills is an important dimension for evaluating a fitter's skill level. To improve the filing skills of practitioners, many improved file products have been developed. However, their functions are relatively limited, only monitoring a single variable such as file levelness or file angle, which cannot scientifically evaluate the quality of the filing process. Chinese invention patent 202110461706.8 discloses a fitter's filing posture detection and correction system. This patent monitors only a single physical quantity, and the device also includes a vise table, making the system bulky, cumbersome, and with low integration. Chinese invention patent 202111517722.0 discloses a file teaching system and method. The core of this patent is to collect the file angle through sensors, and the display end can observe the data and set labels to monitor multiple files. However, it monitors only a single physical quantity, and the evaluation of filing quality can only rely on the teacher's experience. Chinese utility model patent 202421945546.X discloses a file handle and file with intelligent level detection and alarm function. This utility model integrates a level detector and an alarm, but it can only monitor whether the file is in a level state, and its guidance on the filing process is not strong.

[0004] Therefore, there is an urgent need for an invention to solve the problems of the single physical quantity for monitoring the filing process, the subjective and one-sided evaluation of filing quality, the lack of closed-loop dynamic guidance, and the low degree of equipment integration. Summary of the Invention

[0005] To address the problems of existing technologies, this invention proposes a multi-sensor fusion intelligent filing monitoring handle and evaluation method. This invention uses sensors within the filing handle to collect multi-dimensional signals in real time, and a host computer performs multi-source data fusion and real-time analysis. Dynamic prompts are output through a display unit on the handle, achieving a closed-loop dynamic guidance system for monitoring, analysis, and feedback during the filing process. This solves the problems of limited physical quantity monitoring in filing equipment, subjective and one-sided filing quality evaluation, lack of closed-loop dynamic guidance, and low equipment integration in the prior art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A filing monitoring and evaluation method based on multi-sensor fusion is proposed. The evaluation method is embedded in a host computer, which is located on a PC. The filing monitoring and evaluation method includes the following steps: The first step involves the microcontroller receiving raw data from multiple sensors within the intelligent file monitoring handle, inserting delimiters between different data types, concatenating them into a single data frame, and then transmitting it to the host computer via a wireless transmission unit. The details are as follows: The data frames mainly include raw temperature data, raw humidity data, raw pressure data, and raw pose data. All data frames are raw data, and all data will be preprocessed and have features extracted in the host computer.

[0007] The second step involves the host computer splitting the data frames, as detailed below: Step 2.1: Extract the raw temperature data from the data frame and store it using a temporary temperature variable; Step 2.2: Extract the raw humidity data from the data frame and store it using a temporary humidity variable; Step 2.3: Extract the raw pressure data from the data frame and store it using a temporary pressure variable; Step 2.4: Extract the original pose data from the data frame and store it using a pose temporary variable.

[0008] The third step is to preprocess the raw data separated in the second step. The purpose of preprocessing is to eliminate interference from high-frequency impact vibrations and noise, thereby improving the accuracy of the condition assessment. Different preprocessing methods are used for different physical quantities according to their characteristics, as detailed below: Step 3.1: The raw temperature data is processed using a limited moving average filter; Step 3.2: The raw humidity data is processed using a limited moving average filter. Step 3.3: The raw pressure data is processed using Kalman filtering; Step 3.4: The original pose data is processed using Mahony complementary filtering.

[0009] The fourth step involves feature extraction from all the preprocessed raw data from the third step, yielding temperature, humidity, pressure, and pose feature data. This includes the following steps: Step 4.1: Extract the average temperature and temperature change rate by summing all temperature data points within the sampling period and taking the arithmetic mean. This reflects the overall heat load level during the filing process; the temperature derivative with respect to time is approximately calculated using the backward difference method to obtain the temperature change rate. A rapid temperature rise indicates excessive cutting strength or poor heat dissipation.

[0010] (1) (2) in, This represents the total number of sampling time points; This refers to the sampling time point; Let t be the temperature value at time t. The sampling time interval; The temperature at the previous moment.

[0011] Step 4.2: Extract average humidity and humidity variation range. The average humidity is obtained by summing all humidity data points within the sampling period and taking the arithmetic mean. The average humidity level of the filing environment is reflected; the difference between the maximum and minimum humidity values ​​during the sampling period is taken to obtain the humidity change range, which reflects the degree of fluctuation in environmental humidity.

[0012] (3) (4) in, Let be the humidity value at time t; This represents the range of humidity change. Maximum humidity; This represents the minimum humidity level.

[0013] Step 4.3: Extract the average pressure and pressure fluctuation coefficient by summing all pressure data points within the sampling period and taking the arithmetic mean. This reflects the average force applied during the filing process; first, calculate the standard deviation of the pressure. Dividing this value by the average pressure yields the dimensionless fluctuation coefficient, which reflects the uniformity of the applied force. The smaller the value, the more stable the applied force.

[0014] (5) (6) in, Let t be the pressure value at time t; This refers to the pressure fluctuation coefficient. This represents the standard deviation of pressure.

[0015] Step 4.4: Extract the mean, standard deviation, and attitude deviation for each angle, including pitch angle P, roll angle R, and yaw angle Y. Calculate the arithmetic mean of the sampled data from each of the three angle channels to reflect the average attitude during the filing process. Calculate the standard deviation for each of the three angle channels to reflect the stability of the attitude in each direction; a smaller standard deviation indicates a more stable file attitude during filing. Compare the mean of the three angles with the preset ideal attitude (…). Euclidean distance calculations are performed to comprehensively assess the overall deviation between the actual and ideal filing postures. A larger Euclidean distance value indicates a more severe deviation in the filing posture. Specifically: (7) (8) (9) in, This represents the average pitch angle. Let be the pitch angle value at time t; This represents the average roll angle. Let be the roll angle value at time t; This represents the average yaw angle. Let be the yaw angle value at time t; (10) (11) (12) in, The standard deviation of the pitch angle; This represents the standard deviation of the roll angle. The standard deviation of the yaw angle; (13) in, Ideal pitch angle posture; Ideal roll angle posture; The ideal attitude for the yaw angle; The square of the pitch angle deviation; The square of the roll angle deviation; It is the square of the yaw angle deviation.

[0016] The fifth step involves normalizing the temperature, humidity, pressure, and pose feature data extracted in the fourth step, mapping all physical quantities to the [0, 1] interval to eliminate dimensional differences. The formula is as follows: (14) in, Normalized data; Data to be normalized; Minimum reference value for the data; The maximum reference value for the data.

[0017] Step 6: Based on the normalized temperature, humidity, pressure, and pose feature data obtained in Step 5, score the temperature, humidity, pressure, and pose states, including the following steps: Step 6.1: Obtain the temperature status score using the following formula. : (15) in, It is the ideal working temperature; It is the normalized rate of temperature change; It is the first weighting coefficient. ; This is the second weighting coefficient. ; ; Step 6.2, use the following formula to obtain the humidity status score. : (16) in, This is the ideal humidity for operation; This is the normalized humidity variation. It is the third weighting coefficient. ; It is the fourth weighting coefficient. ; ; Step 6.3: Obtain the pressure state score using the following formula. : (17) in, It is the optimal force value; It is the fifth weighting coefficient. ; It is the sixth weighting coefficient. ; ; Step 6.4: Obtain the pose state score using the following formula. : (18) in, It is the normalized attitude deviation; It is the normalized standard deviation of the pitch angle; It is the normalized standard deviation of the roll angle; It is the normalized standard deviation of the yaw angle; It is the seventh weighting coefficient. ; It is the eighth weighting coefficient. ; ; In the scoring process of temperature, humidity, pressure and pose, the values ​​of the participating items for each state score are all in the range [0,1]. If an extreme outlier occurs, a value less than 0 is taken as the cutoff value 0, and a value greater than 1 is taken as the cutoff value 1.

[0018] Step 7: Based on the scores of each state obtained in Step 6, perform a comprehensive score, and calculate the comprehensive state score S using a weighted summation method: (19) in, It is the first weight. It is the second weight. It is the third weight. It is the fourth weight, and each weight satisfies ; .

[0019] The overall condition score S can be used to judge the quality of the filing process. The higher the score, the better the filing quality, and the lower the score, the worse the filing quality.

[0020] Furthermore, the filing monitoring and evaluation method is embedded in the host computer. The host computer acquires data frames through a wireless transmission unit, and the data frames are used for various purposes in the host computer.

[0021] A smart filing monitoring handle with multi-sensor fusion includes a handle top cover 1, a circuit board 2, a handle bottom cover 3, and a side plate 4, which is connected to a host computer located independently on a PC.

[0022] The handle cover 1 has a display unit mounting slot 101 on its surface and an internal cover housing for reserving mounting space for components on the top surface of the circuit board. The four corners have first bolt holes for connecting to the circuit board 2 and the handle cover 3. The outer side of the handle cover 1 has a rounded corner design.

[0023] The circuit board 2 serves as the carrier for each unit module and is installed in the internal space after the upper cover 1 and the lower cover 3 of the handle are joined together. The top surface of the circuit board 2 integrates a display unit, a temperature sensing unit, a humidity sensing unit, a posture sensing unit, a wireless transmission unit, and a microcontroller unit. The bottom surface of the circuit board 2 integrates a pressure sensing unit, a vibration unit, and a power supply unit. The handle lower cover 3 has a universal file mounting head 301 at its head. The pressure sensing module is pre-installed on the upper surface of the file hole. After the universal file is inserted, the pressure sensing module fits against the tail of the file. The pressure sensing module refers to the pressure sensing part, and the pressure sensing unit refers to the conversion part that converts the pressure signal into an electrical signal. The pressure sensing module is connected to the pressure sensing unit on the circuit board 2 through wires. The handle lower cover 3 has a lower cover receiving compartment 302 inside, which is used to reserve installation space for the components on the bottom of the circuit board. There are second bolt holes at the four corners, which are used to connect to the circuit board 2 and the handle upper cover 3. The side has a side plate mounting hole for use when the power supply unit is charging. The outer side of the handle lower cover 3 is designed with rounded corners.

[0024] The host computer has functions for data display, data storage, data processing, command control, and quality evaluation, as detailed below: The data display function is based on the data display module, which is used to display the temperature, humidity, pressure and pose (including pitch, roll and yaw angles) in real time. The data storage function is implemented based on the data storage module, which is used to save the original data of each physical quantity after the data frame is split into files, forming the user's filing training database. The data processing function is implemented based on the data processing module. It preprocesses the received raw data by filtering out noise and eliminating outliers. Only after processing can the data be evaluated by the evaluation method. The quality evaluation function is implemented based on the quality evaluation module. It uses the filing quality evaluation method embedded in the host computer to extract features of each physical quantity data, score each item separately, and then weight and fuse them to give a comprehensive score of filing quality during use. The command control function is implemented based on the command control module. Based on the scores of each item in the quality evaluation function, if the score is too low, a preset threshold will be triggered, and the relevant command will be sent to the microcontroller unit in the controller through the wireless transmission unit. The microcontroller unit controls the relevant units to make corresponding prompts, so as to achieve accurate dynamic feedback.

[0025] Furthermore, the temperature sensing unit, humidity sensing unit, pressure sensing unit, and posture sensing unit will sense each physical quantity in real time and send the collected data to the microcontroller unit. The microcontroller unit can then autonomously display the physical quantities collected by each sensing unit on the display unit. Furthermore, the microcontroller unit splices the data from each sensor unit into data frames and sends them to the host computer via the wireless transmission unit. The host computer scores each item using an embedded quality evaluation method. If the score is too low, a threshold is triggered, and the host computer sends the corresponding instruction control back to the microcontroller unit via the wireless transmission unit. After parsing the corresponding instruction, the microcontroller unit controls the vibration unit to generate vibration to remind the operator. At the same time, the display unit displays the items with low scores, so the operator can clearly understand the problems in this stage of the operation and achieve dynamic feedback.

[0026] The beneficial effects of this invention are: (1) This invention achieves multi-dimensional and all-element monitoring of the filing process through multi-sensor fusion and integrated handle structure design. It can simultaneously collect temperature, humidity, pressure and position signals, effectively solving the problem of single physical quantity monitoring in the existing technology. (2) By performing filtering preprocessing, feature extraction, normalization and weighted scoring on the data through the host computer, the objective quantitative evaluation of filing quality is realized, avoiding the subjective bias caused by human experience judgment. At the same time, through the data display and data storage functions of the host computer, the filing process of the operator can be visualized, traceable and quantitatively analyzed. (3) Based on the wireless transmission and instruction feedback mechanism, a closed-loop dynamic guidance system of monitoring-analysis-feedback was constructed. It can drive the vibration prompts and display output on the handle in real time according to the scoring results, and accurately guide the operator to correct the action. At the same time, all sensing, control and interaction units are highly integrated inside the handle. The device has a high degree of integration and a compact structure, making it more suitable for actual teaching and processing scenarios. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of the handle of the present invention; Figure 2 This is a system block diagram of the present invention; Figure 3 This describes the implementation steps of a filing monitoring and evaluation method based on multi-sensor fusion. Figure 4 It is a closed-loop dynamic feedback block diagram; In the picture: 1. Top cover of the handle; 2. Circuit board; 3. Bottom cover of the handle; 4. Side plate; 101 Display unit mounting slot; 301 File mounting head; 302 Lower cover receiving compartment. Detailed Implementation

[0028] To ensure that the technical solution and advantages of this invention are more clearly understood, the invention and its technical means will be further described below in conjunction with specific embodiments and accompanying drawings, detailing its specific structure, implementation methods, and technical features. In the description, the described embodiments are a subset of the embodiments of this invention. Other embodiments obtained by those skilled in the art without creative effort based on these embodiments are all within the scope of protection of this application.

[0029] Figure 1 This is a schematic diagram of the handle of the present invention. Figure 2 This is a system block diagram of the present invention. Figure 3 This describes the implementation steps of a filing monitoring and evaluation method based on multi-sensor fusion. Figure 4 This is a closed-loop dynamic feedback block diagram. The embodiments of the present invention are described in detail below with reference to the accompanying drawings: A filing monitoring and evaluation method based on multi-sensor fusion is proposed. The evaluation method is embedded in a host computer, which is located on a PC. Figure 3 As shown, it includes the following steps: The experimental subject was a beginner-level student. The workpiece material was Q235 carbon structural steel, with dimensions of 100mm × 100mm × 8mm. The filing time was 5 minutes. Sampling frequencies were: pressure channel 200Hz, posture channel 100Hz, temperature channel 10Hz, and humidity channel 5Hz. The host computer's evaluation cycle was set to 10 seconds (i.e., a comprehensive score was output every 10 seconds), and all sampling points from each channel within that time period were used for calculation. The host computer presets the normalized reference range and ideal value for each physical quantity, as shown in Table 1.

[0030] Table 1: Normalized Reference Range and Ideal Value of Each Physical Quantity The weighting coefficients for each score are determined, as shown in Table 2: Table 2: Values ​​of Each Scoring Weight Coefficient The first step involves each sensor acquiring data in real time. Within a complete 10-second evaluation cycle, the microcontroller receives raw data from multiple sensors within the handle, inserts delimiters between different data types, concatenates them into a single data frame, and transmits it to the host computer via the wireless transmission unit, as follows: The data frames mainly include: raw temperature data, raw humidity data, raw pressure data, and raw pose data. All data frames are raw data, and all data will be preprocessed and have features extracted in the host computer.

[0031] The second step, based on the first step, involves the host computer receiving the integrated data frame from the microcontroller unit. The host computer then performs splitting processing on the data frame, as follows: Step 2.1: Extract the raw temperature data from the data frame and store it using a temporary temperature variable; Step 2.2: Extract the raw humidity data from the data frame and store it using a temporary humidity variable; Step 2.3: Extract the raw pressure data from the data frame and store it using a temporary pressure variable; Step 2.4: Extract the original pose data from the data frame and store it using a pose temporary variable.

[0032] The third step is to preprocess the raw data separated in the second step. The purpose of preprocessing is to eliminate interference from high-frequency impact vibrations and noise, thereby improving the accuracy of the condition assessment. Different preprocessing methods are used for different physical quantities according to their characteristics, as detailed below: Step 3.1: Amplitude-limited moving average filtering is used to eliminate spike noise. After filtering, the temperature data is stable between 42.2℃ and 43.1℃ within a 10-second evaluation period. Step 3.2 involves using a limited-amplitude moving average filter to eliminate spike noise. After filtering, the humidity data stabilizes at 51.2% within a 10-second evaluation period. RH -53.8% RH between; Step 3.3: Kalman filtering is used to process the process noise covariance. The measurement noise covariance R is adaptively adjusted based on the real-time vibration amplitude. After filtering, the pressure signal fluctuates within a range of 8N from the original peak value. 65N smoothing is 12N 58N, high-frequency noise components are effectively suppressed; Step 3.4: Apply Mahony complementary filtering, filter gain... , By combining the low-frequency stability of the accelerometer with the high-frequency response of the gyroscope, the pitch angle jitter amplitude is reduced to within ±1.2° after filtering.

[0033] The fourth step involves feature extraction from all the raw sensor data from the third step to obtain temperature feature data, humidity feature data, pressure feature data, and pose feature data. This includes the following steps: Step 4.1: Using formulas (1) and (2), extract the average temperature and temperature change rate from the temperature features, and take all valid temperature points within the 10-second evaluation period. ) Step 4.2: Using formulas (3) and (4), the average humidity and humidity change amplitude are extracted from the humidity features. All valid humidity points within the 10-second evaluation period are selected. ) Step 4.3: Using formulas (5) and (6), extract the average pressure and pressure fluctuation coefficient from the pressure characteristics, and take all effective pressure points within the 10-second evaluation period. ) Step 4.4: Using formulas (7) to (13), extract the mean, standard deviation, and attitude deviation of each angle from the pose features. The angles include pitch angle P, roll angle R, and yaw angle Y. Take all valid pose points within a 10-second evaluation period. ) Fifth, using formula (14), the temperature feature data, humidity feature data, pressure feature data, and pose feature data extracted in the fourth step are all normalized, and the physical quantities are uniformly mapped to the interval [0, 1] to eliminate dimensional differences, as shown in Table 3: Table 3: Feature Data Normalization Step 6: Using formulas (15) to (18), based on the normalized temperature characteristic data, humidity characteristic data, pressure characteristic data, and pose characteristic data obtained in step 5, score the temperature state, humidity state, pressure state, and pose state, including the following steps: Step 6.1: Obtain the temperature status score using the following formula. : The temperature condition score is 92.7.

[0034] Step 6.2, use the following formula to obtain the humidity status score. : The humidity rating is 93.3.

[0035] Step 6.3: Obtain the pressure state score using the following formula. : The stress level score was 61.6.

[0036] Step 6.4: Obtain the pose state score using the following formula. : The pose score was 69.4.

[0037] Step 7: Using formula (19), based on the scores of each state obtained in step 6, perform a comprehensive score, and calculate the comprehensive state score S using a weighted summation method: The overall score is 69.4.

[0038] In this embodiment, the scoring results show that scores above 90 for temperature and humidity are considered excellent, while scores between 60 and 70 for pressure and posture are considered average. The operator's main problems were specifically diagnosed as follows: Problem 1: Insufficient pressure applied. The actual applied pressure was 38.2 kJ, which deviated from the ideal value of 80 kJ by 52.25%. This resulted in a deduction of 18.3 points from the pressure score, and was the main reason why the overall score fell below 70 points. Question 2: The pressure fluctuates significantly, and the fluctuation coefficient is high. 20.2 points were deducted from the pressure score, indicating that the uniformity of force application still needs improvement; Question 3: Poor positional stability, with a pitch angle standard deviation of 1.8° (normalized value 0.36), contributing the most to the stability item. This indicates that the file exhibits significant back-and-forth jitter in the pitch direction, which is the primary issue related to the operator's positional stability. The roll angle standard deviation of 1.2° (normalized value 0.24) is the second most significant, while the yaw angle standard deviation of 0.9° (normalized value 0.18) is relatively small. Furthermore, the yaw angle has a low weight in the scoring (only 0.2), indicating that the use of forward and cross files has little impact on the workpiece machining quality.

[0039] A smart filing monitoring handle with multi-sensor fusion includes a handle top cover 1, a circuit board 2, a handle bottom cover 3, and a side plate 4; the host computer software is independently located on a PC. The specific models of each unit are as follows: the display unit is OLED, the vibration unit is a vibration motor, the wireless transmission unit is ESP8266 using WIFI technology, the microcontroller unit is STM32C8T6, the posture sensing unit is MPU6050, the pressure sensing module is a resistance strain gauge force sensor, and the temperature sensing unit and humidity sensing unit are integrated together as AHT20. like Figure 1 As shown, the surface of the handle cover 1 has a reserved display unit mounting slot 101, and the interior is designed with a cover housing to reserve mounting space for components on the top surface of the circuit board. The four corners are provided with first bolt holes for connection with the circuit board 2 and the handle cover 3. The outer side of the handle cover 1 adopts a rounded corner design.

[0040] The circuit board 2 serves as the carrier for each unit module and is installed in the internal space after the upper cover 1 and the lower cover 3 of the handle are joined together. The top surface of the circuit board 2 integrates a display unit, a temperature sensing unit, a humidity sensing unit, a posture sensing unit, a wireless transmission unit, and a microcontroller unit. The bottom surface of the circuit board 2 integrates a pressure sensing unit, a vibration unit, and a power supply unit. In this embodiment, the host computer has other functions, including data display, data storage, data processing, command control, and quality evaluation. Specifically: The data display function can display the temperature, humidity, pressure, and posture (including pitch, roll, and yaw angles) in real time, allowing the observer to monitor the data of each physical quantity through the host computer's interface; the data storage function can save the original data of each physical quantity after the data frame is split into data files, forming the user's filing training database, which the host computer will automatically collect and save in a designated file; the data processing and quality evaluation function is the entire process of comprehensive status scoring; the command control function, based on the scores of each item in the quality evaluation function, will trigger a preset threshold if the score is too low, and will send relevant commands to the microcontroller unit in the handle through the wireless transmission unit. The microcontroller unit will control the relevant units to make corresponding prompts, achieving accurate dynamic feedback.

[0041] In this embodiment, as Figure 2 As shown, the temperature sensing unit, humidity sensing unit, pressure sensing unit, and posture sensing unit will sense each physical quantity in real time and send the collected data to the microcontroller unit. The microcontroller unit can autonomously display the physical quantities collected by each sensing unit on the display unit. In this embodiment, as Figure 4 As shown, the microcontroller unit splices the data from each sensor unit into a data frame and sends it to the host computer via the wireless transmission unit. The host computer scores each item using an embedded quality evaluation method. If the score is too low, a threshold is triggered, and the host computer sends the corresponding instruction control back to the microcontroller unit via the wireless transmission unit. After parsing the corresponding instruction, the microcontroller unit controls the vibration unit to generate vibration to remind the operator. At the same time, the display unit displays the items with low scores, so the operator can clearly understand the problems in this stage of the operation and achieve dynamic feedback.

[0042] like Figure 1As shown, the head of the lower cover 3 of the handle is designed with a universal file mounting head 301. The pressure sensing module is pre-installed on the upper surface of the file hole. After the universal file is inserted, the pressure sensing module fits against the tail of the file. The pressure sensing module is connected to the pressure sensing unit on the circuit board 2 through a wire. The lower cover 3 of the handle is designed with a lower cover receiving compartment 302 inside, which is used to reserve installation space for the components on the bottom of the circuit board. There are second bolt holes at the four corners, which are used to connect with the circuit board 2 and the upper cover 3 of the handle. The side is designed with a side plate mounting hole for use when the power supply unit is charging. The outer side of the lower cover of the handle is designed with rounded corners.

[0043] In summary, a temperature sensor, humidity sensor, pressure sensor, posture sensor, display unit, vibration unit, wireless transmission unit, power supply unit, and microcontroller unit are integrated into the file handle. The microcontroller unit transmits the data collected by each sensor unit to a host computer. The host computer embeds a multi-dimensional signal processing and fusion algorithm to make a comprehensive quality evaluation of the filing process. When the score is too low, a threshold is triggered, and a command is fed back to the smart handle. The vibration unit on the handle vibrates to provide tactile warning feedback, and the display unit displays the adjustment suggestions given by the host computer. This solves the problem that existing filing tools have limited monitoring functions and lack dynamic feedback and quality evaluation, realizing a closed loop of detection, analysis, and feedback during the filing process. Based on the filing operation, a multi-dimensional data fusion algorithm is introduced to achieve a scientific evaluation of filing quality.

[0044] The above description is merely a preferred embodiment of the present invention. It should be understood that the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A filing monitoring and evaluation method based on multi-sensor fusion, characterized in that, The filing monitoring and evaluation method includes the following steps: The first step is to acquire raw data collected by multiple sensors inside the smart file monitoring handle, insert separators between different types of data, and splice them into a data frame. The second step is to split the data frames. The third step is to preprocess the raw data separated in the second step; The fourth step is to extract features from all the preprocessed raw data in the third step to obtain temperature feature data, humidity feature data, pressure feature data, and pose feature data. The fifth step is to normalize the temperature, humidity, pressure, and pose feature data extracted in the fourth step. Step 6: Based on the normalized temperature, humidity, pressure, and pose feature data obtained in Step 5, score the temperature, humidity, pressure, and pose states. Specifically: Step 6.1: Obtain the temperature status score using the following formula. : （15） in, It is the ideal working temperature; It is the normalized rate of temperature change; It is the first weighting coefficient; This is the second weighting coefficient; Step 6.2, use the following formula to obtain the humidity status score. : （16） in, This is the ideal humidity for operation; This is the normalized humidity variation. It is the third weighting coefficient; It is the fourth weighting coefficient; Step 6.3: Obtain the pressure state score using the following formula. : （17） in, It is the optimal force value; It is the fifth weighting coefficient. ; It is the sixth weighting coefficient. ; Step 6.4: Obtain the pose state score using the following formula. : （18） in, It is the normalized attitude deviation; It is the normalized standard deviation of the pitch angle; It is the normalized standard deviation of the roll angle; It is the normalized standard deviation of the yaw angle; It is the seventh weighting coefficient; It is the eighth weighting coefficient; During the scoring process of temperature, humidity, pressure and pose, the values ​​of the participating items for each state score are all in the range [0,1]. If an extreme outlier occurs, if the calculated value is less than 0, the cutoff value is 0; if the calculated value is greater than 1, the cutoff value is 1. The seventh step is to perform a comprehensive score based on the scores of each state obtained in the sixth step. The comprehensive state score S is calculated by weighted summation. The comprehensive state score S is used to judge the quality of the machining process during filing. The higher the score, the better the quality of filing.

2. The filing monitoring and evaluation method based on multi-sensor fusion according to claim 1, characterized in that, In the first step, the data frame includes raw temperature data, raw humidity data, raw pressure data, and raw pose data; all data frames are raw data.

3. The filing monitoring and evaluation method based on multi-sensor fusion according to claim 2, characterized in that, The second step is as follows: Step 2.1: Extract the raw temperature data from the data frame and store it using a temporary temperature variable; Step 2.2: Extract the raw humidity data from the data frame and store it using a temporary humidity variable; Step 2.3: Extract the raw pressure data from the data frame and store it using a temporary pressure variable; Step 2.4: Extract the original pose data from the data frame and store it using a pose temporary variable.

4. The filing monitoring and evaluation method based on multi-sensor fusion according to claim 3, characterized in that, In the third step, different preprocessing methods are used for different physical quantities according to their characteristics, as follows: Step 3.1: The raw temperature data is processed using a limited moving average filter; Step 3.2: The raw humidity data is processed using a limited moving average filter. Step 3.3: The raw pressure data is processed using Kalman filtering; Step 3.4: The original pose data is processed using Mahony complementary filtering.

5. The filing monitoring and evaluation method based on multi-sensor fusion according to claim 4, characterized in that, The fourth step is as follows: Step 4.1: Extract the average temperature and temperature change rate by summing all temperature data points within the sampling period and taking the arithmetic mean. This reflects the overall heat load level during the filing process; the temperature derivative with respect to time is approximated using the backward difference method to obtain the temperature change rate; （1） （2） in, This represents the total number of sampling time points; This refers to the sampling time point; Let t be the temperature value at time t. The sampling time interval; The temperature at the previous moment; Step 4.2: Extract average humidity and humidity variation range. The average humidity is obtained by summing all humidity data points within the sampling period and taking the arithmetic mean. This reflects the average humidity level of the filing environment; the difference between the maximum and minimum humidity values ​​during the sampling period is used to obtain the humidity variation range. （3） （4） in, Let be the humidity value at time t; This represents the range of humidity change. Maximum humidity; Minimum humidity; Step 4.3: Extract the average pressure and pressure fluctuation coefficient by summing all pressure data points within the sampling period and taking the arithmetic mean. This reflects the average force applied during the filing process; first, calculate the standard deviation of the pressure. Then divide by the average pressure to obtain the dimensionless fluctuation coefficient; （5） （6） in, Let t be the pressure value at time t; This is the pressure fluctuation coefficient; The standard deviation of pressure; Step 4.4: Extract the mean, standard deviation, and attitude deviation for each angle, including pitch angle P, roll angle R, and yaw angle Y. Calculate the arithmetic mean of the sampled data from each of the three angle channels to reflect the average attitude during the filing process. Calculate the standard deviation for each of the three angle channels to reflect the stability of the attitude in each direction; a smaller standard deviation indicates a more stable file attitude during filing. Compare the mean of the three angles with the preset ideal attitude (…). Euclidean distance calculations are performed to comprehensively assess the overall deviation between the actual and ideal attitudes. A larger Euclidean distance value indicates a more severe deviation in the filing attitude. Specifically: （7） （8） （9） in, This represents the average pitch angle. Let be the pitch angle value at time t; This represents the average roll angle. Let be the roll angle value at time t; This represents the average yaw angle. Let be the yaw angle value at time t; （10） （11） （12） in, The standard deviation of the pitch angle; This represents the standard deviation of the roll angle. The standard deviation of the yaw angle; （13） in, Ideal attitude for pitch angle; Ideal roll angle posture; The ideal attitude for the yaw angle; The square of the pitch angle deviation; The square of the roll angle deviation; It is the square of the yaw angle deviation.

6. The filing monitoring and evaluation method based on multi-sensor fusion according to claim 5, characterized in that, In the fifth step, all physical quantities are uniformly mapped to the interval [0, 1], as shown in the following formula: (14) in, Normalized data; Data to be normalized; Minimum reference value for the data; The maximum reference value for the data.

7. The filing monitoring and evaluation method based on multi-sensor fusion according to claim 6, characterized in that, In the sixth step: In step 6.1, ; ; ; In step 6.2, ; ; ; In step 6.3, ; ; ; In step 6.4, ; ; .

8. The filing monitoring and evaluation method based on multi-sensor fusion according to claim 7, characterized in that, In the seventh step, the comprehensive status score S is: （19） in, It is the first weight. It is the second weight. It is the third weight. It is the fourth weight, and each weight satisfies ; .

9. A smart filing monitoring handle with multi-sensor fusion, characterized in that, The filing monitoring and evaluation method based on multi-sensor fusion as described in any one of claims 1-8 is implemented using an intelligent filing monitoring handle. The intelligent filing monitoring handle includes a handle upper cover (1), a circuit board (2), and a handle lower cover (3), which are connected to a host computer, which is independently located on a PC. The upper cover (1) of the handle has a pre-reserved display unit mounting slot (101) and an internal upper cover receiving compartment. The lower cover (3) of the handle has an internal lower cover receiving compartment (302). The circuit board (2) is installed after the two are put together. The lower cover (3) of the handle is designed with a file mounting head (301). The pressure sensing module is pre-installed on the upper surface of the file hole. After the universal file is inserted, the pressure sensing module is attached to the tail of the file. The pressure sensing module is connected to the pressure sensing unit on the circuit board (2) through a wire. The circuit board (2) serves as the carrier for each unit module and is installed in the internal space after the upper cover (1) and lower cover (3) of the handle are joined. The circuit board (2) integrates a display unit, a temperature sensing unit, a humidity sensing unit, a posture sensing unit, a wireless transmission unit, and a micro control unit. The bottom surface of the circuit board (2) integrates a pressure sensing unit, a vibration unit, and a power supply unit.

10. The intelligent filing monitoring handle with multi-sensor fusion according to claim 9, characterized in that, The host computer has functions for data display, data storage, data processing, command control, and quality evaluation, as detailed below: The data display function is implemented based on the data display module, which is used to display the temperature, humidity, pressure and position status in real time. The data storage function is implemented based on the data storage module, which is used to save the original data of each physical quantity after the data frame is split into files, forming the user's filing training database. The data processing function is implemented based on the data processing module, which preprocesses the raw data so that the processed data can be evaluated using the evaluation method. The quality evaluation function is based on the quality evaluation module, which extracts features from each physical quantity data, scores each item separately, and then weights and fuses them to give a comprehensive score of the filing quality during use. The aforementioned command control function is implemented based on the command control module, and achieves accurate dynamic feedback based on the scores of each sub-item in the quality evaluation function; Temperature sensing unit, humidity sensing unit, pressure sensing unit, and posture sensing unit will sense each physical quantity in real time and send the collected data to microcontroller unit. Microcontroller unit will stitch the data from each sensing unit into a data frame and send it to the host computer through wireless transmission unit to achieve scoring of each item. If the value is too low, a preset threshold will be triggered to achieve dynamic feedback. The threshold is determined according to the actual situation.

Images