Low-power-consumption standby intelligent wake-up control system and method for percussion drill

Abstract

The invention discloses a low-power-consumption standby intelligent awakening control system and method for a percussion drill, and particularly relates to the technical field of percussion drills. A user operation signal and an external interference signal are accurately distinguished by using a built-in classification model in combination with the frequency characteristic, the duration and the mode of a vibration signal, and an accurate wake-up signal is triggered; according to a real-time evaluation result of the classification model, a threshold value of a wake-up condition is dynamically adjusted, system adaptability is improved, and a main processor and a high-power-consumption module are rapidly activated after confirmation of a user operation signal; if the false wake-up behavior is recognized, the device can automatically recover the low-power-consumption mode, and the false wake-up characteristics are recorded, the problems that in the prior art, external interference signals are sensitive, false wake-up is frequent, and consequently battery loss and insufficient endurance are caused are effectively solved, and the operation efficiency and the endurance capacity of the device in the complex environment are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of impact drills, and in particular to a low-power standby intelligent wake-up control system and method for an impact drill. Background Art

[0002] The impact drill's low-power standby intelligent wake-up control is a technical design that allows the impact drill to enter a low-power mode when not in operation, and can quickly wake up the device in an intelligent way to resume normal operation. This design can not only significantly save power and extend the device's standby time, but also improve ease of use and response speed.

[0003] For example, a hammer drill has built-in motion sensors and touch detection. When the user does not operate the device for a long time, the hammer drill will automatically enter low-power mode, keeping only a small number of core sensors in operation to reduce battery loss. If the user picks up the device again or taps the start button, the motion sensor or touch sensor will immediately sense and wake up the device, allowing it to quickly return to full power. For example, when a construction worker pauses and puts down the hammer drill and the device enters standby, when he picks it up again a few minutes later, the device will respond immediately without the need for manual restart, thereby improving work efficiency and saving battery energy.

[0004] The prior art has the following deficiencies:

[0005] In the prior art, motion sensors or touch sensors may be overly sensitive to external interference that is not caused by user operations, such as device vibration, accidental touch, or even slight vibrations in the environment (such as shaking of the ground or the movement of a toolbox). This false wake-up will cause the device to frequently switch from low-power mode to working mode, increase battery consumption, and significantly shorten battery life. This may seriously affect work efficiency for users who work in remote construction scenarios and need to use the device for a long time. Summary of the invention

[0006] The object of the present invention is to provide a low-power standby intelligent wake-up control system and method for an impact drill to solve the deficiencies in the background technology.

[0007] In order to achieve the above object, the present invention provides the following technical solution: a low-power standby intelligent wake-up control method for an impact drill, comprising the following steps:

[0008] S1: Through the built-in motion sensor and touch sensor, the user interaction behavior signal of the device in the working state is collected in real time. If it is detected that the device has no operation behavior within the set time, it will enter the low-power standby mode. In the low-power standby mode, only the core sensor is kept in the low-power working state, and the main motor drive circuit and high-power modules are turned off;

[0009] S2: Detect user operation behavior signals through core sensors, and analyze the collected signals using the built-in classification model based on the frequency characteristics, duration and pattern of vibration signals to distinguish user operation signals from external interference signals. When the signal characteristics meet the preset threshold of user operation behavior, the device wake-up signal is triggered;

[0010] S3: According to the real-time evaluation results of the classification model, the threshold of the wake-up condition is dynamically adjusted, including the minimum duration threshold of the vibration signal and the frequency range of the vibration signal. After the user's actual operation signal is confirmed, the main processor and high-power modules are activated to enter the full-power working mode;

[0011] S4: During the wake-up process, if a false wake-up behavior is identified, the device will automatically re-enter the low-power standby mode and record the false wake-up characteristics to optimize the wake-up logic. By continuously collecting user operation behavior data, the classification threshold is dynamically updated.

[0012] Preferably, in S2, the vibration signal frequency fluctuation index is generated after analyzing the frequency change of the vibration signal, and the method for obtaining the vibration signal frequency fluctuation index is:

[0013] Set the sampled time series vibration signal x[n], the sampling frequency is fs, and select the time window length ,use The sliding window of length splits the signal, and each sliding sampling points, for each signal in the window Perform fast Fourier transform to obtain the corresponding frequency spectrum: ; is the frequency distribution result of the mth window and the spectrum value corresponding to the kth frequency point. w[n] is the window function. is the kernel function of Fourier transform;

[0014] Calculate the main frequency in each window, that is, the frequency point with the largest frequency power: Convert the FFT result into a power spectrum : , find the frequency point with maximum power , ;in, Indicates the main frequency of the mth time window, the change of the main frequency on the time axis for: ; Define the frequency fluctuation index of the vibration signal as the root mean square value of the main frequency change, which represents the overall intensity of the frequency fluctuation. The calculation expression is: ; M is the total number of time windows, is the change of main frequency in the mth time window, and FC is the frequency fluctuation index of vibration signal.

[0015] Preferably, in S2, after analyzing the duration of the vibration signal in the time domain, a vibration signal duration anomaly index is generated, and the method for obtaining the vibration signal duration anomaly index is:

[0016] The sampled vibration signal x[s] sets an amplitude threshold for the vibration signal and a duration threshold for , which is used to define the upper and lower bounds of the normal duration, and divide the vibration signal into multiple active and inactive segments. The duration of each segment is ΔT[i], which represents the duration of the i-th segment signal. , detect the active area of ​​vibration signal: ; is the set of sample points of the i-th segment activity signal;

[0017] For each active area, calculate its duration : ; is the number of samples of the i-th segment activity signal, is the sampling frequency, used to convert the number of samples into time;

[0018] According to the normal operation status or historical data of the equipment, determine the normal duration range: Tnormal = [Tmin, Tmax]; Tmin is the lower limit of the duration, and Tmax is the upper limit of the duration;

[0019] Check whether the duration of each activity ΔT[i] exceeds the normal range [Tmin, Tmax]: ; Calculate the abnormal index of vibration signal duration, the expression is: ; DF is the abnormal index of vibration signal duration, is the total number of active segments, is the average of normal duration.

[0020] Preferably, the vibration signal frequency fluctuation index and the vibration signal duration anomaly index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as inputs of the machine learning model. The machine learning model predicts each sample as a user operation signal or an external interference signal label based on each group of comprehensive feature vectors as a prediction target, and takes minimizing the sum of prediction errors for all samples as user operation signals or external interference signal labels as a training target. The machine learning model is trained until the sum of prediction errors converges, and the model training is stopped. Each sample is determined to be a user operation signal or an external interference signal based on the model output results, wherein the machine learning model is a polynomial regression model.

[0021] Preferably, each sample is determined to be a user operation signal or an external interference signal based on the predicted value yk output by the model, specifically including: yk=1: the signal is determined to be a user operation signal; yk=0: the signal is determined to be an external interference signal; when the signal is determined to be a user operation signal, the system triggers the wake-up mechanism to activate the main processor and the high-power consumption module; if the signal is determined to be an external interference signal, the device maintains a low-power standby mode.

[0022] Preferably, in S3, according to the real-time evaluation result of the classification model, the threshold of the wake-up condition is dynamically adjusted, including the minimum duration threshold of the vibration signal and the frequency range of the vibration signal. is the minimum duration threshold, based on the false wake-up rate of the classification model and leakage wake-up rate Dynamic Adjustment : ; is the adjusted minimum duration threshold, is the initial minimum duration threshold, k is the adjustment coefficient, which is used to control the amplitude of the threshold adjustment. is the current false wake-up rate, is the current leakage wake-up rate; when : Increase , improve the wake-up conditions, when : Decrease , relax the wake-up conditions;

[0023] Assume that the initial main frequency range of the vibration signal is [ , ], dynamically adjust the frequency range through feature importance evaluation of the classification model: ; , are the lower and upper limits of the adjusted frequency range, , It indicates the offset of the main frequency of the user operation signal observed in the model, which respectively represents the frequency fluctuation on the low-frequency side and the high-frequency side. α and β are adjustment coefficients used to control the expansion amplitude of the frequency range.

[0024] Preferably, in S4, the characteristic vector of the current wake-up signal is ,include: ;in, is the RMS value of the amplitude, is the main frequency, is the duration of the signal;

[0025] Through the pre-trained classification model M, input Then we get a predicted probability , indicating the confidence that the signal belongs to the user operation signal, the expression is: ;

[0026] Assume that the judgment threshold of the classification model is θp. If the signal is judged as external interference, the current wake-up is considered to be a false wake-up behavior. For each false wake-up signal, its key feature Record to log ;

[0027] The false wakeup feature set Normal user operation feature set Merge to form a new training set: ; Based on the expanded training set , update the classification model M using existing machine learning methods: ; According to the real-time statistics of false wake-up rate and leakage wake-up rate , dynamically adjust the classification threshold : ; In the formula, is the adjusted classification threshold, and k is the adjustment coefficient, which is used to control the change range of the threshold.

[0028] The present invention also provides a low-power standby intelligent wake-up control system for an impact drill, comprising a low-power management module, a signal detection and classification module, a full-power control module and an optimization module;

[0029] Low-power management module: Through the built-in motion sensor and touch sensor, the user interaction behavior signal of the device in the working state is collected in real time. If it is detected that the device has no operation behavior within the set time, it will enter the low-power standby mode. In the low-power standby mode, only the core sensor is kept in the low-power working state, and the main motor drive circuit and high-power modules are turned off;

[0030] Signal detection and classification module: detects user operation behavior signals through core sensors, and analyzes the collected signals using the built-in classification model based on the frequency characteristics, duration and pattern of vibration signals to distinguish user operation signals from external interference signals. When the signal characteristics meet the preset threshold of user operation behavior, the device wake-up signal is triggered;

[0031] Full power control module: According to the real-time evaluation results of the classification model, the threshold of the wake-up condition is dynamically adjusted, including the minimum duration threshold of the vibration signal and the frequency range of the vibration signal. After the user's actual operation signal is confirmed, the main processor and high-power consumption module are activated to enter the full-power working mode;

[0032] Optimization module: During the wake-up process, if a false wake-up behavior is identified, the device will automatically re-enter the low-power standby mode and record the false wake-up characteristics to optimize the wake-up logic. By continuously collecting user operation behavior data, the classification threshold is dynamically updated.

[0033] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0034] 1. The present invention effectively solves the problem of false awakening caused by excessive sensitivity of motion sensors or touch sensors to external interference in the prior art through an innovative low-power standby intelligent wake-up control method. By real-time acquisition of user operation behavior signals, combined with the frequency characteristics, duration and pattern of vibration signals, the classification model is used to accurately distinguish user operation signals from external interference signals, and dynamically adjust the threshold of the wake-up conditions (such as the minimum duration and frequency range) to achieve intelligent optimization of the wake-up behavior. After the user's actual operation signal is confirmed, the system quickly activates the main processor and high-power modules to ensure the efficient response and operational reliability of the equipment in complex construction environments; in the case of false awakening, the device can automatically restore the low-power standby mode, record the false awakening characteristics, and further reduce the occurrence of false awakening by continuously optimizing the classification model.

[0035] 2. The present invention significantly reduces battery consumption caused by false awakening, prolongs the battery life of the device, and through the dynamic update of the machine learning model, continuously adapts to user behavior and environmental changes, improves the accuracy of awakening and the intelligence level of the device. Especially in remote construction scenarios that require long-term use, the application of the present invention can significantly improve the operating efficiency of the equipment, reduce energy waste, provide users with a more durable and reliable use experience, and greatly improve the energy utilization and working performance of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0037] Figure 1 The figure is a flow chart of the method of the present invention.

[0038] Figure 2 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0040] Example 1, please refer to Figure 1 As shown, the low-power standby intelligent wake-up control method of the impact drill described in this embodiment includes the following steps:

[0041] S1: Through the built-in motion sensor and touch sensor, the user interaction behavior signal of the device in the working state is collected in real time. If it is detected that the device has no operation behavior within the set time, it will enter the low-power standby mode. In the low-power standby mode, only the core sensor is kept in the low-power working state, and the main motor drive circuit and high-power modules are turned off;

[0042] S2: Detect user operation behavior signals through core sensors, and analyze the collected signals using the built-in classification model based on the frequency characteristics, duration and pattern of vibration signals to distinguish user operation signals from external interference signals. When the signal characteristics meet the preset threshold of user operation behavior, the device wake-up signal is triggered;

[0043] S3: According to the real-time evaluation results of the classification model, the threshold of the wake-up condition is dynamically adjusted, including the minimum duration threshold of the vibration signal and the frequency range of the vibration signal. After the user's actual operation signal is confirmed, the main processor and high-power modules are activated to enter the full-power working mode;

[0044] S4: During the wake-up process, if a false wake-up behavior is identified, the device will automatically re-enter the low-power standby mode and record the false wake-up characteristics to optimize the wake-up logic. By continuously collecting user operation behavior data, the classification threshold is dynamically updated.

[0045] In S1, after starting the device, the main controller initializes and enables the following modules:

[0046] a) Motion sensor (such as accelerometer, gyroscope): monitor the motion state and vibration signal of the device in real time.

[0047] b) Touch sensor: detects whether the user interacts through buttons or touch panels.

[0048] c) Main motor drive circuit: in normal operation, still consumes power when in standby mode.

[0049] Set a fixed inactivity time threshold (such as 1-5 minutes) to determine whether the device needs to enter low power mode.

[0050] Real-time collection of user interaction behavior signals:

[0051] Motion sensor signal acquisition: At fixed time intervals (such as 100 milliseconds), the motion signals of the accelerometer and gyroscope are collected to analyze whether the device has obvious displacement, vibration, or tilt.

[0052] The signal processing module filters and suppresses noise on the collected data to eliminate the influence of slight jitter or high-frequency interference signals.

[0053] Touch sensor signal acquisition: Continuously monitor the status of the touch sensor to determine whether there is a user's press or touch operation signal.

[0054] Combine the data from the motion sensor and touch sensor to analyze whether there is any user interaction. If the device does not detect the following signals within the set time (such as 2 minutes), it is judged as "no operation":

[0055] a) Continuous displacement or vibration signal;

[0056] b) Interaction signals from touch sensors (such as button presses or touch detection).

[0057] No operation timer: The system starts a timer. When the accumulated time of the timer exceeds the threshold (such as 2 minutes) and no operation signal is detected, the low power standby mode is triggered.

[0058] Enter low power standby mode:

[0059] Turn off high-power consumption modules: The system turns off the following functional modules with high power consumption in sequence:

[0060] a) Main motor drive circuit: Stop motor output to reduce unnecessary power consumption.

[0061] b) Wireless communication module (such as Bluetooth or Wi-Fi, if available): enters sleep mode and stops data broadcasting or receiving.

[0062] c) Display or indicator lights: Dim or completely turn off the display module to reduce power consumption.

[0063] Keep core sensor working: motion sensor and touch sensor enter low power mode:

[0064] a) Reduce the sampling frequency (e.g. from 100 Hz to 10 Hz) to reduce energy consumption.

[0065] b) Only the basic signal monitoring function is retained to determine the wake-up conditions.

[0066] Low power state continuous detection:

[0067] Low-frequency collection of user interaction signals: In low-power mode, motion sensors and touch sensors continue to detect user interaction behaviors at a low sampling frequency (such as once per second).

[0068] The detection signals include:

[0069] a) Obvious signs of vibration or movement (such as the device being picked up or changed position).

[0070] b) Button touch signal (such as the user touching the device start button).

[0071] User operation behavior judgment: If a user operation behavior is detected, the system immediately triggers the wake-up mechanism to make the device exit the low power mode (the specific wake-up steps are described separately).

[0072] If there is no operation signal, the device continues to maintain low power standby mode and is checked regularly.

[0073] Returns to normal operation status:

[0074] User operation signal confirmation: After detecting the user's clear operation behavior, the system reactivates the main controller.

[0075] Start the main motor drive circuit, high power consumption modules (such as display or wireless modules), and restore the device to full power operation.

[0076] S2: Detect user operation behavior signals through core sensors (such as motion sensors), and combine the frequency characteristics of vibration signals and the duration and pattern of vibration signals to comprehensively analyze the differences between user operation signals and external interference signals. The specific steps are as follows:

[0077] Use an accelerometer (to capture linear acceleration) and a gyroscope (to capture angular velocity) to collect vibration and motion signals of the device in the X, Y, and Z axes in real time. The sampling frequency is set to above 100 Hz to ensure that the complete characteristics of the vibration are captured. Data types include: acceleration change value; frequency range of vibration (frequency domain signal is extracted through FFT fast Fourier transform); vibration duration (the duration of vibration is statistically calculated through time domain signal). Collect data in a variety of usage scenarios, covering the following two signal sources: user operation signal: vibration signal generated by the user picking up, moving, and tapping the start button of the device; external interference signal: construction site vibration, accidental collision of equipment, ground shaking, and slight vibration in the toolbox.

[0078] When the user is operating, the vibration signal frequency is usually in the low frequency range (such as 5-10 Hz), because the human body movement and the inertia of the handheld device produce slower acceleration changes. The amplitude is relatively large and continuous (obvious displacement and angular velocity changes of the device). The vibration frequency distribution is concentrated and stable (reflecting the regularity of user operation). When the user picks up the device, the acceleration signal will show a gradually increasing low-frequency characteristic. If the device tilt angle changes (gyroscope signal), its angular velocity signal will also show obvious low-frequency changes. External interference signals have a wider frequency range and usually include high-frequency components (such as signals caused by vibrating tools or hard objects hitting). The vibration amplitude is small and short-lived, lacking regularity. The frequency domain distribution of interference signals is scattered (random noise characteristics). For example, a slight shake of a tool box will generate a high-frequency signal in the range of 10-30 Hz, but the duration is extremely short (hundreds of milliseconds). The signal generated by ground vibration is usually dominated by high-frequency components and lacks obvious continuous changes.

[0079] After analyzing the frequency change of the vibration signal, the vibration signal frequency fluctuation index is generated. The method for obtaining the vibration signal frequency fluctuation index is as follows:

[0080] Set the sampled time series vibration signal x[n], the sampling frequency is fs; select the time window length (usually 50-100 ms, and window functions (such as Hanning window, rectangular window).

[0081] use The sliding window of length splits the signal, and each sliding sampling points (usually = ). For each signal in the window Perform fast Fourier transform (FFT) to obtain the corresponding frequency spectrum: ; is the frequency distribution result of the mth window and the spectrum value corresponding to the kth frequency point. w[n] is the window function. is the kernel function of Fourier transform;

[0082] Calculate the main frequency in each window, that is, the frequency point with the largest frequency power: Convert the FFT result into a power spectrum : , find the frequency point with maximum power , ;in, Indicates the main frequency of the mth time window, the change of the main frequency on the time axis for: ; Define the frequency fluctuation index of the vibration signal as the root mean square value of the main frequency change, which represents the overall intensity of the frequency fluctuation. The calculation expression is: ; M is the total number of time windows, is the change of main frequency in the mth time window, and FC is the frequency fluctuation index of vibration signal.

[0083] Duration and pattern analysis of vibration signals include: The duration of the signal for user operations is relatively long (usually greater than 500 milliseconds). The vibration signal has a distinct pattern, such as gradual acceleration when picking up the device, smooth and continuous changes. The signal is usually stable after the action is completed (such as the device no longer vibrates after being held). When the user taps the start button, the touch sensor records a short signal, while the motion sensor captures the subsequent hand-held vibration (duration greater than 1 second). When the user moves the device, the signal strength shows a gradual change.

[0084] The duration of interference signals is usually short (less than 300 milliseconds). There is a lack of clear patterns, and signal changes are random or sudden, such as sudden spike signals caused by hard objects hitting. The energy of interference signals is concentrated in a short period of time and lacks subsequent changes. For example, ground vibrations may trigger a high-intensity acceleration signal in a short period of time, but lack continuous motion changes. Statistics on the duration of vibration signals in the time domain: Set a signal duration threshold (such as 300 milliseconds) to determine the source of the vibration signal; if the signal duration is short and lacks pattern changes, it is determined to be an interference signal.

[0085] After analyzing the duration of the vibration signal in the time domain, the vibration signal duration anomaly index is generated. The method for obtaining the vibration signal duration anomaly index is as follows:

[0086] The sampled vibration signal x[s] sets an amplitude threshold for the vibration signal to define the activity state of the signal. Signals below the threshold are considered background noise or inactive areas. Duration threshold : Used to define the upper and lower bounds of normal duration, such as [Tmin,Tmax].

[0087] The vibration signal is divided into multiple active segments and inactive segments. The duration of each segment is ΔT[i], which represents the duration of the i-th segment signal. , detect the active area of ​​vibration signal: ; is the set of sample points of the i-th segment activity signal.

[0088] For each active area, calculate its duration : ; is the number of samples of the i-th segment activity signal, is the sampling frequency, which is used to convert the number of samples into time.

[0089] According to the normal operating status or historical data of the equipment, determine the normal duration range: Tnormal = [Tmin, Tmax]; Tmin is the lower limit of the duration, and Tmax is the upper limit of the duration.

[0090] Check whether the duration of each activity ΔT[i] exceeds the normal range [Tmin, Tmax]: ; Calculate the abnormal index of vibration signal duration, the expression is: ; DF is the abnormal index of vibration signal duration, is the total number of active segments, is the average of normal duration.

[0091] The vibration signal frequency fluctuation index and the vibration signal duration anomaly index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as the input of the machine learning model. The machine learning model predicts each sample as a user operation signal or an external interference signal label based on each group of comprehensive feature vectors as the prediction target, and takes minimizing the sum of prediction errors for all samples as user operation signals or external interference signal labels as the training target. The machine learning model is trained until the sum of prediction errors converges and the model training is stopped. Each sample is determined to be a user operation signal or an external interference signal based on the model output results, wherein the machine learning model is a polynomial regression model.

[0092] According to the predicted value yk output by the model, each sample is determined to be a user operation signal or an external interference signal, specifically including: yk=1: the signal is judged to be a user operation signal; yk=0: the signal is judged to be an external interference signal; when the signal is judged to be a user operation signal, the system triggers the wake-up mechanism to activate the main processor and high-power consumption modules; if the signal is judged to be an external interference signal, the device maintains a low-power standby mode to avoid power waste caused by false wake-up.

[0093] User operation behaviors usually have the following key characteristics:

[0094] Frequency characteristics: The main frequency of the vibration signal is within a certain range (e.g. 5-10 Hz, which is suitable for the vibration range of handheld devices).

[0095] Duration characteristics: The signal duration is long enough (e.g. >500 ms) to distinguish it from short-term external interference.

[0096] Amplitude feature: The amplitude of the vibration signal exceeds the threshold of the background noise.

[0097] Pattern characteristics: The change pattern of the signal is consistent with the regularity of user operations (such as picking up and pressing).

[0098] According to the above characteristics, the following threshold parameters are set:

[0099] Main frequency range: [fmin,fmax];

[0100] Amplitude Threshold: ;

[0101] Pattern similarity threshold: (such as correlation with the operation template signal);

[0102] When the signal characteristics meet the preset threshold of the user's operation behavior, the device wake-up signal is triggered.

[0103] S3: According to the real-time evaluation results of the classification model, the threshold of the wake-up condition is dynamically adjusted, including the minimum duration threshold of the vibration signal and the frequency range of the vibration signal. After the user's actual operation signal is confirmed, the main processor and high-power modules are activated to enter the full-power working mode;

[0104] set up is the minimum duration threshold, based on the false wake-up rate of the classification model and leakage wake-up rate Dynamic Adjustment : ; is the adjusted minimum duration threshold, is the initial minimum duration threshold, k is the adjustment coefficient, which is used to control the amplitude of the threshold adjustment (determined by experiments), is the current false wake-up rate (the ratio of external interference signals misidentified as user operation signals). is the current missed wake-up rate (the ratio of user operation signals that are not correctly recognized).

[0105] when : Increase , improve the wake-up conditions, when : Decrease , relax the wake-up condition; assume that the initial main frequency range of the vibration signal is [ , ], dynamically adjust the frequency range through feature importance evaluation of the classification model: ; , are the lower and upper limits of the adjusted frequency range. , It indicates the offset of the main frequency of the user operation signal observed in the model, which respectively represents the frequency fluctuation on the low-frequency side and the high-frequency side. α and β are adjustment coefficients used to control the expansion amplitude of the frequency range.

[0106] The main frequency of the user operation signal tends to expand: increase the frequency range to ensure that the boundary signal is not missed. The false wake-up rate is high (the interference signal is concentrated in the boundary area of ​​the frequency range): reduce the frequency range and filter out the interference signal in the boundary frequency band.

[0107] When the user's actual operation signal is confirmed after being detected by the sensor and judged by the system logic, the device immediately switches from low-power standby mode to full-power working mode, activating the main processor and high-power modules, including the main motor drive circuit, display module, and wireless communication module (if any), to ensure that the device can quickly respond to user operation needs. This process reduces switching delays through optimized hardware control and efficient algorithm processing, ensuring that the device can provide complete performance output when the user operates, improving user experience and work efficiency.

[0108] S4: Assume that the feature vector of the current wake-up signal is ,include: ;in, is the RMS value of the amplitude, is the main frequency, is the duration of the signal.

[0109] Through the pre-trained classification model M, input Then we get a predicted probability , indicating the confidence that the signal belongs to the user operation signal, the expression is: ;

[0110] Assume that the judgment threshold of the classification model is θp (usually 0.5 or dynamically adjusted), and the condition of false awakening behavior is as follows: if the signal is judged as external interference (confidence is lower than the threshold), the current awakening is considered to be a false awakening behavior.

[0111] For each false wake-up signal, its key features Record to log ;

[0112] Log data can include the following:

[0113] Frequency fluctuation index and duration anomaly index of the vibration signal. Background noise intensity in the current environment (can be calculated by RMS value). The main frequency of the signal and duration .

[0114] The false wakeup feature set Normal user operation feature set Merge to form a new training set: ; Based on the expanded training set , update the classification model M using existing machine learning methods (such as random forest, SVM, or deep learning): ; Updated model It can better identify false awakening features and improve classification accuracy.

[0115] According to the real-time statistics of false wake-up rate and leakage wake-up rate , dynamically adjust the classification threshold : ; In the formula, is the adjusted classification threshold, and k is the adjustment coefficient, which is used to control the change range of the threshold.

[0116] In this embodiment, the user interaction behavior signal is collected in real time through the motion sensor and the touch sensor. When there is no operation within the set time, the device enters the low-power standby mode, only the core sensor is kept working and the high-power module is turned off; the user operation signal is detected by the core sensor, and the classification model is used to analyze the signal in combination with the frequency characteristics, duration and pattern of the vibration signal to distinguish the user operation signal from the external interference signal, and the device is triggered to wake up when the signal characteristics meet the preset threshold; according to the real-time evaluation results of the classification model, the wake-up conditions are dynamically adjusted, including the minimum duration threshold and frequency range. After the user operation signal is confirmed, the main processor and the high-power module are activated to make the device enter the full-power working mode; if the wake-up process is identified as a false wake-up, the device will automatically restore the low-power standby mode and record the false wake-up characteristics to optimize the wake-up logic. At the same time, the classification model and threshold are dynamically updated by continuously collecting user behavior data, thereby improving the accuracy of device wake-up and operating efficiency.

[0117] Example 2, please refer to Figure 2 As shown, the low-power standby intelligent wake-up control system of the impact drill described in this embodiment includes a low-power management module, a signal detection and classification module, a full-power control module and an optimization module;

[0118] Low-power management module: Through the built-in motion sensor and touch sensor, the user interaction behavior signal of the device in the working state is collected in real time. If it is detected that the device has no operation behavior within the set time, it will enter the low-power standby mode. In the low-power standby mode, only the core sensor is kept in the low-power working state, and the main motor drive circuit and high-power modules are turned off;

[0119] Signal detection and classification module: detects user operation behavior signals through core sensors, and analyzes the collected signals using the built-in classification model based on the frequency characteristics, duration and pattern of vibration signals to distinguish user operation signals from external interference signals. When the signal characteristics meet the preset threshold of user operation behavior, the device wake-up signal is triggered;

[0120] Full power control module: According to the real-time evaluation results of the classification model, the threshold of the wake-up condition is dynamically adjusted, including the minimum duration threshold of the vibration signal and the frequency range of the vibration signal. After the user's actual operation signal is confirmed, the main processor and high-power consumption module are activated to enter the full-power working mode;

[0121] Optimization module: During the wake-up process, if a false wake-up behavior is identified, the device will automatically re-enter the low-power standby mode and record the false wake-up characteristics to optimize the wake-up logic. By continuously collecting user operation behavior data, the classification threshold is dynamically updated.

[0122] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0123] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.

[0124] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0125] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A low-power standby intelligent wake-up control method for an impact drill, characterized in that: The following steps are involved: S1: Through the built-in motion sensor and touch sensor, the user interaction behavior signal of the device in the working state is collected in real time. If it is detected that the device has no operation behavior within the set time, it will enter the low-power standby mode. In the low-power standby mode, only the core sensor is kept in the low-power working state, and the main motor drive circuit and high-power modules are turned off; S2: Detect user operation behavior signals through core sensors, and analyze the collected signals using the built-in classification model based on the frequency characteristics, duration and pattern of vibration signals to distinguish user operation signals from external interference signals. When the signal characteristics meet the preset threshold of user operation behavior, the device wake-up signal is triggered; S3: According to the real-time evaluation results of the classification model, the threshold of the wake-up condition is dynamically adjusted, including the minimum duration threshold of the vibration signal and the frequency range of the vibration signal. After the user's actual operation signal is confirmed, the main processor and high-power modules are activated to enter the full-power working mode; S4: During the wake-up process, if a false wake-up behavior is identified, the device will automatically re-enter the low-power standby mode and record the false wake-up characteristics to optimize the wake-up logic. By continuously collecting user operation behavior data, the classification threshold is dynamically updated.

2. The low-power standby intelligent wake-up control method for an impact drill according to claim 1, characterized in that: In S2, the vibration signal frequency fluctuation index is generated after analyzing the frequency change of the vibration signal. The method for obtaining the vibration signal frequency fluctuation index is: Set the sampled time series vibration signal x[n], the sampling frequency is fs, and select the time window length ,use The sliding window of length splits the signal, and each sliding sampling points, for each signal in the window Perform fast Fourier transform to obtain the corresponding frequency spectrum: ; is the frequency distribution result of the mth window and the spectrum value corresponding to the kth frequency point. w[n] is the window function. is the kernel function of Fourier transform; Calculate the main frequency in each window, that is, the frequency point with the largest frequency power: Convert the FFT result into a power spectrum : , find the frequency point with maximum power , ;in, Indicates the main frequency of the mth time window, the change of the main frequency on the time axis for: ; Define the frequency fluctuation index of the vibration signal as the root mean square value of the main frequency change, which represents the overall intensity of the frequency fluctuation. The calculation expression is: ; M is the total number of time windows, is the change of main frequency in the mth time window, and FC is the frequency fluctuation index of vibration signal.

3. The low-power standby intelligent wake-up control method for an impact drill according to claim 2, characterized in that: In S2, the duration of the vibration signal in the time domain is analyzed to generate a vibration signal duration anomaly index. The method for obtaining the vibration signal duration anomaly index is as follows: The sampled vibration signal x[s] sets an amplitude threshold for the vibration signal and a duration threshold for , which is used to define the upper and lower bounds of the normal duration, and divide the vibration signal into multiple active and inactive segments. The duration of each segment is ΔT[i], which represents the duration of the i-th segment signal. , detect the active area of ​​vibration signal: ; is the set of sample points of the i-th segment activity signal; For each active area, calculate its duration : ; is the number of samples of the i-th segment activity signal, is the sampling frequency, used to convert the number of samples into time; According to the normal operation status or historical data of the equipment, determine the normal duration range: Tnormal = [Tmin, Tmax]; Tmin is the lower limit of the duration, and Tmax is the upper limit of the duration; Check whether the duration of each activity ΔT[i] exceeds the normal range [Tmin, Tmax]: ; Calculate the abnormal index of vibration signal duration, the expression is: ; DF is the abnormal index of vibration signal duration, is the total number of active segments, is the average of normal duration.

4. The low-power standby intelligent wake-up control method for an impact drill according to claim 3, characterized in that: The vibration signal frequency fluctuation index and the vibration signal duration anomaly index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as the input of the machine learning model. The machine learning model predicts each sample as a user operation signal or an external interference signal label based on each group of comprehensive feature vectors as the prediction target, and takes minimizing the sum of prediction errors for all samples as user operation signals or external interference signal labels as the training target. The machine learning model is trained until the sum of prediction errors converges and the model training is stopped. Each sample is determined to be a user operation signal or an external interference signal based on the model output results, wherein the machine learning model is a polynomial regression model.

5. The low-power standby intelligent wake-up control method for an impact drill according to claim 4, characterized in that: According to the predicted value yk output by the model, each sample is determined to be a user operation signal or an external interference signal, specifically including: yk=1: the signal is determined to be a user operation signal; yk=0: the signal is determined to be an external interference signal; when the signal is determined to be a user operation signal, the system triggers the wake-up mechanism to activate the main processor and high-power consumption modules; if the signal is determined to be an external interference signal, the device remains in low-power standby mode.

6. The low-power standby intelligent wake-up control method for an impact drill according to claim 1, characterized in that: In S3, according to the real-time evaluation results of the classification model, the threshold of the wake-up condition is dynamically adjusted, including the minimum duration threshold of the vibration signal and the frequency range of the vibration signal. is the minimum duration threshold, based on the false wake-up rate of the classification model and leakage wake-up rate Dynamic Adjustment : ; is the adjusted minimum duration threshold, is the initial minimum duration threshold, k is the adjustment coefficient, which is used to control the amplitude of the threshold adjustment. is the current false wake-up rate, is the current leakage wake-up rate; when : Increase , improve the wake-up conditions, when : Decrease , relax the wake-up conditions; Assume that the initial main frequency range of the vibration signal is [ , ], dynamically adjust the frequency range through feature importance evaluation of the classification model: ； , are the lower and upper limits of the adjusted frequency range, , It indicates the offset of the main frequency of the user operation signal observed in the model, which respectively represents the frequency fluctuation on the low-frequency side and the high-frequency side. α and β are adjustment coefficients used to control the expansion amplitude of the frequency range.

7. The low-power standby intelligent wake-up control method for an impact drill according to claim 1, characterized in that: In S4, the feature vector of the current wake-up signal is assumed to be ,include: ;in, is the RMS value of the amplitude, is the main frequency, is the duration of the signal; Through the pre-trained classification model M, input Then we get a predicted probability , indicating the confidence that the signal belongs to the user operation signal, the expression is: ; Assume that the judgment threshold of the classification model is θp. If the signal is judged as external interference, the current wake-up is considered to be a false wake-up behavior. For each false wake-up signal, its key feature Record to log ; The false wakeup feature set Normal user operation feature set Merge to form a new training set: ; Based on the expanded training set , update the classification model M using existing machine learning methods: ; According to the real-time statistics of false wake-up rate and leakage wake-up rate , dynamically adjust the classification threshold : ; In the formula, is the adjusted classification threshold, and k is the adjustment coefficient, which is used to control the change range of the threshold.

8. A low-power standby intelligent wake-up control system for an impact drill, used to implement the low-power standby intelligent wake-up control method for an impact drill according to any one of claims 1 to 7, characterized in that: Including low power management module, signal detection and classification module, full power control module and optimization module; Low-power management module: Through the built-in motion sensor and touch sensor, the user interaction behavior signal of the device in the working state is collected in real time. If it is detected that the device has no operation behavior within the set time, it will enter the low-power standby mode. In the low-power standby mode, only the core sensor is kept in the low-power working state, and the main motor drive circuit and high-power modules are turned off; Signal detection and classification module: detects user operation behavior signals through core sensors, and analyzes the collected signals using the built-in classification model based on the frequency characteristics, duration and pattern of vibration signals to distinguish user operation signals from external interference signals. When the signal characteristics meet the preset threshold of user operation behavior, the device wake-up signal is triggered; Full power control module: According to the real-time evaluation results of the classification model, the threshold of the wake-up condition is dynamically adjusted, including the minimum duration threshold of the vibration signal and the frequency range of the vibration signal. After the user's actual operation signal is confirmed, the main processor and high-power consumption module are activated to enter the full-power working mode; Optimization module: During the wake-up process, if a false wake-up behavior is identified, the device will automatically re-enter the low-power standby mode and record the false wake-up characteristics to optimize the wake-up logic. By continuously collecting user operation behavior data, the classification threshold is dynamically updated.

Images