Digital metering system of building robot mixing device

Abstract

The invention discloses a digital metering system for a mixing device of a building robot, which relates to the technical field of engineering automation and comprises a data acquisition module used for synchronously acquiring an original weighing signal carrying vibration noise and output by a weighing sensor of the mixing device, the real-time vibration signal is output by a vibration sensor which is installed on the building robot body and used for representing the mechanical vibration state. According to the invention, the weighing signal and the vibration signal are synchronously collected, the state identification module is combined to accurately judge the operation state, the corresponding dynamic correlation model is called to calculate and eliminate the vibration noise in real time, and the limitation of physical damping is avoided; during signal processing, model smooth switching is adopted, output jump is restrained, and metering precision and dynamic response speed are both considered; the model self-updating module can correct static state model parameters, and errors are further reduced.

Description

Technical Field

[0001] This invention relates to the field of engineering automation technology, specifically to a digital metering system for a construction robot mixing device. Background Technology

[0002] In the field of construction automation, construction robots, especially mobile and tracked concrete and mortar mixing robots, are increasingly widely used. The accuracy of their material proportions is crucial, and the weighing sensor, as the core component of the digital metering of the mixing device, directly determines the quality of the final mixture through its reading accuracy.

[0003] However, during operation, such as stirring or moving, the interaction between the power system and the ground inevitably generates continuous and complex mechanical vibrations in these robots. These vibrations originating from the equipment itself are directly transmitted to the rigidly connected load cells. Current mainstream methods for dealing with vibration interference focus on physical damping measures, such as using rubber damping pads or spring suspension systems. Physical damping has limited effectiveness in suppressing high-frequency vibrations, especially during dynamic robot movement, such as tracked robots climbing slopes or performing high-intensity tasks; the lag in the response of damping devices cannot suppress instantaneous impact vibrations.

[0004] In construction engineering, the mixing precision requirements for special concretes such as high-performance grouting materials and precision mortars such as tile adhesives are extremely high. A 1% error in mixing 10 tons of concrete amounts to 100 kg, which is enough to cause the strength grade to drop by 1-2 grades, such as from C30 to C25. More importantly, the instantaneous fluctuations in weighing readings caused by vibration can create cumulative errors that are difficult to eliminate during continuous measurement, causing the weighing error to significantly exceed the allowable range. This severely limits the ability of construction robots to achieve high-precision mixing under complex and dynamic working conditions. Therefore, the impact of vibration and noise on measurement accuracy has practical engineering hazards and is not a minor issue that can be ignored. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a digital metering system for a construction robot mixing device, which solves the problems of insufficient suppression of high-frequency vibration by physical damping measures during use, simple low-pass filtering sacrificing dynamic response speed and failing to distinguish between actual weight and vibration noise, and vibration causing cumulative measurement errors.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a digital metering system for a construction robot mixing device, the system comprising: The data acquisition module is used to synchronously acquire the original weighing signal carrying vibration noise output by the weighing sensor of the mixing device, as well as the real-time vibration signal output by the vibration sensor installed on the weighing platform body of the mixing device to characterize the vibration state of the weighing platform. The state recognition module is used to identify the current operating state of the construction robot as one of a plurality of preset operating states based on the amplitude and frequency characteristics of the real-time vibration signal; the plurality of operating states include a stationary state, a moving state, and a stirring state. The model management module is used to store a library of dynamic correlation models corresponding to the multiple operating states; wherein, each dynamic correlation model represents the mathematical mapping relationship between the real-time vibration signal and the vibration noise component in the original weighing signal under the operating state. The signal processing module is used to call the corresponding dynamic correlation model from the model management module according to the current operating state identified by the state recognition module, and calculate the vibration noise component contained in the original weighing signal in real time based on the model and the input real-time vibration signal. The influence of the vibration noise component is eliminated through compensation operation to generate a compensated weighing signal. The compensation operation includes the calculation of the ratio correction between the amplitude and frequency of the correlation noise component based on the peak value obtained from the original weighing. The metering output module is used to determine the real-time weight value of the material based on the compensated weighing signal, and output the real-time weight value to the control unit of the construction robot.

[0007] Furthermore, the data acquisition module includes: The first acquisition unit is used to perform analog-to-digital conversion on the analog voltage signal output by the weighing sensor at a first preset sampling frequency to generate time series data of the original weighing signal; The second acquisition unit is used to perform analog-to-digital conversion on the analog signal output by the vibration sensor at a second preset sampling frequency synchronized with the first preset sampling frequency, thereby generating time series data of the real-time vibration signal; wherein, the vibration sensor may be a triaxial accelerometer or a displacement vibration sensor.

[0008] Furthermore, the identification process of the state recognition module includes: The vibration spectrum is obtained by performing a fast Fourier transform on the real-time vibration signal. The dominant frequency band energy value and total energy value in the vibration spectrum are extracted as state characteristic parameters; In the state recognition process, in addition to extracting the main frequency band energy value and the total energy value, the peak factor and kurtosis of the vibration signal are also integrated as feature parameters. The peak factor is the ratio of the vibration peak value to the effective value, and the kurtosis is used to characterize the signal impact. The current operating state of the construction robot is determined by comparing the state feature parameters with the preset threshold range of feature parameters under the multiple operating states.

[0009] If the maximum vibration value exceeds the safety threshold or the tilt angle calculated by the triaxial accelerometer is greater than 5°, the current environment is determined to be unsuitable for weighing, triggering the metering output module to stop weighing and send an alarm signal.

[0010] Furthermore, the process of obtaining the threshold range of the feature parameters includes: During the calibration phase, the construction robot is controlled to perform baseline actions such as stationary movement, uniform speed movement, variable speed movement, low-speed mixing, and high-speed mixing. When performing each benchmark action, the real-time vibration signal is recorded synchronously, and the corresponding state characteristic parameters are calculated, thereby establishing a unique characteristic parameter threshold range for each operating state.

[0011] Furthermore, the process of obtaining the dynamic correlation model includes: During the calibration phase, the construction robot sequentially traverses the multiple operating states under the preset ratio of rated load conditions of the mixing device. In each operating state, raw weighing signals and real-time vibration signals for a period of time are synchronously collected as calibration datasets; For each operating state's calibration dataset, a system identification algorithm is used, taking the real-time vibration signal as input and the original weighing signal as output, to identify the transfer function, and the transfer function is stored in the model management module as the dynamic correlation model for that operating state.

[0012] Furthermore, the system identification algorithm is an adaptive identification algorithm based on recursive least squares.

[0013] Furthermore, the specific processing procedure of the signal processing module includes: Receive the current running status identifier output by the status recognition module; Based on the current running status identifier, retrieve and load the corresponding dynamic correlation model from the model management module; The real-time vibration signal at the current moment is used as the input to the dynamic correlation model; The dynamic correlation model is used to perform calculations and output a predicted vibration noise signal. The peak value of the original weighing signal at the current moment is compensated based on the predicted vibration noise signal to obtain the compensated weighing signal value at the current moment.

[0014] Furthermore, the signal processing module is also used for: When the state recognition module detects a change in the operating state, it smoothly switches the dynamic correlation model from the model management module. The smooth transition switching process includes, within a preset time window after the switching occurs, performing a weighted average of the vibration noise components calculated based on the old model and the vibration noise components calculated based on the new model. The weight of the weighted average is dynamically adjusted over time to suppress output jumps that may be caused by model switching. The weight adjustment of the weighted average is based on the stability index of the vibration signal, namely the signal variance: when the environmental variable fluctuates greatly and the variance is greater than the preset threshold, the time window is extended to 2 seconds, and the weight of the old model is slowly reduced, initially 0.8, decreasing by 0.4 per second, to reduce the impact of sudden environmental changes on the output; when the environment is stable and the variance is less than or equal to the preset threshold, the time window is kept at 1 second, and the weights are quickly transitioned to ensure dynamic response speed.

[0015] Furthermore, the metering output module is also used for: The compensated weighing signal is subjected to low-pass filtering to eliminate residual high-frequency random noise and generate the final weighing signal. Based on the rate of change of the final weighing signal, determine whether the feeding process is currently underway; When it is determined that the material feeding process is in progress, the final weighing signal is accumulated to calculate the total weight of the material that has been added, and the total weight of the material is used as the real-time weight value.

[0016] Furthermore, the system also includes: The model self-updating module is used to take the original weighing signal at this time as the zero-point reference and the real-time vibration signal at this time as the background noise during the time period when the construction robot is stationary and the weighing sensor reading is stable. The model self-updating module is further configured to correct the parameters of the dynamic correlation model corresponding to the static state stored in the model management module based on the zero-point reference and the background noise.

[0017] Beneficial effects: This invention simultaneously collects weighing and vibration signals, combines them with a state recognition module to accurately determine the operating status, and calls the corresponding dynamic correlation model to calculate and eliminate vibration noise in real time, avoiding the limitations of physical vibration reduction. During signal processing, smooth model switching is used to suppress output jumps, balancing measurement accuracy and dynamic response speed. The model self-update module can correct static state model parameters, further reducing errors. The measurement output module effectively eliminates residual noise and accurately measures the total weight of materials through low-pass filtering and material addition judgment accumulation, significantly improving the measurement accuracy of the construction robot under dynamic working conditions, reducing accumulated errors, ensuring the accuracy of material proportioning, and enhancing the system's adaptability and stability in complex operating environments. Attached Figure Description

[0018] Figure 1 This is a system structure diagram of the present invention; Figure 2 This is a flowchart of the weighing signal of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Please see Figure 1 This invention provides a digital metering system for a construction robot mixing device, the system comprising: The data acquisition module is used to synchronously acquire the original weighing signal carrying vibration noise output by the weighing sensor of the mixing device, as well as the real-time vibration signal output by the vibration sensor installed on the weighing platform body of the mixing device to characterize the vibration state of the weighing platform. The state recognition module is used to identify the current operating state of the construction robot as one of a number of preset operating states based on the amplitude and frequency characteristics of the real-time vibration signal; the multiple operating states include stationary state, moving state and stirring state. The model management module is used to store a library of dynamic correlation models corresponding to multiple operating states. Each dynamic correlation model represents the mathematical mapping relationship between the real-time vibration signal and the vibration noise component in the original weighing signal under that operating state. The signal processing module is used to call the corresponding dynamic correlation model from the model management module according to the current operating state identified by the state recognition module, and calculate the vibration noise component contained in the original weighing signal in real time based on the model and the input real-time vibration signal. Through compensation operation, a compensated weighing signal is generated. The compensation operation includes the calculation of the ratio of the amplitude and frequency of the correlation noise component based on the peak value obtained from the original weighing. The metering output module is used to determine the real-time weight value of the material based on the compensated weighing signal and output the real-time weight value to the control unit of the construction robot.

[0021] Specifically, in the data acquisition module, the weighing sensor can be a strain gauge type, installed on the load-bearing structure of the mixing device. Its output raw weighing signal is transmitted to the data acquisition interface via a shielded cable. In this embodiment, a piezoelectric triaxial accelerometer or a displacement vibration sensor is selected as the vibration sensor to ensure that the acquired vibration signal is directly correlated with the vibration of the weighing platform, avoiding signal distortion caused by distance from the weighing area. The signals from the two sensors are synchronized using the same clock source, ensuring that the acquired raw weighing signal and real-time vibration signal correspond in time.

[0022] When analyzing real-time vibration signals, the state recognition module first extracts the amplitude and frequency distribution of the signal. For example, in a stationary state, the amplitude of the real-time vibration signal is small and the frequency components are simple; in a moving state, the amplitude increases and has specific frequency components; and in a stirring state, the amplitude and frequency changes are more complex. Based on these characteristics, the current operating state is identified as the corresponding state. The dynamic correlation model library in the model management module is built through a large amount of experimental data. Each operating state corresponds to a model, which can accurately reflect the relationship between the real-time vibration signal and the vibration noise component in the original weighing signal under this state.

[0023] After the signal processing module calls the corresponding model, it inputs the real-time vibration signal into the model, calculates the vibration noise component, and then performs a compensation operation based on the noise component from the original weighing signal to obtain a more accurate compensated weighing signal. The compensation operation method is the same as the compensation described above.

[0024] The metering output module calculates the real-time weight of the material based on the compensated weighing signal and transmits it to the control unit of the construction robot, enabling the control unit to more accurately control the mixing process. This system effectively reduces the impact of vibration on the metering results, improves the accuracy of material proportioning, and particularly meets the mixing accuracy requirements during operation or non-stationary conditions, thereby increasing work efficiency.

[0025] The present invention further proposes a data acquisition module, including: The first acquisition unit is used to perform analog-to-digital conversion on the analog voltage signal output by the weighing sensor at a first preset sampling frequency to generate time series data of the original weighing signal. The second acquisition unit is used to perform analog-to-digital conversion on the analog signal output by the vibration sensor at a second preset sampling frequency synchronized with the first preset sampling frequency, thereby generating time series data of the real-time vibration signal.

[0026] Specifically, the first acquisition unit samples and quantizes the analog voltage signal output by the weighing sensor, converting it into a digital signal to generate time-series data of the original weighing signal. Each data point includes a corresponding timestamp and voltage value. The second acquisition unit performs analog-to-digital conversion on the analog signal output by the triaxial accelerometer, generating time-series data of the real-time vibration signal. Each data point includes a timestamp and acceleration values ​​in three directions. Through this setup, the data acquisition module can accurately and synchronously acquire time-series data of both signals, providing a reliable data foundation for subsequent signal processing and state identification, thus helping to improve the overall measurement accuracy of the system.

[0027] This invention further proposes a state recognition module, the recognition process of which includes: The vibration spectrum is obtained by performing a fast Fourier transform on the real-time vibration signal. The dominant frequency band energy value and total energy value in the vibration spectrum are extracted as state characteristic parameters; In the state recognition process, in addition to extracting the main frequency band energy value and the total energy value, the peak factor and kurtosis of the vibration signal are also integrated as feature parameters. The peak factor is the ratio of the vibration peak value to the effective value, and the kurtosis is used to characterize the signal impact. The current operating state of the construction robot is determined by comparing the state feature parameters with the preset threshold range of feature parameters under multiple operating states.

[0028] Specifically, when processing real-time vibration signals, the state recognition module first performs a Fast Fourier Transform (FFT) on the signals. This transform converts the time-domain signal into a frequency-domain vibration spectrum, obtaining the signal amplitude at different frequencies.

[0029] In spectral analysis, it is crucial to extract the primary frequencies corresponding to large vibration amplitudes: by traversing the amplitudes of each frequency point in the spectrum, the top three characteristic frequencies with the largest amplitudes, f1, f2, and f3, are selected and converted into angular frequencies. , recorded as The angular frequency with the largest amplitude is denoted as . The main frequency corresponding to the maximum vibration amplitude will be used as the core characteristic parameter for subsequent quantitative calculation of vibration noise.

[0030] Next, extract the corresponding values ​​from the vibration spectrum. The main frequency band energy value and total energy value of the frequency band, the main frequency band energy value is obtained by analyzing the energy value of the main frequency band. The total energy value is obtained by integrating the square of the spectral amplitude within a range of ±5. The total energy value is obtained by integrating the square of the spectral amplitude across the entire frequency range. The peak factor and kurtosis of the vibration signal are also incorporated as supplementary characteristic parameters. The peak factor is the ratio of the peak value to the effective value of the vibration, and the kurtosis is used to characterize the impulsiveness of the signal.

[0031] Finally, the above characteristic parameters are compared with the preset threshold ranges for characteristic parameters under various operating conditions: Combining traditional vibration reduction methods with the large mass impact of building materials, the oscillation frequency and amplitude of the weighing platform are smaller than those of typical mechanical vibrations. Although the frequency differs under different vibration states, the dominant angular frequency can be considered... If the characteristic parameters fall within the threshold range of a certain operating state, then the current operating state is determined to be that state. This is combined with the main angular frequency. and amplitude The identification method can more accurately distinguish different operating states, providing core parameter support for subsequent dynamic correlation model calls.

[0032] This invention further proposes a threshold range for feature parameters, the acquisition process of which includes: During the calibration phase, the construction robot is controlled to perform baseline actions such as stationary movement, uniform movement, variable speed movement, low-speed mixing, and high-speed mixing. When performing each benchmark action, the real-time vibration signal is recorded synchronously, and the corresponding state characteristic parameters are calculated, thereby establishing a unique characteristic parameter threshold range for each operating state.

[0033] Specifically, during the calibration phase, the construction robot's control system controls it to perform various baseline actions. For example, when the robot is stationary, its power system is ensured to be off and it is on a stable surface; during uniform movement, the robot is controlled to travel at a fixed speed on a flat surface; during variable-speed movement, the robot accelerates and decelerates; and for low-speed and high-speed mixing, the mixing mechanism is controlled to operate at different fixed speeds. During each baseline action, the data acquisition module synchronously records real-time vibration signals, and the state recognition module processes these signals to calculate the corresponding dominant frequency band energy value and total energy value, among other state characteristic parameters. For the stationary state, multiple sets of data are collected to determine the range of characteristic parameters, forming a threshold interval for that state; for the moving state, the characteristic parameters during uniform and variable-speed movement are combined to determine the threshold interval for the moving state; and for the mixing state, the characteristic parameters during low-speed and high-speed mixing are combined to determine the threshold interval for the mixing state. Low-speed stirring refers to a stirring mode with a stirring mechanism speed of no more than 150 r / min, while high-speed stirring refers to a stirring mode with a stirring mechanism speed of 200-400 r / min. The threshold range of the exclusive characteristic parameters for each operating state is thus established. This calibration method enables subsequent state identification to be more accurate and reliable, laying the foundation for the stable operation of the system.

[0034] This invention further proposes a dynamic correlation model, the acquisition process of which includes: In this embodiment, during the calibration phase, the construction robot sequentially traverses multiple operating states under the conditions of the mixing device being at 0%, 25%, 50%, and 75% of its rated load, respectively. In each operating state, raw weighing signals and real-time vibration signals for a period of time are synchronously collected as calibration datasets; For each operating state's calibration dataset, a system identification algorithm is used, taking the real-time vibration signal as input and the original weighing signal as output, to identify the transfer function. The transfer function is then stored in the model management module as the dynamic correlation model for that operating state.

[0035] Specifically, during the calibration phase, the mixing device was placed under operating conditions of 0%, 25%, 50%, and 75% of its rated load, respectively. These loads correspond to actual mass... These masses are the material masses that need to be weighed in this system. Based on this, the construction robot is controlled to sequentially enter a stationary state, a moving state, and a mixing state. The moving state includes two modes: constant speed and variable speed, and the mixing state includes two modes: low speed and high speed.

[0036] In each operating state, the running time is maintained at no less than 5 minutes, during which the data acquisition module synchronously collects the raw weighing signal. and real-time vibration signals Simultaneously, the main angular frequency of the current state is recorded in real time through the state recognition module. and vibration amplitude vibration amplitude By analyzing vibration signals The peak value was extracted in the time domain, and the final result contained " "A multidimensional calibration dataset".

[0037] For each operating state's calibration dataset, an adaptive identification algorithm based on recursive least squares is used to identify the "real-time vibration signal". Main angular frequency ,amplitude "As a combined input, the original weighing signal" The difference between the actual mass and the theoretical weighing value is output as the "difference between the actual mass and the theoretical weighing value". The actual material mass, i.e., the mass that needs to be weighed in this system. , Let gravitational acceleration be the acceleration due to gravity, and its value be [value]. .

[0038] The transfer function is identified using this algorithm. The transfer function needs to be explicitly stated. and The specific expression for the impact on vibration and noise is: in This refers to the weighing error component caused by vibration noise, and this component must satisfy the simple harmonic motion error law of this system: when the construction robot is in a stationary state, its body vibration can be simplified to single-frequency simple harmonic motion. The dominant angular frequency, The vibration amplitude and the expression for the vibration displacement signal are: , where t is a time variable; at this time This formula is derived from the acceleration formula. acceleration And weighing value Therefore, the error When it becomes clear Then, it can be based on its factors such as Size, clear impact factor size, impact factor During weighing The adjusted accurate value can be obtained by multiplying the peak value by the influence factor C; Finally, this transfer function is stored in the model management module as the dynamic correlation model for the corresponding operating state, ensuring that the model can quantitatively correlate the main frequencies. ,amplitude This avoids the logical flaw of simply subtracting the frequency signal from the weighing value, thus mitigating vibration and noise errors.

[0039] This invention further proposes an adaptive identification algorithm based on recursive least squares for system identification.

[0040] Specifically, when the adaptive identification algorithm based on recursive least squares is used to obtain the dynamic correlation model, its basic formula is: , in, for The parameter estimation vector at time step [time]. for The parameter estimation vector at time step [time]. Here is the gain matrix. for The output signal at that moment, i.e., the original weighing signal. for The input vector at any given time, i.e., the vector composed of real-time vibration signals. for The transpose of .

[0041] Gain matrix The calculation formula is: , in The forgetting factor takes a value between 0 and 1. for The covariance matrix at time t.

[0042] The formula for updating the covariance matrix is: , in It is the identity matrix. In practical applications, the forgetting factor... The value can be set to 0.95 to balance the algorithm's tracking capability and stability. This algorithm adaptively tracks changes in system characteristics and accurately identifies the transfer function by continuously recursively updating parameter estimates. The dynamic correlation model obtained using this algorithm has high accuracy and better reflects the relationship between real-time vibration signals and vibration noise components, thereby improving the noise compensation effect.

[0043] This invention further proposes a specific processing procedure for the signal processing module, including: Receive the current running status identifier output by the status recognition module; Based on the current running status identifier, retrieve and load the corresponding dynamic correlation model from the model management module; Use the real-time vibration signal at the current moment as the input to the dynamic correlation model; The dynamic correlation model is used to calculate and output a predicted vibration noise signal. Multiply the peak value of the original weighing signal at the current moment by the influence factor of the predicted vibration noise signal to obtain the compensated weighing signal value at the current moment.

[0044] Specifically, the signal processing module first receives the current operating status identifier output by the status recognition module, such as "high-speed stirring status," and simultaneously receives the core parameters corresponding to this status, including the main angular frequency. ,amplitude and actual load quality .

[0045] Based on the running status identifier, retrieve and load the corresponding dynamic correlation model from the model management module. The real-time vibration signal at the current moment , , As the joint input to the model, the model outputs the weighing error component corresponding to the vibration noise. : If the current vibration is a single-frequency simple harmonic motion, then the error calculation based on the simple harmonic motion error model of this system is as follows: The calculation formula is: ,in The initial phase of the vibration is obtained through time-domain analysis of the vibration signal; the influence factor for single-frequency simple harmonic vibration is C=g / (g+ ; If the current vibration is not a single-frequency simple harmonic motion, according to the Fourier transform principle, it can also be regarded as a superposition of multiple-frequency simple harmonic motions. Assuming it is simplified to a superposition of vibrations of two frequencies, the vibration signal expression is: At this point, the error components are calculated using a superposition operation, and the expression is: .

[0046] The influencing factor of dual-frequency simple harmonic vibration is ; The compensation correction formula is: ,in This is the error after vibration compensation. This is the original error. Vibration influence factor; This process strictly follows the logical chain of "frequency signal → angular frequency conversion → error component calculation → compensation correction", avoiding the problem of "directly subtracting the frequency weighing unit to obtain the accurate weighing value", and achieving accurate restoration of the converted value.

[0047] The present invention further proposes that the signal processing module is also used for: When the state recognition module detects a change in the running state, it smoothly switches the dynamic correlation model from the model management module. The smooth transition switching process includes, within a preset time window after the switching occurs, performing a weighted average of the vibration and noise components calculated based on the old model and the vibration and noise components calculated based on the new model. The weight of the weighted average is dynamically adjusted over time to suppress output jumps that may be caused by model switching. The weight adjustment of the weighted average is based on the stability index of the vibration signal, namely the signal variance: when the environmental variable fluctuates greatly and the variance is greater than the preset threshold, the time window is extended to 2 seconds, and the weight of the old model is slowly reduced, initially 0.8, decreasing by 0.4 per second, to reduce the impact of sudden environmental changes on the output; when the environment is stable and the variance is less than or equal to the preset threshold, the time window is kept at 1 second, and the weights are quickly transitioned to ensure dynamic response speed.

[0048] Specifically, when the state recognition module detects a change in operating state, such as switching from a moving state to a stirring state, the signal processing module initiates a smooth transition switching mechanism: within a preset time window, it processes the data calculated based on the old model. and calculations based on the new model Perform a weighted average, where Including the old state, , Including new states, The weight adjustment is based on the stability index of the vibration signal, namely the signal variance: When environmental fluctuations are significant and the signal variance exceeds a preset threshold, the time window is extended to 2 seconds. The initial weights of the old model are set to 0.8, decreasing by 0.4 per second, while simultaneously updating... and The transition parameters are expressed as follows: This reduces the impact of sudden changes in frequency and amplitude on the output; When the environment is stable and the signal variance does not exceed the preset threshold, the time window is maintained at 1 second, and the weights are quickly transitioned to ensure dynamic response speed.

[0049] The present invention further proposes that the metering output module is also used for: The compensated weighing signal is low-pass filtered to eliminate residual high-frequency random noise and generate the final weighing signal. Based on the rate of change of the final weighing signal, determine whether the feeding process is currently underway; When it is determined that the material is being added, the final weighing signal is accumulated to calculate the total weight of the material that has been added, and the total weight of the material is used as the real-time weight value.

[0050] Specifically, the metering output module first processes the compensated weighing signal. Butterworth low-pass filtering is performed with a cutoff frequency of 10Hz to eliminate residual high-frequency random noise. This high-frequency random noise refers to vibration interference signals with frequencies above 50Hz and randomly fluctuating amplitudes. This type of noise mainly originates from high-frequency vibrations of the equipment's motor or external environmental interference. These noises are non-... Interference in the frequency band ultimately generates the final weighing signal. .

[0051] Subsequently according to The rate of change is used to determine whether a feeding process is underway; the expression for the rate of change is: ,in The final rate of change of the weighing signal. This represents the final change in the weighing signal. This is the change in weight over time during the weighing process; when When this value is a preset threshold, it is determined that the current state is in the feeding state, and at this time... Perform a sliding window summation or integration, with the summation or integration window size set to 0.5 seconds. Calculate the total weight of the added material by accumulating the summation or integration. For example, the total weight expression is: .

[0052] In this process, it is necessary to clearly introduce the concept of error: Based on the calibration data of this system and the vibration error calculation model, the weighing error of the current system mainly comes from the dynamic error caused by vibration, and the maximum amplitude of this error satisfies the following formula: in The actual material mass, i.e., the mass that needs to be weighed in this system. , The dominant angular frequency, For vibration amplitude, This is the acceleration due to gravity. Verification was performed through a calibration experiment: The limit is ≤10 rad / s. The error is limited to ≤0.5cm, and the maximum error before adjustment does not exceed 5%. After adjustment by the influencing factor, this error range can meet the proportioning accuracy requirements in most construction projects. For scenarios with higher requirements for proportioning accuracy, such as special concrete (e.g., high-performance grouting materials) and precision mortar (e.g., tile adhesive), the error can be increased by improving the accuracy of the proportioning. The requirement is to constrain the weighing scenario to achieve accurate weighing.

[0053] To further ensure weighing accuracy, optimization is performed through error calculation and calibration: During the calibration phase, different load masses are considered. and different principal angular frequencies Calibration was conducted with load capacities covering 0%, 25%, 50%, and 75% of rated load, and main angular frequencies covering 10-50 rad / s. The corresponding error coefficients were pre-calculated. The error coefficient is linearly correlated with the aforementioned influencing factor C, and is adjusted after calibration. Simultaneously, a " "Three-dimensional calibration table. During actual weighing, the metering output module calls this calibration table in real time, based on the current..." and actual measurement (This value is approximate) The final weighing value is then corrected a second time using the following formula:

[0054] When applying the correction formula, The peak value of the weighing time window should be taken. In this way, even with fluctuations in vibration frequency and amplitude, the cumulative error can still be controlled within ±0.3% through dynamic adjustment of the error coefficient. This value is far lower than the industry standard of ±1%, ultimately resulting in a corrected error. (or The weight value is output to the control unit of the construction robot as a real-time weight value.

[0055] It should be noted that this system only addresses weighing errors caused by mechanical vibration. Mechanical vibration is caused by… and The error caused by non-vibration factors such as material residue on the inner wall of the mixing device and sensor aging needs to be avoided through routine cleaning and maintenance or periodic sensor calibration, which does not conflict with the vibration error compensation logic of this solution.

[0056] The present invention further proposes a model self-updating module, which is used to take the original weighing signal at this time as the zero point reference and the real-time vibration signal at this time as the background noise during the time period when the construction robot is stationary and the weighing sensor reading is stable. The model self-update module is also used to correct the parameters of the dynamic correlation model corresponding to the static state stored in the model management module based on the zero-point reference and background noise.

[0057] Specifically, the model self-updating module monitors the operating status of the construction robot and the weighing sensor readings in real time: when it detects that the robot is stationary and the weighing sensor readings change within 30 seconds... When the load cell reading is stable, the load cell reading is considered stable. The determination of a stationary state is based on the principal angular frequency. And amplitude .

[0058] At this point, the current original weighing signal will be... As the zero-point reference, the expression for the zero-point reference is: Simultaneously, collect the current real-time vibration signal. Extract the angular frequency corresponding to the background noise in the signal. and amplitude ,generally and And calculate the background noise error component. The calculation expression is: ,in This refers to the unloaded mass of the mixing device.

[0059] Subsequently, based on the zero-point reference... and background noise error The dynamic correlation model parameters in the static state of the model management module are corrected: the original model is adjusted. Updated to The updated formula is as follows This ensures the model accurately reflects the influence of extremely low-frequency background noise under static conditions, further reducing errors caused by zero-point drift, and is consistent with the previous model based on... and The error compensation logic remains consistent.

[0060] In summary, this invention simultaneously acquires weighing and vibration signals, accurately determines the operating status using a state recognition module, and calls the corresponding dynamic correlation model to calculate and eliminate vibration noise in real time, thus avoiding the limitations of physical vibration reduction. During signal processing, smooth model switching is employed to suppress output jumps, balancing measurement accuracy and dynamic response speed. The model self-update module can correct static state model parameters, further reducing errors. The measurement output module effectively eliminates residual noise and accurately measures the total weight of materials through low-pass filtering and material addition judgment accumulation, significantly improving the measurement accuracy of the construction robot under dynamic working conditions, reducing accumulated errors, ensuring the accuracy of material proportioning, and enhancing the system's adaptability and stability in complex operating environments.

[0061] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0062] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A digital metering system for a construction robot mixing device, characterized in that, The system includes: The data acquisition module is used to synchronously acquire the original weighing signal carrying vibration noise output by the weighing sensor of the mixing device, as well as the real-time vibration signal output by the vibration sensor installed on the weighing platform body of the mixing device to characterize the vibration state of the weighing platform. The state recognition module is used to identify the current operating state of the construction robot as one of a plurality of preset operating states based on the amplitude and frequency characteristics of the real-time vibration signal; the plurality of operating states include a stationary state, a moving state, and a stirring state. The model management module is used to store a library of dynamic correlation models corresponding to the multiple operating states; wherein, each dynamic correlation model represents the mathematical mapping relationship between the real-time vibration signal and the vibration noise component in the original weighing signal under the operating state. The signal processing module is used to call the corresponding dynamic correlation model from the model management module according to the current operating state identified by the state recognition module, and calculate the vibration noise component contained in the original weighing signal in real time based on the model and the input real-time vibration signal. Through compensation operation, a compensated weighing signal is generated. The compensation operation includes the calculation of the ratio correction of the amplitude and frequency of the correlation noise component based on the peak value obtained from the original weighing. The metering output module is used to determine the real-time weight value of the material based on the compensated weighing signal, and output the real-time weight value to the control unit of the construction robot.

2. The digital metering system for the construction robot mixing device according to claim 1, characterized in that, The data acquisition module includes: The first acquisition unit is used to perform analog-to-digital conversion on the analog voltage signal output by the weighing sensor at a first preset sampling frequency to generate time series data of the original weighing signal; The second acquisition unit is used to perform analog-to-digital conversion on the analog signal output by the vibration sensor at a second preset sampling frequency synchronized with the first preset sampling frequency, thereby generating time series data of the real-time vibration signal; wherein the vibration sensor is a triaxial accelerometer or a displacement vibration sensor.

3. The digital metering system for the construction robot mixing device according to claim 1, characterized in that, The identification process of the status identification module includes: The vibration spectrum is obtained by performing a fast Fourier transform on the real-time vibration signal. The dominant frequency band energy value and total energy value in the vibration spectrum are extracted as state characteristic parameters; In the state recognition process, in addition to extracting the main frequency band energy value and the total energy value, the peak factor and kurtosis of the vibration signal are also integrated as feature parameters. The peak factor is the ratio of the vibration peak value to the effective value, and the kurtosis is used to characterize the signal impact. The current operating state of the construction robot is determined by comparing the state feature parameters with the preset threshold range of feature parameters under the multiple operating states.

4. The digital metering system for the construction robot mixing device according to claim 3, characterized in that, The process of obtaining the threshold range of the feature parameters includes: During the calibration phase, the construction robot is controlled to perform baseline actions such as stationary movement, uniform speed movement, variable speed movement, low-speed mixing, and high-speed mixing. When performing each benchmark action, the real-time vibration signal is recorded synchronously, and the corresponding state characteristic parameters are calculated, thereby establishing a unique characteristic parameter threshold range for each operating state.

5. The digital metering system for the construction robot mixing device according to claim 1, characterized in that, The process of obtaining the dynamic correlation model includes: During the calibration phase, the construction robot sequentially traverses the multiple operating states under the preset ratio of rated load conditions of the mixing device. In each operating state, raw weighing signals and real-time vibration signals for a period of time are synchronously collected as calibration datasets; For each operating state's calibration dataset, a system identification algorithm is used, taking the real-time vibration signal as input and the original weighing signal as output, to identify the transfer function, and the transfer function is stored in the model management module as the dynamic correlation model for that operating state.

6. The digital metering system for the construction robot mixing device according to claim 5, characterized in that, The system identification algorithm is an adaptive identification algorithm based on recursive least squares.

7. The digital metering system for the construction robot mixing device according to claim 1, characterized in that, The specific processing steps of the signal processing module include: Receive the current running status identifier output by the status recognition module; Based on the current running status identifier, retrieve and load the corresponding dynamic correlation model from the model management module; The real-time vibration signal at the current moment is used as the input to the dynamic correlation model; The dynamic correlation model is used to perform calculations and output a predicted vibration noise signal. The peak value of the original weighing signal at the current moment is compensated based on the predicted vibration noise signal to obtain the compensated weighing signal value at the current moment.

8. The digital metering system for the construction robot mixing device according to claim 7, characterized in that, The signal processing module is also used for: When the state recognition module detects a change in the operating state, it smoothly switches the dynamic correlation model from the model management module. The smooth transition switching process includes performing a weighted average of the vibration noise components calculated based on the old model and the vibration noise components calculated based on the new model within a preset time window after the switching occurs. The weight of the weighted average is dynamically adjusted over time to suppress output jumps that may be caused by model switching.

9. The digital metering system for the construction robot mixing device according to claim 1, characterized in that, The metering output module is also used for: The compensated weighing signal is subjected to low-pass filtering to eliminate residual high-frequency random noise and generate the final weighing signal. Based on the rate of change of the final weighing signal, determine whether the feeding process is currently underway; When it is determined that the material feeding process is in progress, the final weighing signal is accumulated to calculate the total weight of the material that has been added, and the total weight of the material is used as the real-time weight value.

10. The digital metering system for the construction robot mixing device according to claim 1, characterized in that, The system also includes: The model self-updating module is used to take the original weighing signal at this time as the zero-point reference and the real-time vibration signal at this time as the background noise during the time period when the construction robot is stationary and the weighing sensor reading is stable. The model self-updating module is further configured to correct the parameters of the dynamic correlation model corresponding to the static state stored in the model management module based on the zero-point reference and the background noise.

Images