Urban rail transit floor cleaning and maintaining device control method and system

Abstract

The present application relates to the technical field of urban rail transit floor cleaning and maintenance equipment control, and provides a kind of urban rail transit floor cleaning and maintenance equipment control method and system, by real-time acquisition of the load information of motor driving cleaning brush disc and frequency analysis is carried out, identify the load fluctuation characteristics related to the rotation frequency of cleaning brush disc;According to the amplitude and phase change of load fluctuation characteristics, determine that the bristles of cleaning brush disc are in non-uniform wear state, adaptively adjust the closed-loop control parameters of motor / overlap the compensation signal synchronous with the rotation angle of cleaning brush disc in the control command of motor;The motor is controlled by the adjusted closed-loop control parameters and / or control command of motor, so as to dynamically optimize the motor control according to the actual wear condition of bristles, so that the cleaning force applied to the ground by the cleaning brush disc remains uniform during the entire rotation process, significantly improving the cleaning quality.

Description

Technical Field

[0001] This application relates to the field of control technology for floor cleaning and maintenance equipment in urban rail transit, and more specifically, to a control method and system for floor cleaning and maintenance equipment in urban rail transit. Background Technology

[0002] Automated floor cleaning equipment is widely used in urban rail transit to maintain platform cleanliness. Its operation relies on closed-loop precise control of the drive motor to ensure stable brush rotation speed and accurate path. However, over long-term operation, the brushes experience uneven wear due to the complex surface and uneven friction, causing periodic fluctuations in motor load. The equipment's preset fixed control parameters are difficult to adapt to such dynamic load changes, resulting in slight fluctuations in motor speed, uneven cleaning intensity, and uncleaned areas appearing on the floor after cleaning, resembling "stripes" or "patches."

[0003] Over time, these stains harden, forming stubborn dirt. Because the wear patterns are relatively fixed, repeated ineffective cleaning of the same areas leads to a continuous accumulation of problems. Ultimately, manual secondary cleaning or localized treatment is required, increasing operating costs and workload, and reducing the efficiency of the automated system. This reflects the inadequacy of current control methods in responding to dynamic changes in the internal state of equipment.

[0004] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0005] This application discloses a control method and system for urban rail transit floor cleaning and maintenance equipment, which aims to solve the technical problems of uneven cleaning force, reduced cleaning effect, and increased manual intervention caused by uneven wear of the cleaning brush bristles during long-term operation of urban rail transit floor cleaning and maintenance equipment.

[0006] The technical solution of this application is as follows: In a first aspect, this application discloses a control method for urban rail transit floor cleaning and maintenance equipment, the cleaning and maintenance equipment including a cleaning brush and a motor, the method comprising: The load information of the motor driving the cleaning brush is collected in real time; this load information includes current information and torque information. Frequency analysis was performed on the load information to identify load fluctuation characteristics related to the rotation frequency of the cleaning brush. Based on the amplitude and phase changes of the load fluctuation characteristics, it is determined that the bristles of the cleaning brush are in a non-uniform wear state. Based on the fact that the bristles are in a non-uniform wear state, the closed-loop control parameters of the motor are adaptively adjusted and / or a compensation signal synchronized with the rotation angle of the cleaning brush is superimposed on the control command of the motor to counteract the periodic load fluctuations caused by the non-uniform wear of the bristles. Based on the closed-loop control parameters and / or the motor control command, the motor is controlled to ensure that the cleaning brush applies uniform cleaning force to the floor.

[0007] Furthermore, frequency analysis is performed on the load information to identify load fluctuation characteristics related to the rotation frequency of the cleaning brush, including: performing real-time frequency analysis on the collected load information and extracting specific frequency components related to the rotation frequency of the cleaning brush as the load fluctuation characteristics.

[0008] Based on this, according to the amplitude and phase changes of the load fluctuation characteristics, the brush bristles of the cleaning brush are determined to be in a non-uniform wear state, including: after the initial debugging of the equipment or after replacing the brush, establishing a reference load frequency spectrum of the motor, and storing the amplitude and phase of the specific frequency component as reference amplitude and reference phase; during the operation of the equipment, calculating the current amplitude and current phase of the current specific frequency component in real time; when the increase of the current amplitude relative to the reference amplitude continuously exceeds a first preset threshold; and the offset of the current phase relative to the reference phase continuously exceeds a second preset threshold, the non-uniform wear state of the brush bristles of the cleaning brush is determined.

[0009] In some preferred embodiments, the motor employs a PID closed-loop controller; the method further includes: determining the severity and location of the uneven wear while determining that the bristles of the cleaning brush are in a non-uniform wear state; the adaptive adjustment of the closed-loop control parameters of the motor includes: dynamically adjusting at least one of the proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd of the PID closed-loop controller according to the severity and location of the uneven wear and a preset parameter mapping table.

[0010] As an optional solution, based on the non-uniform wear state of the brush bristles, the closed-loop control parameters of the motor are adaptively adjusted and / or a compensation signal synchronized with the rotation angle of the cleaning brush disc is superimposed on the control command of the motor. This includes: collecting real-time load data of the motor and the location information of the cleaning and maintenance equipment; performing frequency analysis, transient characteristic analysis, and spectral entropy analysis on the real-time load data to obtain load fluctuation characteristics; determining whether the load fluctuation characteristics are caused by non-uniform wear of the brush bristles or localized ground contamination; if the load fluctuation characteristics are determined to be caused by non-uniform wear of the brush bristles, adjusting the closed-loop control parameters of the drive motor, or obtaining the rotation angle of the cleaning brush disc and generating a compensation signal synchronized with the rotation angle, superimposing the compensation signal on the control command of the motor; if the load fluctuation characteristics are determined to be caused by localized ground contamination, adjusting the enhanced cleaning strategy.

[0011] Furthermore, determining whether the load fluctuation characteristic originates from uneven wear of the brush bristles or localized ground contamination includes: simultaneously acquiring real-time load data of the motor and the rotation angle of the cleaning brush disc; performing preliminary analysis of the real-time load data to identify periodic load fluctuations; applying a perturbation signal to the motor and acquiring the motor's transient response data to the perturbation signal; analyzing the transient response data, extracting response features, and determining the physical source of the periodic load fluctuation based on the response features; the physical source includes uneven wear of the brush bristles and localized ground contamination.

[0012] Based on the above, the perturbation signal is a high-frequency narrow pulse current signal; the frequency of the high-frequency narrow pulse current signal is higher than the highest harmonic of the rotation frequency of the cleaning brush. A perturbation signal is applied to the motor, and the transient response data of the motor to the perturbation signal is collected, including: acquiring the current ambient noise and the brush speed of the cleaning brush; adjusting the amplitude and duration of the perturbation signal according to the ambient noise and the brush speed to ensure that the energy of the perturbation signal is higher than the ambient noise and the duration of the perturbation signal is less than the time required for the cleaning brush to rotate one angle interval; and simultaneously collecting motor current, torque and brush vibration data as the transient response data.

[0013] More specifically, determining the physical source of the periodic load fluctuation based on the response characteristics includes: acquiring the rotation angle of the cleaning brush and the response characteristics of the transient response data, and correlating the response characteristics with the rotation angle of the cleaning brush; determining the physical source of the load fluctuation based on the distribution pattern of the response characteristics at the rotation angle of the cleaning brush.

[0014] In one embodiment, the response characteristics include response time, peak overshoot, decay rate, and oscillation frequency; wherein, analyzing the transient response data, extracting the response characteristics, and determining the physical source of the periodic load fluctuation based on the response characteristics includes: if the response time is less than a first response time threshold and the peak overshoot is less than a first overshoot threshold, then the load fluctuation is determined to originate from uneven wear of the brush bristles; if the response time is greater than a second response time threshold and the decay rate is less than a first decay rate threshold, then the load fluctuation is determined to originate from localized ground contamination.

[0015] Secondly, this application also discloses a control system for urban rail transit floor cleaning and maintenance equipment. The cleaning and maintenance equipment includes a cleaning brush and a motor. The system includes: a data acquisition module for real-time acquisition of load information of the motor driving the cleaning brush; the load information includes current information and torque information; an analysis module for frequency analysis of the load information to identify load fluctuation characteristics related to the rotation frequency of the cleaning brush; a determination module for determining that the bristles of the cleaning brush are in a non-uniform wear state based on the amplitude and phase changes of the load fluctuation characteristics; an adjustment module for adaptively adjusting the closed-loop control parameters of the motor and / or superimposing a compensation signal synchronized with the rotation angle of the cleaning brush into the control command of the motor to counteract the periodic load fluctuations caused by the non-uniform wear of the bristles, based on the non-uniform wear state of the bristles; and a control module for controlling the motor to ensure uniform cleaning force of the cleaning brush on the floor, based on the closed-loop control parameters and / or the control command of the motor. Beneficial effects

[0016] The control method for urban rail transit floor cleaning and maintenance equipment disclosed in this application, by real-time acquisition of the load information of the motor driving the cleaning brush, and frequency analysis of this load information, can identify load fluctuation characteristics related to the rotation frequency of the cleaning brush. Based on this, according to the amplitude and phase changes of the load fluctuation characteristics, it can accurately determine whether the bristles of the cleaning brush are in a non-uniform wear state. Once non-uniform wear is identified, the method can adaptively adjust the closed-loop control parameters of the motor, or superimpose a compensation signal synchronized with the rotation angle of the cleaning brush onto the motor control command, thereby effectively counteracting the periodic load fluctuations caused by non-uniform bristle wear. Finally, through the adjusted closed-loop control parameters and / or the motor control command, the motor is controlled to ensure uniform cleaning force of the cleaning brush on the floor.

[0017] This technical solution effectively solves the problems of uneven cleaning force, reduced cleaning effect, and increased manual intervention caused by uneven bristle wear in existing technologies. Through real-time monitoring and frequency analysis of motor load information, this application can accurately capture subtle load changes caused by bristle wear and make intelligent judgments accordingly. Compared with existing solutions using fixed control parameters, this application's adaptive adjustment mechanism can dynamically optimize motor control based on the actual wear of the bristles, ensuring that the cleaning force applied to the ground by the cleaning brush disc remains uniform throughout its rotation, significantly improving cleaning quality. Furthermore, by offsetting periodic load fluctuations, it avoids the formation of "stripes" or "patches" of incomplete cleaning areas caused by uneven cleaning force, thereby reducing reliance on secondary manual cleaning, lowering operating costs and the workload of maintenance personnel, and greatly improving the autonomy and efficiency of automated equipment. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the steps of the control method for urban rail transit floor cleaning and maintenance equipment disclosed in an embodiment of the present invention; Figure 2 This is a schematic diagram of the control system structure of the urban rail transit floor cleaning and maintenance equipment disclosed in an embodiment of the present invention. Detailed Implementation

[0019] The implementation details of the technical solution in this embodiment are described in detail below: Traditional urban rail transit floor cleaning and maintenance equipment inevitably experiences uneven wear on its brush bristles during prolonged operation due to the complex ground environment and continuous friction. This wear causes periodic fluctuations in the load on the drive motor, which the equipment's original fixed closed-loop control parameters cannot effectively handle. Consequently, the cleaning force of the brush bristles on the ground is uneven, resulting in poor cleaning performance and increasing the frequency and cost of manual intervention.

[0020] In response, this application proposes a control method for urban rail transit floor cleaning and maintenance equipment, wherein the cleaning and maintenance equipment includes a cleaning brush and a motor; for example... Figure 1 As shown, the method includes: S101, Real-time acquisition of load information of the motor driving the cleaning brush; the load information includes current information and torque information; S102, perform frequency analysis on the load information to identify load fluctuation characteristics related to the rotation frequency of the cleaning brush. S103, based on the amplitude and phase changes of the load fluctuation characteristics, it is determined that the bristles of the cleaning brush are in a non-uniform wear state. S104, based on the fact that the brush bristles are in a non-uniform wear state, adaptively adjust the closed-loop control parameters of the motor and / or superimpose a compensation signal synchronized with the rotation angle of the cleaning brush disc into the control command of the motor, so as to counteract the periodic load fluctuations caused by the non-uniform wear of the brush bristles. S105, according to the closed-loop control parameters and / or the control command of the motor, control the motor to ensure that the cleaning brush has a uniform cleaning force on the ground.

[0021] This application, through real-time monitoring of motor load information and frequency analysis, can accurately identify periodic load fluctuations caused by uneven wear of the brush bristles. By adaptively adjusting the closed-loop control parameters of the motor or superimposing a synchronous compensation signal, this application can effectively counteract the negative impact of uneven wear, thereby ensuring uniform cleaning force of the cleaning brush on the floor, significantly improving cleaning effect, reducing manual intervention, and extending the service life of the equipment.

[0022] To better understand the technical solutions proposed in this application, some key terms are explained first. Cleaning and maintenance equipment typically refers to automated equipment used for cleaning the floors of urban rail transit systems. Its core components include cleaning brushes and the motors that drive them. The cleaning brushes are the components that directly contact the ground for cleaning, and the wear state of their bristles directly affects the cleaning effect. The motor is the power source that drives the rotation of the cleaning brushes, and the stability and accuracy of its operation are crucial to the cleaning process. Load information refers to the information about the external resistance experienced by the motor during operation, typically including current and torque information. Current information reflects the electrical energy consumed by the motor, while torque information directly reflects the magnitude of the torque output by the motor; both can be used to assess the load on the motor. Frequency analysis is a signal processing technique that converts time-domain signals to the frequency domain to identify periodic components in the signal and their corresponding frequencies, amplitudes, and phases. Closed-loop control parameters refer to the parameters used to adjust the controller output in a closed-loop control system to achieve the desired system state, such as the proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd in a PID controller. A compensation signal is a signal added to control commands to counteract the effects of internal or external disturbances on system performance.

[0023] The core of the control method for urban rail transit floor cleaning and maintenance equipment proposed in this application lies in addressing the problem of non-uniform wear of the cleaning brush bristles through intelligent means. First, the method includes real-time acquisition of load information from the motor driving the cleaning brush. This load information can include current and torque information. For example, current information can be acquired in real-time by connecting a current sensor in series in the motor circuit, or torque information can be acquired in real-time by installing a torque sensor on the motor output shaft. These sensors continuously monitor the motor's operating status and transmit the acquired data to the control system. As a preferred embodiment, current information can also be directly acquired through the current sampling circuit inside the motor driver, and torque information can be estimated using a motor model, thus avoiding the need for additional sensors. Second, frequency analysis is performed on the load information to identify load fluctuation characteristics related to the rotation frequency of the cleaning brush. For example, a Fast Fourier Transform (FFT) algorithm can be used to perform spectral analysis on the real-time acquired current or torque information. By analyzing the spectrum, frequency components corresponding to the rotation frequency of the cleaning brush (and its harmonic frequencies) can be identified. The amplitude and phase changes of these specific frequency components constitute the load fluctuation characteristics. For example, when the cleaning brush rotates at a fixed speed, if the bristles are unevenly worn, the effective cleaning area and friction of the bristles in contact with the ground will change periodically during one rotation of the brush, resulting in the motor load exhibiting periodic fluctuations synchronized with the rotation frequency of the brush.

[0024] Furthermore, based on the amplitude and phase changes of the load fluctuation characteristics, it is determined that the bristles of the cleaning brush are in a non-uniform wear state. For example, after initial equipment debugging or replacement of the brush, a reference load frequency spectrum of the motor under normal, uniform load can be recorded, and the reference amplitude and reference phase of specific frequency components can be stored. During daily equipment operation, the current amplitude and current phase of the current specific frequency component are calculated in real time. When it is detected that the increase in the current amplitude relative to the reference amplitude continuously exceeds a certain preset threshold, and the offset of the current phase relative to the reference phase continuously exceeds another preset threshold, it can be determined that the bristles of the cleaning brush are in a non-uniform wear state. This method can effectively distinguish between periodic fluctuations and occasional load changes caused by bristle wear.

[0025] Finally, based on the non-uniform wear state of the brush bristles, the closed-loop control parameters of the motor are adaptively adjusted and / or a compensation signal synchronized with the rotation angle of the cleaning brush disc is superimposed on the control commands of the motor to counteract the periodic load fluctuations caused by the non-uniform wear of the brush bristles. For example, if the motor uses a PID closed-loop controller, when it is determined that the brush bristles are in a non-uniform wear state, at least one of the proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd of the PID controller can be dynamically adjusted according to the severity and location of the wear. Another approach is to obtain the real-time rotation angle of the cleaning brush disc and generate a compensation signal synchronized with that rotation angle. The amplitude and phase of this compensation signal can be calculated based on the load fluctuation characteristics and then superimposed on the control commands of the motor. For example, when brush bristle wear causes a decrease in cleaning force in a certain angle region, the compensation signal can increase the output torque of the motor in that angle region, thereby counteracting the effects of wear.

[0026] Therefore, based on the closed-loop control parameters and / or the motor control commands, the motor is controlled to ensure uniform cleaning force from the cleaning brush disc on the floor. Through the aforementioned adaptive adjustment or compensation mechanism, the motor can respond more precisely to load changes, maintaining a stable rotation speed and uniform cleaning force even under conditions of uneven brush bristle wear. For example, when brush bristle wear leads to insufficient cleaning force in certain areas, the system can promptly adjust the motor output to increase the cleaning force in that area; conversely, when brush bristle wear leads to excessive friction in certain areas, the system can appropriately reduce the motor output to avoid over-cleaning or causing additional wear to the brush disc.

[0027] The overall working principle of this application lies in the real-time and refined monitoring and analysis of the load information of the motor driving the cleaning brush disc, thereby identifying periodic load fluctuations caused by uneven wear of the brush bristles. Traditional methods often struggle to distinguish whether such periodic fluctuations originate from bristle wear or other external factors, making targeted optimization impossible. This application, through frequency analysis, can accurately extract specific frequency components related to the rotation frequency of the cleaning brush disc from complex load signals. The amplitude and phase changes of these components directly reflect the wear state of the bristles. Once it is determined that the bristles are in a state of uneven wear, the system no longer relies on fixed control parameters but intelligently adjusts the closed-loop control parameters of the motor according to the specific wear conditions, or superimposes a compensation signal synchronized with the rotation angle of the cleaning brush disc onto the motor control command. This adaptive adjustment or compensation mechanism can proactively counteract the periodic load fluctuations caused by uneven bristle wear, thereby ensuring that the motor can always provide stable and uniform cleaning force throughout the cleaning process. In this way, this application effectively solves the problems of uneven cleaning force, reduced cleaning effect, and the need for frequent manual intervention caused by uneven wear of brush bristles in the prior art, and realizes the intelligent and efficient operation of cleaning and maintenance equipment.

[0028] Compared to existing technologies, the core innovation of this application lies in its ability to identify the non-uniform wear state of the cleaning brush bristles in real time and make adaptive control adjustments accordingly. In existing technologies, when the bristles experience non-uniform wear, the periodic fluctuations in the motor load cause the motor speed to decrease slightly in certain angular areas, while it may return to normal or slightly overshoot in other areas, leaving "stripes" or "patches" on the ground indicating incomplete cleaning. This application, by acquiring the motor's current and torque information in real time and performing frequency analysis, can accurately capture the load fluctuation characteristics related to the rotation frequency of the cleaning brush. By analyzing the amplitude and phase changes of these characteristics, the system can accurately determine whether the bristles are in a non-uniform wear state. Once non-uniform wear is identified, this application can adaptively adjust the motor's closed-loop control parameters, such as the proportional, integral, and derivative coefficients of the PID controller, or superimpose a compensation signal synchronized with the rotation angle of the cleaning brush onto the motor control command. This proactive, real-time adjustment mechanism enables the motor to counteract periodic load fluctuations caused by uneven wear of the brush bristles, thereby ensuring that the cleaning force of the cleaning brush disc on the floor remains uniform at all times.

[0029] Furthermore, this application proposes the following steps for performing frequency analysis on load information to identify load fluctuation characteristics related to the rotation frequency of the cleaning brush: performing real-time frequency analysis on the collected load information and extracting specific frequency components related to the rotation frequency of the cleaning brush as the load fluctuation characteristics.

[0030] Specifically, the real-time frequency analysis can be understood as performing instantaneous or quasi-instantaneous spectral decomposition on continuously or periodically acquired load information (including current and torque information). For example, signal processing techniques such as Fast Fourier Transform (FFT), Short-Time Fourier Transform (STFT), or wavelet transform can be used to convert the time-domain load data into a frequency-domain spectrum. The purpose is to reveal the various frequency components contained in the load information and their corresponding amplitude and phase information. Extracting specific frequency components related to the rotation frequency of the cleaning brush refers to identifying and separating the frequency components corresponding to the rotation frequency and harmonic frequencies of the cleaning brush in the spectrum obtained from the real-time frequency analysis. For example, if the rotation frequency of the cleaning brush is f_brush, then load fluctuations at frequencies f_brush, 2f_brush, 3f_brush, etc., need to be considered. The amplitude and phase changes of these specific frequency components can directly reflect the periodic load changes experienced by the cleaning brush during rotation. The aim is to distinguish periodic load fluctuations associated with uneven wear of the cleaning brush from other noise or non-periodic load changes, thereby improving the accuracy of identification.

[0031] The solution presented in this application continuously monitors the spectral characteristics of the motor load by performing real-time frequency analysis on the collected load information. It is precisely this real-time analysis that enables the system to promptly capture dynamic changes in load fluctuations. Furthermore, by specifically extracting specific frequency components related to the rotation frequency of the cleaning brush, the system can effectively focus on periodic load changes caused by the characteristics of the cleaning brush itself (such as bristle wear), thereby filtering out other irrelevant interference signals. This focused analysis allows the system to more accurately identify load fluctuation characteristics directly related to uneven bristle wear, providing a reliable data foundation for subsequent wear condition assessment.

[0032] This application further proposes a specific method for determining that the bristles of the cleaning brush are in a non-uniform wear state based on the amplitude and phase changes of load fluctuation characteristics, which includes: After initial equipment debugging or replacement of the brush disk, a reference load frequency spectrum of the motor is established, and the amplitude and phase of the specific frequency components are stored as reference amplitude and reference phase. During equipment operation, the current amplitude and current phase of a specific frequency component are calculated in real time. When the increase in the current amplitude relative to the reference amplitude continuously exceeds a first preset threshold, and the offset of the current phase relative to the reference phase continuously exceeds a second preset threshold, the non-uniform wear state of the cleaning brush is determined.

[0033] Specifically, in this embodiment, establishing a reference load frequency spectrum for the motor refers to a comprehensive collection of load information and frequency analysis conducted before the urban rail transit floor cleaning and maintenance equipment is put into use, or after the cleaning brushes are replaced with brand new ones, while the brushes are in a normal, uniform wear state. During this process, the load information of the motor driving the cleaning brushes, such as current and torque information, is collected in real time. Subsequently, frequency analysis is performed on this load information to extract specific frequency components related to the rotation frequency of the cleaning brushes. The amplitude and phase of these specific frequency components are considered as reference data for the brushes under ideal or normal working conditions, and are therefore stored as reference amplitude and reference phase. The establishment of this reference data provides a reliable reference point for subsequent judgment of non-uniform wear of the brush bristles.

[0034] During equipment operation, the system continuously monitors the motor's load information and performs real-time frequency analysis by calculating the current amplitude and phase of specific frequency components. This allows for the dynamic acquisition of the current amplitude and phase of specific frequency components related to the cleaning brush's rotation frequency. This real-time data reflects the actual load fluctuations of the brush under its current operating conditions.

[0035] When the increase in current amplitude relative to a reference amplitude continuously exceeds a first preset threshold, and the offset of current phase relative to a reference phase continuously exceeds a second preset threshold, the system determines that the bristles of the cleaning brush are in a state of non-uniform wear. The first and second preset thresholds are parameters pre-set based on experience, experiments, or simulations, used to define the degree of amplitude increase and phase offset sufficient to indicate non-uniform wear of the bristles. For example, the first preset threshold can be set as a certain percentage of the reference amplitude, and the second preset threshold can be set as a certain angle value. The phrase "continuously exceeding" aims to avoid misjudgments caused by instantaneous interference or accidental fluctuations, ensuring the robustness of the judgment.

[0036] This application's solution effectively addresses the potential ambiguity and inaccuracy issues in identifying uneven brush wear by introducing a reference load frequency spectrum and a threshold-based continuous judgment mechanism. Specifically, the reference amplitude and phase established after initial equipment debugging or replacement of the brush disc provide a clear reference point for the load fluctuation characteristics of the cleaning brush disc under normal operating conditions. When uneven wear occurs in the brush bristles, the force they exert on the ground becomes uneven, causing the motor load to exhibit periodic fluctuations different from the normal state during brush disc rotation. These fluctuations are reflected in the increase in amplitude and phase shift of specific frequency components. By calculating the current amplitude and phase in real time and comparing them with preset reference values, this change can be quantified. When these changes (i.e., the increase and shift) continuously exceed preset thresholds, it can be reliably determined that the brush bristles are in an uneven wear state. This mechanism based on quantitative indicators and continuous judgment significantly improves the accuracy and robustness of uneven wear state identification, avoiding misjudgments caused by subjective judgment or instantaneous interference.

[0037] In some preferred embodiments, a specific example is given below. Assume that an urban rail transit floor cleaning and maintenance device undergoes initial commissioning before being put into use. During this process, brand-new cleaning brushes are installed, and the motor is driven to run at a standard speed. The system collects the motor's current and torque information in real time and performs frequency analysis to extract the amplitude and phase of specific frequency components related to the rotation frequency of the cleaning brushes. For example, at a specific frequency f1, a reference amplitude of A_base and a reference phase of P_base are measured. These values ​​are stored as reference data.

[0038] During routine operation, the system continuously collects motor load information in real time and calculates the current amplitude A_current and current phase P_current at a specific frequency f1. For example, after a period of use, the system detects that the increase in A_current relative to A_base continuously exceeds a preset first threshold (e.g., 15% of A_base), and the offset of P_current relative to P_base continuously exceeds a preset second threshold (e.g., 10 degrees). When these conditions are met continuously for a certain period (e.g., 10 consecutive sampling cycles), the system determines that the bristles of the cleaning brush are in a state of non-uniform wear. Based on this judgment, the control system can immediately trigger subsequent adaptive adjustment mechanisms, such as adjusting the closed-loop control parameters of the motor or superimposing compensation signals, to counteract the effects of non-uniform wear and ensure uniform cleaning force.

[0039] Furthermore, this application also proposes an optimization scheme, wherein the motor adopts a PID closed-loop controller; the method further includes: determining the severity and location of the non-uniform wear while determining that the bristles of the cleaning brush are in a non-uniform wear state; the adaptive adjustment of the closed-loop control parameters of the motor includes: dynamically adjusting at least one of the proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd of the PID closed-loop controller according to the severity and location of the non-uniform wear and a preset parameter mapping table.

[0040] Specifically, the PID closed-loop controller is a feedback controller widely used in industrial control. It adjusts the control output by calculating the weighted sum of proportional, integral, and derivative components to make the system output as close as possible to the setpoint. In this application, the PID controller is used to precisely control the motor driving the cleaning brush to maintain stable speed and torque, thereby ensuring uniform cleaning force from the cleaning brush on the floor. The severity and location of non-uniform wear can be determined based on further analysis of the amplitude and phase changes of load fluctuation characteristics. For example, by comparing the current load fluctuation characteristics with a reference spectrum, the degree of amplitude deviation can be quantified as an indicator of wear severity, while the phase shift can indicate the specific rotational direction of wear occurrence. For instance, the rotation cycle of the cleaning brush can be divided into multiple angular intervals, and the load fluctuation characteristics within each interval can be analyzed to accurately locate the direction of non-uniform wear.

[0041] In practical applications, the preset parameter mapping table is a pre-established database or lookup table that stores optimized combinations of PID control parameters (Kp, Ki, Kd) corresponding to different degrees and locations of non-uniform wear. Once the system determines that the brush bristles are in a non-uniform wear state and quantifies its severity and location, it can query this mapping table to obtain the most suitable PID parameters for the current wear condition and dynamically adjust at least one of the proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd of the PID closed-loop controller. The proportional coefficient Kp mainly affects the system's response speed and steady-state error; the integral coefficient Ki is used to eliminate steady-state error and improve system accuracy; and the derivative coefficient Kd is used to suppress oscillations and improve system stability. By dynamically adjusting these parameters, the PID controller can better adapt to the periodic load changes caused by non-uniform brush bristle wear.

[0042] This application's solution addresses the potential for insufficient compensation accuracy in basic solutions by simultaneously determining the non-uniform wear state of the bristles and further identifying the severity and location of this non-uniform wear, combined with dynamic parameter adjustment using a PID closed-loop controller. Specifically, when the bristles of the cleaning brush experience non-uniform wear, the pressure they exert on the ground changes periodically during rotation, causing periodic fluctuations in the drive motor's load information related to the brush's rotation frequency. By accurately identifying the amplitude and phase of these load fluctuations, the severity and specific location of the non-uniform wear can be inferred. For example, a larger amplitude may indicate more severe wear, while a specific phase corresponds to the angular region where wear occurs.

[0043] Based on this, a PID closed-loop controller is used to control the motor, and the proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd of the PID controller are dynamically adjusted according to the determined severity and location of non-uniform wear. This dynamic adjustment allows the controller to specifically compensate for non-uniform wear of different degrees and locations. For example, if wear is more severe in a certain angular area, resulting in a large load fluctuation in that area, the PID parameters can be adjusted to apply a stronger compensating torque in that angular area to offset the additional load, thereby ensuring that the cleaning force of the cleaning brush on the floor remains uniform throughout the entire rotation cycle. The preset parameter mapping table provides a scientific basis and an efficient implementation method for this dynamic adjustment, ensuring the accuracy and real-time performance of parameter adjustment.

[0044] In some preferred embodiments, a specific example is given below. Assume that during the operation of an urban rail transit floor cleaning and maintenance equipment, the system collects real-time load information of the motor driving the cleaning brush, and after frequency analysis, identifies load fluctuation characteristics related to the rotation frequency of the cleaning brush. Further, by analyzing the amplitude and phase changes of these load fluctuation characteristics, the system determines that the brush bristles of the cleaning brush are in a state of non-uniform wear. For example, if the system detects that the amplitude of the load fluctuation is significantly higher than the reference value and the phase continuously deviates within the range of 90 to 180 degrees of brush rotation angle, this indicates that the brush bristles experience severe non-uniform wear in this specific area.

[0045] At this point, the control system queries a preset parameter mapping table based on the severity (e.g., the increase in amplitude exceeds a certain threshold) and orientation (e.g., 90 to 180 degrees) of the detected non-uniform wear. This mapping table may predefine PID parameter adjustment strategies for cases of "severe wear in a specific angular area." For example, the mapping table might indicate that in this case, the proportional coefficient Kp needs to be appropriately increased to improve response speed, the integral coefficient Ki fine-tuned to eliminate steady-state error, and the derivative coefficient Kd slightly adjusted to maintain system stability. The system then dynamically adjusts the Kp, Ki, and Kd parameters of the PID closed-loop controller. Through this targeted parameter adjustment, the PID controller can apply a more precise compensating torque within the wear area, effectively counteracting the periodic load fluctuations caused by non-uniform brush wear, thereby ensuring that the cleaning force of the cleaning brush on the floor remains uniform throughout the cleaning process, avoiding cleaning blind spots or over-cleaning caused by localized wear.

[0046] Furthermore, this application proposes an optimization scheme that, based on the non-uniform wear state of the brush bristles, adaptively adjusts the closed-loop control parameters of the motor and / or superimposes a compensation signal synchronized with the rotation angle of the cleaning brush disc onto the motor control command, specifically including: Collect real-time load data of the motor and location information of the cleaning and maintenance equipment, and perform frequency analysis, transient characteristic analysis and spectral entropy analysis on the real-time load data to obtain load fluctuation characteristics; Determine whether the load fluctuation is caused by uneven wear of the brush bristles or localized contamination of the ground. If the load fluctuation is caused by uneven wear of the brush bristles, adjust the closed-loop control parameters of the drive motor, or obtain the rotation angle of the cleaning brush and generate a compensation signal synchronized with the rotation angle, and superimpose the compensation signal into the motor control command. If the load fluctuation is caused by localized contamination of the ground, adjust the enhanced cleaning strategy.

[0047] Specifically, in this embodiment, real-time load data of the motor and location information of the cleaning and maintenance equipment are collected to provide comprehensive input for subsequent load fluctuation characteristic analysis. Real-time load data may include current and torque information, reflecting the actual load on the motor during operation. The location information of the cleaning and maintenance equipment helps correlate load fluctuations with specific ground areas or the rotation angle of the cleaning brush. Frequency analysis of the real-time load data reveals the frequency components, amplitude, and phase of periodic fluctuations, which is particularly important for identifying features related to the rotation frequency of the cleaning brush. Transient characteristic analysis focuses on capturing non-periodic, sudden changes in the load data, such as the impact response when the cleaning brush suddenly encounters an obstacle or localized heavy contamination. Spectral entropy analysis quantifies the complexity and randomness of the load signal, providing additional criteria for distinguishing load fluctuations from different sources. By comprehensively utilizing these three analysis methods, load fluctuation characteristics can be obtained more comprehensively and accurately, providing a rich data foundation for subsequent judgments.

[0048] One of the core innovations of this solution is determining whether load fluctuations originate from uneven brush bristle wear or localized ground contamination. Uneven brush bristle wear typically leads to periodic load fluctuations highly correlated with the rotation frequency of the cleaning brush disc, with amplitude and phase changes exhibiting certain regularity. Localized ground contamination, such as sticky stains or obstacles, can cause transient, non-periodic load shocks or localized, sustained load increases, with potentially more complex frequency characteristics or higher spectral entropy. By comprehensively interpreting different analytical results, these two main sources of interference can be effectively distinguished.

[0049] In practical applications, if the load fluctuation is determined to originate from uneven wear of the brush bristles, a strategy similar to the above method is adopted. This involves adjusting the closed-loop control parameters of the drive motor, such as the proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd of the PID controller, to compensate for the torque imbalance caused by bristle wear. Alternatively, the rotation angle of the cleaning brush disc is obtained, and a compensation signal synchronized with this rotation angle is generated. This compensation signal is then superimposed on the motor's control command to actively counteract periodic load fluctuations and ensure uniform cleaning force from the cleaning brush disc on the floor. If the load fluctuation is determined to originate from localized contamination on the floor, the adjustment is no longer to the balance of the cleaning brush disc, but rather to adjust the enhanced cleaning strategy. Enhanced cleaning strategies may include, but are not limited to: increasing the downward pressure of the cleaning brush disc, increasing the rotation speed of the cleaning brush disc, increasing the amount of cleaning agent sprayed, or having the equipment perform multiple reciprocating cleaning cycles in the contaminated area to effectively remove stubborn stains and improve the overall cleaning effect.

[0050] This application's solution overcomes the limitations of traditional methods in distinguishing between non-uniform brush wear and localized ground contamination by introducing multi-dimensional load data analysis (frequency analysis, transient feature analysis, and spectral entropy analysis) and an intelligent judgment mechanism for the source of load fluctuations. Specifically, when the motor load of the cleaning brush fluctuates, the system no longer simply attributes it to brush wear. Instead, it constructs a more refined load fluctuation model by comprehensively extracting and analyzing features from real-time load data. For example, periodic fluctuations highly correlated with rotation frequency are more likely to indicate brush wear, while sudden, high-amplitude transient impacts or a significant increase in local spectral entropy may indicate localized ground contamination. Thus, the system can accurately identify the physical source of load fluctuations. Once the specific source of the load fluctuation is determined, the system can adopt targeted control strategies: for non-uniform brush wear, the cleaning force uniformity of the cleaning brush is restored by adjusting the motor closed-loop control parameters or superimposing compensation signals; for localized ground contamination, the stains are effectively removed by adjusting the enhanced cleaning strategy. This differentiated response mechanism ensures that the equipment maintains efficient and accurate cleaning performance in the face of complex and changing operating environments.

[0051] In some preferred embodiments, it is assumed that the urban rail transit floor cleaning and maintenance equipment is performing routine cleaning operations. At a certain moment, the load data of the motor driving the cleaning brush fluctuates. The system first collects real-time load data such as motor current and torque, as well as the current position information of the equipment. Subsequently, frequency analysis is performed on these real-time load data, revealing a periodic fluctuation related to the rotation frequency of the cleaning brush, but its amplitude and phase changes do not entirely conform to the typical non-uniform wear pattern of the bristles. Simultaneously, transient feature analysis detects a brief, high-amplitude load impact at a specific location, and spectral entropy analysis shows a significant increase in the complexity of the load signal in this area.

[0052] Based on these comprehensive analysis results, the system determines that the load fluctuation is not simply due to uneven brush bristle wear, but rather to a localized area of ​​stubborn, sticky dirt or a small obstruction on the ground. In this case, instead of incorrectly adjusting the motor's closed-loop control parameters to compensate for brush bristle wear, the system immediately adjusts its enhanced cleaning strategy. Specifically, the control system instructs the cleaning brush to increase downward pressure at its current position, while slightly reducing the equipment's travel speed and increasing the amount of cleaning agent sprayed. It may even perform a short, reciprocating cleaning motion in that localized area. Through this targeted enhanced cleaning, stubborn dirt is effectively removed, the load fluctuation returns to normal, and the brush wear compensation mechanism remains unchanged, avoiding unnecessary adjustments and thus ensuring cleaning effectiveness and equipment operational stability.

[0053] Furthermore, this application also proposes a method for determining whether the aforementioned load fluctuation characteristics originate from uneven wear of the brush bristles or localized ground contamination, which specifically includes: While collecting the real-time load data of the motor, the rotation angle of the cleaning brush disc is also collected. A preliminary analysis of the real-time load data is performed to identify periodic load fluctuations; A perturbation signal is applied to the motor, and the transient response data of the motor to the perturbation signal is collected; The transient response data is analyzed to extract response features, and the physical source of the periodic load fluctuation is determined based on the response features; the physical source includes non-uniform wear of brush bristles and localized ground contamination.

[0054] Specifically, in this embodiment, while acquiring real-time load data of the motor, the rotation angle of the cleaning brush disc is also acquired. This aims to ensure that in subsequent analysis, load fluctuations can be correlated with specific rotational positions of the cleaning brush disc, providing spatial information to distinguish fluctuations from different sources. Preliminary analysis of the real-time load data is performed to identify periodic load fluctuations. Frequency domain analysis methods such as Fourier transform, or time-frequency analysis methods such as wavelet analysis, can be used to identify periodic components related to the rotational frequency and harmonics of the cleaning brush disc. The purpose is to determine whether there are periodic load fluctuations requiring further diagnosis. The perturbation signal is a transient, controllable external excitation, such as a short-duration current pulse or voltage step. Applying the perturbation signal to the motor and acquiring the motor's transient response data to the perturbation signal allows for the observation of the motor's response to this known excitation, thus probing the dynamic characteristics of the motor-brush disc-ground system. The transient response data can include motor current, torque, speed, or brush disc vibration, aiming to obtain the dynamic behavior of the system under specific excitations to reveal its intrinsic physical characteristics. Analyzing transient response data and extracting response features, such as response time, peak overshoot, decay rate, and oscillation frequency, quantifies the system's dynamic response to perturbation signals. Based on these response features, the physical source of periodic load fluctuations can be determined. By comparing the differences in response characteristics under perturbation signals from different physical sources (uniform brush wear or localized ground contamination), discrimination rules can be established. For example, non-uniform brush wear may cause localized changes in the system's response characteristics at a specific rotation angle, while localized ground contamination may cause the system to exhibit different transient responses upon contact with that area.

[0055] This application's solution effectively distinguishes the physical sources of load fluctuations by introducing perturbation signals and analyzing their transient responses. Specifically, when the bristles of the cleaning brush experience non-uniform wear, the contact force between the brush and the ground changes periodically during rotation. This change affects the dynamic response characteristics of the motor at specific rotation angles. When a perturbation signal is applied to the motor, the contact stiffness or damping characteristics of the non-uniformly worn bristles may differ from those of the normal area, resulting in differences in the motor's transient response to the perturbation signal at different rotation angles. For example, in the worn area, the motor response may decay faster or slower, and the peak overshoot may be larger or smaller. Conversely, when load fluctuations originate from localized ground contamination, such as the presence of sticky substances or unevenness, this external disturbance directly affects the contact interface between the brush and the ground, causing instantaneous changes in the motor load. However, this ground contamination typically does not alter the physical properties of the brush itself. Therefore, when a perturbation signal is applied, the motor's transient response to the perturbation signal may remain relatively consistent throughout the entire rotation cycle, or it may exhibit a different response pattern in the contaminated area compared to the bristle wear. By performing detailed analysis of these transient response characteristics and combining them with the rotation angle information of the cleaning brush, a unique "fingerprint" of uneven bristle wear and localized ground contamination can be established, thereby enabling accurate judgment of the source of load fluctuations.

[0056] In some preferred embodiments, specifically, during the operation of the urban rail transit floor cleaning and maintenance equipment, the control system first collects load data such as the motor current and torque driving the cleaning brush in real time, and simultaneously acquires the rotation angle of the cleaning brush. Then, a preliminary Fourier transform analysis is performed on the collected real-time load data to identify whether there are periodic load fluctuations related to the brush rotation frequency. If such fluctuations are detected, the system applies a preset high-frequency narrow-pulse current signal to the motor as a perturbation signal. Simultaneously with applying the perturbation signal, the system rapidly collects the motor current, torque, and brush vibration data as transient response data. Next, these transient response data are analyzed to extract response characteristics such as response time, peak overshoot, decay rate, and oscillation frequency. For example, if it is found that within a specific rotation angle range of the cleaning brush, the motor's response time to the perturbation signal is significantly shortened and the peak overshoot decreases, this may indicate that brush wear in that area has led to a change in contact stiffness. Conversely, if in a certain area, the response time is prolonged and the decay rate is slowed, it may indicate the presence of sticky contamination on the ground in that area. By correlating these response characteristics with the rotation angle of the cleaning brush and comparing them with pre-established characteristic models of brush wear and ground contamination, the system can accurately determine the physical source of the current periodic load fluctuations, thus providing a basis for subsequent precise control.

[0057] In some embodiments described above in this application, a method is proposed to apply a perturbation signal to a motor and collect transient response data of the motor to the perturbation signal in order to determine the physical source of periodic load fluctuations. However, in practical applications, how to select a suitable type of perturbation signal, ensure that it effectively excites the system response in complex noise environments, and accurately and comprehensively collect transient response data are key factors affecting the accuracy and reliability of subsequent judgments.

[0058] In response, this application further proposes that the aforementioned perturbation signal is a high-frequency narrow-pulse current signal; the frequency of the high-frequency narrow-pulse current signal is higher than the highest harmonic of the rotation frequency of the cleaning brush; the perturbation signal is applied to the motor, and the transient response data of the motor to the perturbation signal is collected, including: acquiring the current ambient noise and the brush rotation speed of the cleaning brush; adjusting the amplitude and duration of the perturbation signal according to the ambient noise and the brush rotation speed to ensure that the energy of the perturbation signal is higher than the ambient noise and the duration of the perturbation signal is less than the time required for the cleaning brush to rotate one angle interval; and simultaneously collecting motor current, torque, and brush vibration data as the transient response data.

[0059] Specifically, the perturbation signal is designed as a high-frequency, narrow-pulse current signal. "High-frequency" means its frequency is much higher than the normal rotation frequency of the cleaning brush and its potential harmonic frequencies. For example, it can be set to several times or even tens of times the highest harmonic of the cleaning brush's rotation frequency. The purpose is to separate the spectrum of the perturbation signal from the load fluctuation spectrum of the cleaning brush during normal operation, avoiding signal aliasing. This ensures that the perturbation signal can independently excite the transient response of the motor and brush, unaffected by normal load fluctuations. "Narrow pulse" means the signal has an extremely short duration. Its purpose is to quickly and instantaneously inject energy into the motor to obtain a clear transient response without significantly interfering with the normal operation of the cleaning and maintenance equipment.

[0060] When applying the perturbation signal and acquiring transient response data, it is first necessary to obtain the current ambient noise and the rotational speed of the cleaning brush. Ambient noise can be understood as various random or non-random interference signals present in the equipment's operating environment, such as noise from the motor itself, the transmission mechanism, ground friction, and the external environment. Its purpose is to provide a reference for adjusting the parameters of the perturbation signal. The brush rotational speed is the actual current rotational speed of the cleaning brush, and its purpose is to provide a basis for calculating the time required for the cleaning brush to rotate within a certain angle range.

[0061] Furthermore, based on the acquired ambient noise and brush rotation speed, the amplitude and duration of the perturbation signal are adaptively adjusted. Specifically, the amplitude of the perturbation signal is adjusted to ensure its energy is higher than the current ambient noise. This aims to ensure that the perturbation signal can still be clearly identified and effectively stimulate the system response even in a noisy environment, preventing the response signal from being submerged by noise. The duration of the perturbation signal is adjusted to ensure it is less than the time required for the cleaning brush to rotate within a certain angle range. This aims to ensure that the stimulation and response process of the perturbation signal is completed within a local area of ​​the brush, avoiding the averaging or blurring of response characteristics due to brush rotation, thereby more accurately reflecting local characteristics.

[0062] In addition, motor current, torque, and brush disc vibration data are simultaneously acquired as transient response data. Motor current and torque directly reflect the electrical and mechanical dynamic response of the motor under the influence of perturbation signals, aiming to capture the transient behavior of the system from the perspective of the drive source. Brush disc vibration data directly reflects the mechanical vibration characteristics of the cleaning brush disc in relation to the ground or the brush bristles themselves under the influence of perturbation, aiming to provide direct evidence of physical contact and mechanical state from the perspective of the actuator. The simultaneous acquisition of these multimodal data aims to provide a comprehensive and complementary transient response view, providing richer and more reliable information for subsequent response characteristic analysis and physical source determination.

[0063] This application's solution effectively solves the problems of insufficient perturbation signal effectiveness and inaccurate response data acquisition under complex working conditions by introducing a high-frequency narrow-pulse current signal as a perturbation signal and combining it with environmental noise and brush rotation speed for adaptive parameter adjustment. Specifically, the characteristics of the high-frequency narrow-pulse current signal separate it from the normal operating frequency and harmonics of the cleaning brush in the frequency spectrum, thereby avoiding signal interference and ensuring that the excited transient response is purely caused by the perturbation signal. At the same time, its narrow-pulse characteristic ensures instantaneous and localized impact on the cleaning process, avoiding long-term interference with normal cleaning operations.

[0064] By acquiring the current ambient noise and adjusting the amplitude of the perturbation signal accordingly, this application ensures that the energy of the perturbation signal is always higher than the ambient noise, enabling the system's response to the perturbation to be clearly separated from the noise, thus greatly improving the signal-to-noise ratio and reliability of the transient response data. Furthermore, by adjusting the duration of the perturbation signal according to the brush rotation speed to be less than the time required for the cleaning brush to rotate within a certain angle range, this strategy ensures that the transient response acquisition is completed within a local area of ​​the brush, avoiding the averaging effect of continuous brush rotation on the transient response characteristics. This allows for more accurate capture of transient features related to local physical states (such as uneven bristle wear or localized ground contamination). Simultaneous acquisition of motor current, torque, and brush vibration data provides multi-dimensional and complementary transient response information. Motor current and torque data reflect the changes in electrical and mechanical load on the motor under instantaneous excitation, while brush vibration data directly reveals the mechanical dynamic behavior of the brush in contact with the ground or the bristles themselves. This comprehensive analysis of multimodal data makes the judgment of the physical source of periodic load fluctuations more comprehensive and accurate, because different physical sources (non-uniform wear of brush bristles or local contamination of the ground) will leave unique "fingerprints" in these different types of transient response data.

[0065] Furthermore, by adaptively adjusting the amplitude and duration of the perturbation signal based on environmental noise and brush rotation speed, the system ensures that the perturbation signal can effectively excite the system response under various complex operating conditions, and that the response signal is not drowned out by environmental noise. This also avoids the blurring of transient response characteristics by brush rotation, thereby significantly improving the quality of transient response data and the reliability of diagnostics. Simultaneous acquisition of multimodal data such as motor current, torque, and brush vibration provides more comprehensive and richer system dynamic information, enabling subsequent response characteristic analysis to be cross-validated from multiple dimensions. This allows for a more accurate distinction between the two different physical sources: non-uniform brush wear and localized ground contamination, providing a solid data foundation for subsequent adaptive control or enhanced cleaning strategies.

[0066] In some preferred embodiments, this application is implemented as follows. Assume that urban rail transit floor cleaning and maintenance equipment is cleaning a subway platform. To determine whether periodic load fluctuations during equipment operation are due to uneven wear of the cleaning brush bristles or localized ground contamination, the control system triggers the application of a perturbation signal. Specifically, firstly, the system acquires the current ambient noise level through a built-in microphone or vibration sensor, for example, detecting high noise from nearby trains entering the station. Simultaneously, the current rotational speed of the cleaning brush is acquired in real time via a motor encoder, for example, 120 RPM. Based on this information, the control unit dynamically calculates and generates a high-frequency narrow-pulse current signal. For example, if the highest harmonic of the brush rotation frequency is 20 Hz, the frequency of the perturbation signal might be set to 100 Hz. The amplitude of the perturbation signal is adjusted according to the current ambient noise level to ensure its energy is sufficient to produce a detectable transient response in a noisy environment. Simultaneously, based on the brush rotation speed of 120 RPM, the time required for the brush to rotate one angular interval (e.g., 10 degrees) is calculated, and the duration of the perturbation signal is set to be much shorter than this time, for example, 5 milliseconds. When this parameter-tuned high-frequency narrow-pulse current signal is applied to the motor driving the cleaning brush, high-precision sensors simultaneously collect data on the motor's current, torque, and the cleaning brush's vibration. This data is recorded at high speed, forming a multi-dimensional transient response dataset. This dataset is then fed into an analysis module to extract response characteristics such as response time, peak overshoot, decay rate, and oscillation frequency, thereby determining the specific physical source of periodic load fluctuations. For example, if the transient response exhibits rapidly decaying oscillations in current and torque, and the brush vibration data also shows a specific pattern, it may indicate non-uniform wear of the brush bristles.

[0067] Furthermore, this application also proposes to determine the physical source of the periodic load fluctuation based on the response characteristics, including: acquiring the rotation angle of the cleaning brush and the response characteristics of the transient response data, and associating the response characteristics with the rotation angle of the cleaning brush; and determining the physical source of the load fluctuation based on the distribution pattern of the response characteristics on the rotation angle of the cleaning brush.

[0068] Specifically, in this embodiment, the response characteristics of the aforementioned transient response data refer to parameters extracted from the transient response data of the motor to perturbation signals that characterize the dynamic properties of the system, such as response time, peak overshoot, decay rate, and oscillation frequency. Correlating these response characteristics with the rotation angle of the cleaning brush disc means that while acquiring the transient response data and extracting its response characteristics, the corresponding rotation angle of the cleaning brush disc is recorded, thereby establishing a one-to-one correspondence between the response characteristics and the brush disc rotation angle. This can be achieved through time-synchronized sampling or data mapping.

[0069] Determining the physical source of load fluctuations based on the distribution pattern of the response characteristics at the rotation angle of the cleaning brush disc refers to identifying the source by analyzing the variation pattern and spatial distribution characteristics of the response characteristics within one or more rotations of the cleaning brush disc. For example, if the response characteristics exhibit a periodic, repetitive variation pattern at a specific rotation angle of the cleaning brush disc, and this pattern is synchronized with the rotation cycle of the cleaning brush disc, it usually indicates that the load fluctuations originate from the non-uniformity of the cleaning brush disc itself, such as non-uniform wear of the bristles. This is because the wear state of the bristles is fixed at different angles of the brush disc, and when the worn area contacts the ground, it causes a specific change in the motor response characteristics. Conversely, if the variation pattern of the response characteristics has a weak correlation with the rotation angle of the cleaning brush disc, or exhibits localized, non-periodic changes more related to the position of the equipment on the ground, it may indicate that the load fluctuations originate from localized ground contamination.

[0070] This application's solution, by correlating transient response characteristics with the rotation angle of the cleaning brush and analyzing its distribution pattern at the rotation angle, can more precisely capture the physical characteristics of the load fluctuation source. Non-uniform wear of the brush bristles is an inherent characteristic of the cleaning brush itself, and the resulting load fluctuations repeat at the same angular position in each rotation cycle of the cleaning brush, forming a stable angular distribution pattern. Localized ground contamination, on the other hand, is an external environmental factor, and its impact on the load is not fixedly correlated with the rotation angle of the cleaning brush, but rather with the equipment's position on the ground. Through this angular correlation and pattern analysis, these two different sources of periodic load fluctuations can be effectively distinguished, thus avoiding misjudgment.

[0071] In some preferred embodiments, it is assumed that the bristles of the cleaning brush exhibit non-uniform wear in a specific area. When the cleaning and maintenance equipment is running, the system continuously applies a perturbation signal and collects transient response data from the motor. By correlating the response characteristics (e.g., peak overshoot) of each collected transient response data point with the current rotation angle of the cleaning brush, a curve of peak overshoot versus rotation angle is plotted. If the curve shows a sustained and periodic significant decrease in peak overshoot near a fixed rotation angle (e.g., 180 degrees) of the cleaning brush, and this decrease pattern repeats with each brush rotation, it can be determined that the bristles of the cleaning brush are in a state of non-uniform wear. Conversely, if the change in peak overshoot is random, or its pattern is related to the equipment's position on the ground rather than to the fixed rotation angle of the cleaning brush, it can be determined that the load fluctuation originates from localized ground contamination.

[0072] Furthermore, this application proposes specific judgment criteria to achieve accurate identification of the physical source of load fluctuations by quantifying response characteristics and setting thresholds. In this application, the response characteristics include response time, peak overshoot, decay rate, and oscillation frequency. The transient response data is analyzed to extract response characteristics, and the physical source of the periodic load fluctuations is determined based on these characteristics. Specifically, if the response time is less than a first response time threshold and the peak overshoot is less than a first overshoot threshold, the load fluctuation is determined to originate from uneven wear of the brush bristles; if the response time is greater than a second response time threshold and the decay rate is less than a first decay rate threshold, the load fluctuation is determined to originate from localized ground contamination.

[0073] Specifically, in this embodiment, the response characteristics refer to the dynamic characteristics exhibited by the transient response data of the motor after receiving a perturbation signal. The response time can be understood as the time required for the motor to reach a steady state from receiving the perturbation signal, or the time required to reach a preset response level. Peak overshoot refers to the difference between the maximum instantaneous value and the steady-state value of the motor response, usually expressed as a percentage of the steady-state value, reflecting the degree of overshoot of the system to the input. The decay rate refers to the rate at which the motor response gradually decreases from peak overshoot to a steady state, reflecting the system's damping characteristics. The oscillation frequency refers to the periodic fluctuation frequency that may occur in the motor response during decay. These characteristics collectively depict the dynamic behavior of the motor system when subjected to external disturbances. The first response time threshold, the first overshoot threshold, the second response time threshold, and the first decay rate threshold are all preset parameters, set to distinguish load fluctuations caused by different physical sources. These thresholds can be calibrated and optimized through extensive experimental data, simulation analysis, and expert experience. For example, during the initial equipment commissioning phase, a large amount of transient response data can be collected by simulating two scenarios: non-uniform brush bristle wear and localized ground contamination. The response characteristics can then be analyzed to determine the optimal thresholds that effectively distinguish between these two scenarios. The first response time threshold and the first overshoot threshold are typically set to relatively small values ​​to capture the rapid, slight response changes caused by non-uniform brush bristle wear; while the second response time threshold and the first decay rate threshold are set to relatively large values ​​to identify the slower response and slower decay characteristics caused by localized ground contamination.

[0074] This application's solution achieves precise differentiation of the physical source of periodic load fluctuations by quantitatively analyzing the response characteristics of motor transient response data and combining this with preset thresholds. Specifically, when the bristles of the cleaning brush experience non-uniform wear, the frictional force between the bristles and the ground changes periodically during the brush's rotation. This change typically manifests as a relatively fast and small-amplitude transient response on the motor system. Therefore, if a short response time and small peak overshoot are observed, it can be reasonably inferred that the load fluctuation mainly originates from the non-uniform wear of the bristles. This is because the contact force change caused by bristle wear is a structural and relatively stable periodic disturbance within the system, which the motor control system can respond to and suppress relatively quickly, but will still leave slight transient characteristics.

[0075] Conversely, when there is localized contamination in the cleaning area, such as sticky stains or obstacles, the cleaning brush experiences significant, transient external resistance as it passes through these areas. This external resistance typically manifests as a transient response with a longer response time and a slower decay rate on the motor system. This is because resistance changes caused by localized contamination are usually more drastic and may last longer, requiring the motor system more time to adjust its output to overcome this external disturbance, and its response decays relatively slowly during the adjustment process. By setting a reasonable threshold, this application can effectively distinguish the load fluctuation characteristics caused by these two different physical mechanisms, avoiding ambiguity and improving diagnostic accuracy.

[0076] In some preferred embodiments, a specific example is given below. Assume that in the actual operation of a floor cleaning and maintenance equipment for urban rail transit, a high-frequency narrow-pulse current signal is applied to the motor driving the cleaning brush, and the transient response data of the motor is collected. After analyzing this data, response characteristics such as response time, peak overshoot, decay rate, and oscillation frequency are extracted. Specifically, if the system detects that the transient response time of the motor is 0.05 seconds, the peak overshoot is 5%, and the preset first response time threshold is 0.1 seconds and the first overshoot threshold is 10%, then since 0.05 seconds is less than 0.1 seconds and 5% is less than 10%, the system will determine that the current periodic load fluctuation is caused by uneven wear of the cleaning brush bristles. In this case, the control system will adaptively adjust the closed-loop control parameters of the motor according to the state of uneven bristle wear, or superimpose a compensation signal synchronized with the rotation angle of the cleaning brush into the motor control command to counteract the periodic load fluctuation caused by uneven bristle wear, ensuring that the cleaning force of the cleaning brush on the floor is uniform. On the other hand, if the system detects a transient response time of 0.3 seconds and a decay rate of 20% / second for the motor, and the preset second response time threshold is 0.2 seconds and the first decay rate threshold is 30% / second, then since 0.3 seconds is greater than 0.2 seconds and 20% / second is less than 30% / second, the system will determine that the current periodic load fluctuation originates from localized ground contamination. In this case, the control system will adjust and enhance the cleaning strategy, such as increasing the downward pressure of the brush, increasing the brush speed, or performing multiple cleaning cycles in specific areas, to effectively remove localized ground contamination and restore the cleanliness of the ground.

[0077] Furthermore, specific embodiments of this application also disclose a control system for urban rail transit floor cleaning and maintenance equipment, wherein the cleaning and maintenance equipment includes a cleaning brush and a motor; as shown below. Figure 2 As shown, the system includes: The acquisition module 201 is used to acquire the load information of the motor driving the cleaning brush in real time; the load information includes current information and torque information. Analysis module 202 is used to perform frequency analysis on the load information to identify load fluctuation characteristics related to the rotation frequency of the cleaning brush. The determination module 203 is used to determine that the bristles of the cleaning brush are in a non-uniform wear state based on the amplitude and phase changes of the load fluctuation characteristics. The adjustment module 204 is used to adaptively adjust the closed-loop control parameters of the motor and / or superimpose a compensation signal synchronized with the rotation angle of the cleaning brush disc into the control command of the motor according to the non-uniform wear state of the brush bristles, so as to counteract the periodic load fluctuations caused by the non-uniform wear of the brush bristles. The control module 205 controls the motor to perform operations according to the closed-loop control parameters and / or the motor control commands to ensure that the cleaning brush has a uniform cleaning force on the floor.

[0078] The core of the urban rail transit floor cleaning and maintenance equipment control system proposed in this application lies in the intelligent identification and adaptive control of uneven wear of the cleaning brush bristles through the coordinated work of various functional modules.

[0079] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A control method for urban rail transit floor cleaning and maintenance equipment, the cleaning and maintenance equipment comprising a cleaning brush and a motor, characterized in that, The method includes: The load information of the motor driving the cleaning brush is collected in real time; the load information includes current information and torque information. Frequency analysis is performed on the load information to identify load fluctuation characteristics related to the rotation frequency of the cleaning brush. Based on the amplitude and phase changes of the load fluctuation characteristics, it is determined that the bristles of the cleaning brush are in a non-uniform wear state. Based on the fact that the bristles are in a non-uniform wear state, the closed-loop control parameters of the motor are adaptively adjusted and / or a compensation signal synchronized with the rotation angle of the cleaning brush is superimposed on the control command of the motor to counteract the periodic load fluctuations caused by the non-uniform wear of the bristles. The motor is controlled to perform actions according to the closed-loop control parameters and / or the motor control commands to ensure that the cleaning brush applies uniform cleaning force to the floor.

2. The control method for urban rail transit floor cleaning and maintenance equipment according to claim 1, characterized in that, Frequency analysis is performed on the load information to identify load fluctuation characteristics related to the rotation frequency of the cleaning brush, including: performing real-time frequency analysis on the collected load information and extracting specific frequency components related to the rotation frequency of the cleaning brush as the load fluctuation characteristics.

3. The control method for urban rail transit floor cleaning and maintenance equipment according to claim 2, characterized in that, The step of determining that the bristles of the cleaning brush are in a non-uniform wear state based on the amplitude and phase changes of the load fluctuation characteristics includes: After initial equipment debugging or replacement of the brush disk, a reference load frequency spectrum of the motor is established, and the amplitude and phase of the specific frequency components are stored as reference amplitude and reference phase. During equipment operation, the current amplitude and current phase of a specific frequency component are calculated in real time. When the increase in the current amplitude relative to the reference amplitude continuously exceeds a first preset threshold, and the offset of the current phase relative to the reference phase continuously exceeds a second preset threshold, the non-uniform wear state of the cleaning brush is determined.

4. The control method for urban rail transit floor cleaning and maintenance equipment according to claim 1, characterized in that, The motor uses a PID closed-loop controller; the method further includes: determining the severity and location of the uneven wear while determining that the bristles of the cleaning brush are in a non-uniform wear state. The adaptive adjustment of the closed-loop control parameters of the motor includes: dynamically adjusting at least one of the proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd of the PID closed-loop controller according to the severity and location of the non-uniform wear and a preset parameter mapping table.

5. The control method for urban rail transit floor cleaning and maintenance equipment according to claim 1, characterized in that, Based on the fact that the brush bristles are in a non-uniform wear state, the closed-loop control parameters of the motor are adaptively adjusted and / or a compensation signal synchronized with the rotation angle of the cleaning brush disc is superimposed on the control command of the motor, including: The real-time load data of the motor and the location information of the cleaning and maintenance equipment are collected. Frequency analysis, transient feature analysis and spectral entropy analysis are performed on the real-time load data to obtain load fluctuation characteristics. If the load fluctuation is determined to be caused by uneven wear of the brush bristles or localized contamination of the ground, the closed-loop control parameters of the drive motor are adjusted, or the rotation angle of the cleaning brush is obtained and a compensation signal synchronized with the rotation angle is generated and superimposed on the control command of the motor. If the load fluctuation is determined to be caused by localized contamination of the ground, the enhanced cleaning strategy is adjusted.

6. The control method for urban rail transit floor cleaning and maintenance equipment according to claim 5, characterized in that, Determining whether the load fluctuation characteristics originate from uneven wear of the brush bristles or localized ground contamination includes: While collecting the real-time load data of the motor, the rotation angle of the cleaning brush disc is also collected. A preliminary analysis of the real-time load data is performed to identify periodic load fluctuations; A perturbation signal is applied to the motor, and the transient response data of the motor to the perturbation signal is collected; The transient response data is analyzed to extract response features, and the physical source of the periodic load fluctuation is determined based on the response features; the physical source includes non-uniform wear of brush bristles and localized ground contamination.

7. The control method for urban rail transit floor cleaning and maintenance equipment according to claim 6, characterized in that, The perturbation signal is a high-frequency narrow pulse current signal; the frequency of the high-frequency narrow pulse current signal is higher than the highest harmonic of the rotation frequency of the cleaning brush. Applying a perturbation signal to the motor and acquiring transient response data of the motor to the perturbation signal, including: Obtain the current ambient noise and the rotational speed of the cleaning brush; Based on the ambient noise and the brush rotation speed, adjust the amplitude and duration of the perturbation signal to ensure that the energy of the perturbation signal is higher than the ambient noise and the duration of the perturbation signal is less than the time required for the cleaning brush to rotate one angle range; The motor current, torque, and brush vibration data are collected synchronously as the transient response data.

8. The control method for urban rail transit floor cleaning and maintenance equipment according to claim 6, characterized in that, Determining the physical source of the periodic load fluctuations based on the response characteristics includes: The rotation angle of the cleaning brush disc and the response characteristics of the transient response data are obtained, and the response characteristics are correlated with the rotation angle of the cleaning brush disc. Based on the distribution pattern of the response characteristics at the rotation angle of the cleaning brush, the physical source of the load fluctuation is determined.

9. The control method for urban rail transit floor cleaning and maintenance equipment according to claim 6, characterized in that, The response characteristics include response time, peak overshoot, decay rate, and oscillation frequency; Analyzing the transient response data, extracting response features, and determining the physical source of the periodic load fluctuations based on the response features include: If the response time is less than the first response time threshold and the peak overshoot is less than the first overshoot threshold, then the load fluctuation is determined to be caused by uneven wear of the brush bristles. If the response time is greater than the second response time threshold and the decay rate is less than the first decay rate threshold, then the load fluctuation is determined to originate from local ground pollution.

10. A control system for urban rail transit floor cleaning and maintenance equipment, the cleaning and maintenance equipment comprising a cleaning brush and a motor, characterized in that, The system includes: The data acquisition module is used to acquire the load information of the motor driving the cleaning brush in real time; the load information includes current information and torque information. The analysis module is used to perform frequency analysis on the load information to identify load fluctuation characteristics related to the rotation frequency of the cleaning brush. The determination module is used to determine whether the bristles of the cleaning brush are in a non-uniform wear state based on the amplitude and phase changes of the load fluctuation characteristics. The adjustment module is used to adaptively adjust the closed-loop control parameters of the motor and / or superimpose a compensation signal synchronized with the rotation angle of the cleaning brush disc into the control command of the motor according to the non-uniform wear state of the brush bristles, so as to counteract the periodic load fluctuations caused by the non-uniform wear of the brush bristles. The control module controls the motor to perform operations based on the closed-loop control parameters and / or the motor control commands to ensure that the cleaning brush has a uniform cleaning force on the floor.

Images