Sweeping robot control method and system

Abstract

The invention discloses a sweeping robot control method and system, and the method comprises the steps: calling a pre-established fault feature library according to a fault type outputted by an abnormality recognition module, and carrying out the matching of corresponding abnormality feature data, and obtaining a preliminary abnormality confirmation result; according to the action recovery instruction, an execution unit of the sweeping robot is driven to adjust the wheel rotating speed or reversely rotate a brush, state feedback data after execution is obtained, and whether the recovery action is completed or not is judged; a new cleaning path scheme is generated according to the condition that the current obstacle avoidance capability is insufficient through a path adjustment demand, and the feasibility of an alternative path is determined; through real-time state data in a task execution process, cleaning efficiency and frequent fault problems are continuously monitored, a feature library is circulated, and operation stability is judged.

Description

Technical Field

[0001] The present invention relates to the technical field of household appliance control, and in particular to a method and system for controlling a sweeping robot. Background Art

[0002] As a crucial component of the smart home sector, robot vacuum control technology plays a significant role in improving convenience and cleaning efficiency. This technology not only contributes to a clean and tidy home environment but also plays a key role in the trend toward autonomous and intelligent systems. Its development directly impacts user experience and product competitiveness. However, current robot vacuum control methods on the market still have significant shortcomings. Many products often exhibit poor adaptability and frequent failures in complex environments. In particular, their ability to respond to unusual circumstances is insufficient, leading to interrupted cleaning tasks and inefficient operation.

[0003] Against this backdrop, the field of robot vacuum control faces numerous challenges. The primary issue is how to effectively identify and handle abnormal conditions during the cleaning process, such as wheel slippage or brush jams. These anomalies often directly lead to the robot's inability to operate normally. When an anomaly occurs, without an effective automatic recovery strategy, the robot may become stuck or even come to a complete standstill, further increasing the risk of mission failure. A deeper challenge stems from the lack of flexibility in path planning. When the original path becomes unfeasible due to obstacles or faults, the robot is often unable to adjust its strategy in a timely manner, which in turn affects the overall cleaning effect. These problems are interconnected. Inadequate anomaly recognition leads to the lack of recovery actions, and the lack of recovery capabilities amplifies the impact of path planning defects. Summary of the Invention

[0004] In order to solve the above-mentioned technical problems, the present invention provides a sweeping robot control method and system.

[0005] The technical solution of the present invention is achieved as follows: A method for controlling a sweeping robot, comprising: Obtain real-time status information of the sweeping robot during operation, continuously monitor key parameters, and determine whether there are any abnormal signals; If there is an abnormal signal, the abnormal type is classified to determine the specific fault category; According to the fault type, call the pre-established fault feature library, match the corresponding abnormal feature data, and obtain the preliminary abnormality confirmation result; Based on the abnormal confirmation results, the automatic recovery mechanism is activated, and the corresponding recovery action instructions are generated according to the classification of the abnormal type, and the execution order of the recovery actions is determined; According to the recovery action instruction, the sweeping robot is driven to adjust the working status of its corresponding components, and the status feedback data after execution is obtained to determine whether the recovery action is completed.

[0006] Furthermore, the determining whether there is an abnormal signal includes: Obtain the operating status information of the sweeping robot, record the wheel speed, brush resistance and core parameters in real time, and obtain a preliminary data stream; Based on the preliminary data stream, the wheel speed and brush resistance are compared using a preset threshold range. If a parameter is detected to be outside the preset threshold range, it is determined that an abnormal signal exists in the parameter.

[0007] For the abnormal signal identified, obtain the time point and parameter type corresponding to the abnormal signal to determine the specific context information of the abnormality; Based on the context information, the pre-established support vector machine model is called to classify the abnormal signal and obtain the potential classification of the abnormal signal; Based on the classification results, the duration and change trend of the abnormal signal are extracted to determine whether the abnormal signal has a lasting impact; For abnormal signals with persistent impact, obtain the corresponding operating status data to determine whether to trigger the protection mechanism; If the protection mechanism is triggered, the operating state of the sweeping robot is adjusted through the preset control instructions to obtain the adjusted operating parameters.

[0008] Furthermore, determining the specific fault type includes: Acquire the real-time data stream of wheel speed and brush resistance from the running state, perform signal processing on the real-time data stream, and obtain a standardized data sequence; If the wheel speed in the standardized data sequence is lower than the preset threshold, the wheel slip anomaly detection process is triggered. The data sequence is classified using a pre-established logistic regression model to determine the wheel slip fault category. According to the wheel slip fault category, the time point and duration of the abnormality are extracted. The changing trend of the abnormal signal is calculated through time series analysis method to obtain the dynamic characteristics of the abnormality. If the abnormal dynamic characteristics display duration exceeds the normal range, the motor power output is adjusted through the preset control logic to obtain the adjusted operating parameters; Real-time data of brush resistance is obtained from the operating status. If the brush resistance exceeds the preset threshold, the brush stuck anomaly detection process is triggered. The data is classified through the decision tree model to determine the fault category of the brush stuck.

[0009] According to the fault type of brush jam, the frequency distribution and intensity change of the abnormal signal are analyzed. The frequency domain characteristics of the abnormal signal are extracted by Fourier transform method to obtain the frequency domain analysis results. If the frequency domain analysis results show that the intensity of the abnormal signal is continuously higher than the normal range, the operating state of the brush motor is adjusted through a preset instruction sequence to obtain optimized operating parameters.

[0010] Furthermore, obtaining a preliminary abnormality confirmation result includes: Obtain fault category data, search the pre-established feature library for the data, extract the corresponding abnormal feature data through comparison, and obtain preliminary feature matching results; Based on the preliminary feature matching results, redundant or irrelevant information is removed, and the feature data is formatted through standardization to obtain a standardized feature data set; Based on the normalized feature data set, the preset logical judgment rules are called. If the key indicators in the feature data set exceed the preset threshold range, the data set is marked and the priority sequence of abnormal features is determined; Based on the priority sequence of abnormal features, the feature data is grouped and processed using a classification method to obtain the grouped feature subsets, and the correlation strength between each subset and the fault category is determined; According to the grouped feature subsets and their correlation strength, the decision tree model is used to further analyze the subsets with high correlation to obtain refined abnormal feature classification results; Based on the refined abnormal feature classification results, the performance differences of abnormal features in different time periods are analyzed through the time series comparison method, the dynamic change trend of abnormal features is obtained, and the final abnormality confirmation results are determined.

[0011] According to the final abnormality confirmation result, the corresponding adjustment instruction sequence is generated, and the relevant equipment parameters are optimized and adjusted through the internal control logic to obtain the updated operating status data.

[0012] Furthermore, determining the execution order of the recovery actions includes: Obtain fault category data from the anomaly confirmation results. Based on the classification of wheel slippage or brush jamming, call the preset recovery action library, extract the corresponding recovery action instruction sequence through comparison, and determine the initial action allocation plan; According to the initial action allocation plan, data cleaning methods are used to remove redundant or irrelevant instruction information, and the format of the instruction sequence is adjusted through standardization processing to obtain a standardized action instruction data set; If the key parameters in the normalized action instruction data set exceed the preset threshold range, the instruction sequence is marked through the logical judgment rules to determine the priority sequence of the action instructions; According to the priority sequence of the action instructions, the instruction data set is grouped and processed to obtain the grouped instruction subsets, and the correlation strength between each subset and the fault category is determined; Based on the grouped instruction subsets and their correlation strength, the highly correlated subsets are further analyzed using a decision tree model to obtain refined action instruction classification results. Based on the refined action instruction classification results, the time series analysis method is used to compare the execution performance differences of the action instructions in different time periods, obtain the dynamic adjustment trend of the action instructions, and determine the final recovery action execution sequence; According to the final recovery action execution sequence, the corresponding equipment adjustment parameters are generated through the internal control logic to obtain the updated equipment operation status data.

[0013] Furthermore, the determining whether the recovery action is completed includes: By recovering the action instructions, parsing the instruction content, obtaining the specific parameters for adjusting the wheel speed or reverse rotation of the brush, generating the execution unit drive signal, and determining the drive signal sequence; According to the drive signal sequence, the robot vacuum's execution unit is controlled to run, and the wheel speed or brush rotation status data is collected in real time to obtain the device operation status data set; Adopting the state feedback data analysis method, key operating parameters are extracted from the equipment operation status data set, and a standardized state parameter set is generated through standardization processing; If the key parameters in the normalized state parameter set exceed the preset threshold, the abnormal state is marked through the logical judgment rules to determine the abnormal state classification result; According to the abnormal state classification results, the decision tree model is used to classify and analyze the abnormal parameters, generate the adjusted instruction parameter set, and obtain the optimized drive signal sequence; Through the optimized drive signal sequence, the execution unit drive parameters are updated, and the updated state feedback data is collected to obtain the adjusted equipment operation state; According to the adjusted equipment operating status, the completion of the recovery action is judged by comparing with the preset threshold and the action execution result data is generated.

[0014] Furthermore, after determining the execution order of the recovery actions, the method further includes: If the status feedback data still shows abnormalities after the recovery action is completed, the obstacle distribution is re-evaluated in combination with the current environmental information to determine the path adjustment requirements; Based on the path adjustment requirements, a new cleaning path plan is generated to address the current insufficient obstacle avoidance capabilities and determine the feasibility of the alternative path; Based on the feasibility results of the alternative path, the navigation data of the sweeping robot is updated, and the mobile unit is driven to continue the cleaning task along the new path, and the real-time status of the task execution process is obtained; Through real-time status data during task execution, we continuously monitor cleaning efficiency and fault frequency, circulate the feature library, and judge the stability of operation.

[0015] Furthermore, obtaining the path adjustment requirement includes: According to the abnormal distribution range, the current environmental data is integrated and the preset decision tree model is used to re-evaluate the obstacle distribution to obtain the updated result of the obstacle distribution; Based on the updated obstacle distribution results and the path adjustment requirements, a preliminary path planning scheme is generated to determine the direction and scope of path adjustment. Based on the direction and range of path adjustment, key terrain features are extracted from environmental data, the impact of terrain features on path planning is analyzed, and path feasibility assessment data is obtained; If the path feasibility assessment data is lower than the preset threshold, the path planning scheme is iteratively optimized, integrating the abnormal feature parameters and terrain characteristics to determine the optimized path plan; Generate the corresponding control instruction sequence through the optimized path plan, and verify the instructions with the current environment data to obtain the final path execution parameters; Based on the final path execution parameters, the device operation instructions are updated, and status feedback data and environmental data are collected in real time to determine whether the path adjustment meets the expected requirements.

[0016] Furthermore, determining the feasibility of the alternative path includes: Based on the path adjustment requirements, obstacle distribution information is extracted from the current environmental data and combined with the current status of obstacle avoidance to obtain a preliminary path adjustment direction. Based on the preliminary path adjustment direction, the path planning function is called to conduct key analysis on areas with insufficient capacity, integrate the environmental terrain characteristics, and generate a new cleaning path sketch; For the new cleaning path sketch, an alternative path screening mechanism is used, combined with an avoidance strategy, to conduct a multi-dimensional comparison of the paths and determine the feasibility score of the alternative paths; If the feasibility score of the alternative path is lower than the preset threshold, the cleaning path sketch is partially modified through the solution generation tool, and the dynamic information of obstacle avoidance is integrated to obtain the optimized path solution; Based on the optimized path plan, the verification process is activated to extract real-time feedback information from environmental data to determine whether the path plan meets the requirements of the cleaning task; Based on the verification results of the path plan, the corresponding execution instruction sequence is generated. The security of the instruction sequence is checked in combination with the avoidance strategy to determine the final path execution plan. According to the final path execution plan, the equipment operating parameters are updated, and environmental change data is collected in real time to determine whether the path execution meets the expected cleaning goals.

[0017] A sweeping robot control system, comprising: The sensor and anomaly monitoring module obtains real-time status information of the sweeping robot during operation, including key parameters such as wheel speed and brush resistance. It determines whether there are abnormal signals, such as wheel speed falling below a preset threshold or brush resistance exceeding the normal range. If an abnormal signal is detected, it classifies the abnormality as wheel slippage or brush jamming and determines the specific fault type. It continuously monitors cleaning efficiency and fault frequency, cyclically updates the feature library, and determines the stability of system operation. The fault diagnosis and recovery module, based on the fault categories provided by the sensors and anomaly monitoring module, calls a pre-established fault feature library, matches the corresponding anomaly feature data, and obtains a preliminary anomaly confirmation result. Based on the anomaly confirmation result, it generates corresponding recovery action instructions and determines the execution order of the recovery actions. It drives the robot's execution unit to adjust the wheel speed or reverse the brush rotation, obtains status feedback data after execution, and determines whether the recovery action is complete. The environmental perception and path planning module, if the status feedback data still shows an abnormality after the recovery action is completed, re-evaluates the obstacle distribution based on the current environmental information and determines the path adjustment requirements. Based on the path adjustment requirements, it generates a new cleaning path plan and determines the feasibility of the alternative path. It also updates the navigation data of the sweeping robot, drives the mobile unit to continue the cleaning task along the new path, and obtains the real-time status of the task execution process. The navigation and execution control module updates navigation data based on the new path plan provided by the environmental perception and path planning module; drives the robot's mobile unit to continue cleaning along the new path; obtains real-time status data during task execution and feeds it back to the sensor and anomaly monitoring module; The system management and optimization module continuously monitors cleaning efficiency and fault frequency, and determines the stability of system operation through real-time status data provided by sensors and abnormality monitoring modules; Establish a feature library to optimize the performance of the fault diagnosis and recovery module; manage and coordinate data interaction and task execution between modules to ensure efficient operation of the system.

[0018] Compared with the prior art, the present invention has the following beneficial effects: 1. By real-time monitoring of key parameters such as wheel speed and brush resistance, the present invention can accurately identify common abnormal conditions such as wheel slippage and brush jamming. Once an abnormality is detected, the automatic recovery mechanism is quickly activated to generate and execute targeted recovery actions, such as adjusting the wheel speed or rotating the brush in reverse. This rapid response mechanism effectively avoids situations where the robot stagnates or tasks are interrupted due to abnormalities, significantly improving the adaptability and stability of the equipment in complex environments.

[0019] 2. When the abnormal situation is still not resolved through the initial recovery action, the present invention further activates the environmental adaptability analysis module and re-evaluates the obstacle distribution in combination with the current environmental information. Based on this, the present invention can generate a new cleaning path plan to ensure that the robot can flexibly bypass complex obstacles and continue to complete the cleaning task. This dynamic path planning capability not only improves cleaning efficiency, but also reduces repeated cleaning or missed areas caused by unreasonable path planning, further optimizing the cleaning effect.

[0020] 3. The present invention continuously accumulates and learns new abnormal features and processing experience by cyclically updating the abnormal feature library. As the usage time increases, the robot's ability to identify and handle various abnormal situations will continue to improve, and it can more accurately judge the type of fault and generate more effective recovery strategies, so that the robot can continuously adapt to new challenges when facing complex and changing environments, further improving its intelligence level. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 This is a flowchart of a method for controlling a sweeping robot according to Example 1; Figure 2 This is a framework diagram of a sweeping robot control in Example 2. DETAILED DESCRIPTION

[0022] In order to make the purposes, features, and advantages of the present invention more obvious and easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described below are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. Example 1

[0023] like Figure 1 As shown, this embodiment provides a method for controlling a sweeping robot, including: Obtain real-time status information of the sweeping robot during operation, continuously monitor key parameters, and determine whether there are any abnormal signals; If there is an abnormal signal, the abnormal type is classified to determine the specific fault category; According to the fault type, call the pre-established fault feature library, match the corresponding abnormal feature data, and obtain the preliminary abnormality confirmation result; Based on the abnormal confirmation results, the automatic recovery mechanism is activated, and the corresponding recovery action instructions are generated according to the classification of the abnormal type, and the execution order of the recovery actions is determined; According to the recovery action instruction, the sweeping robot is driven to adjust the working status of its corresponding components, and the status feedback data after execution is obtained to determine whether the recovery action is completed.

[0024] Furthermore, the determining whether there is an abnormal signal includes: Obtain the operating status information of the sweeping robot, record the wheel speed, brush resistance and core parameters in real time, and obtain a preliminary data stream; Based on the preliminary data stream, the wheel speed and brush resistance are compared using a preset threshold range. If a parameter is detected to be outside the preset threshold range, it is determined that an abnormal signal exists in the parameter.

[0025] For the abnormal signal identified, obtain the time point and parameter type corresponding to the abnormal signal to determine the specific context information of the abnormality; Based on the context information, the pre-established support vector machine model is called to classify the abnormal signal and obtain the potential classification of the abnormal signal; Based on the classification results, the duration and change trend of the abnormal signal are extracted to determine whether the abnormal signal has a lasting impact; For abnormal signals with persistent impact, obtain the corresponding operating status data to determine whether to trigger the protection mechanism; If the protection mechanism is triggered, the operating state of the sweeping robot is adjusted through the preset control instructions to obtain the adjusted operating parameters.

[0026] Specifically, in one embodiment, to obtain status information of the sweeping robot in real time, first, a speed sensor installed on the wheel and a pressure sensor on the brush can be used to collect data at a frequency of 100 times per second. For example, the wheel speed data is measured by an encoder, and the speed should be maintained in the range of 300 to 500 revolutions per minute during normal operation. The brush resistance is obtained by a pressure sensor, and the normal value should be between 2.5 and 5.0 Newtons. These key parameters are then continuously monitored, and the collected data is processed in real time by an embedded microcontroller. A sliding window algorithm is used to average the speed and resistance data over the last five seconds. For example, if the average wheel speed is less than 200 revolutions per minute, there may be a risk of jamming or motor failure. If the average brush resistance exceeds 6.0 Newtons, it may indicate that the brush is entangled with foreign matter. To further determine whether there are abnormal signals, based on preset thresholds and historical data analysis, a standard deviation algorithm is used to detect data fluctuations. If the speed standard deviation exceeds 50 rpm or the resistance standard deviation exceeds 1.5 Newtons, it is determined to be an abnormal signal. At the same time, the abnormal data is uploaded to the cloud database for pattern matching and compared with historical fault cases. For example, a machine learning model can predict whether the current abnormality is caused by motor aging, with a prediction accuracy of over 85%. Finally, an exception report is automatically generated and an early warning mechanism is triggered. For example, an instruction is sent to the main control module through the internal communication protocol to reduce the rotation speed to 100 revolutions per minute to protect the hardware. The time and parameter values ​​of the exception are recorded to form a complete log for subsequent optimization algorithms. The above links are processed automatically through data flow, forming a closed-loop logic from collection to analysis to early warning, ensuring real-time monitoring of the operating status of the sweeping robot and timely response to abnormalities.

[0027] Furthermore, determining the specific fault type includes: Acquire the real-time data stream of wheel speed and brush resistance from the running state, perform signal processing on the real-time data stream, and obtain a standardized data sequence; If the wheel speed in the standardized data sequence is lower than the preset threshold, the wheel slip anomaly detection process is triggered. The data sequence is classified using a pre-established logistic regression model to determine the wheel slip fault category. According to the wheel slip fault category, the time point and duration of the abnormality are extracted. The changing trend of the abnormal signal is calculated through time series analysis method to obtain the dynamic characteristics of the abnormality. If the abnormal dynamic characteristics display duration exceeds the normal range, the motor power output is adjusted through the preset control logic to obtain the adjusted operating parameters; Real-time data of brush resistance is obtained from the operating status. If the brush resistance exceeds the preset threshold, the brush stuck anomaly detection process is triggered. The data is classified through the decision tree model to determine the fault category of the brush stuck.

[0028] According to the fault type of brush jam, the frequency distribution and intensity change of the abnormal signal are analyzed. The frequency domain characteristics of the abnormal signal are extracted by Fourier transform method to obtain the frequency domain analysis results. If the frequency domain analysis results show that the intensity of the abnormal signal is continuously higher than the normal range, the operating state of the brush motor is adjusted through a preset instruction sequence to obtain optimized operating parameters.

[0029] Specifically, in one embodiment, during the operation of the sweeping robot, if it is detected that the wheel speed is lower than a preset threshold or the brush resistance is outside the normal range, the abnormality recognition module is automatically triggered to classify the abnormality type and determine the specific fault category through a series of information technology means; For example, when the wheel speed drops below 180 rpm or the brush resistance exceeds 6.5 Newtons, the data analysis process is initiated. First, speed and resistance data samples from the last 10 seconds are extracted to form a data set at a sampling rate of 50 times per second. Then, the short-term trend values ​​of speed and resistance are calculated using a weighted average algorithm. The speed weight is set to 0.6 and the resistance weight is set to 0.4 to highlight the impact of speed changes on anomalies. If the calculation results show that the speed trend value continues to decrease and is lower than 150 rpm, and the resistance trend value is stable above 6.8 Newtons, it is preliminarily judged that the possibility of wheel slippage is high; Next, the built-in decision tree algorithm is used to compare data features with historical failure patterns. For example, by analyzing the speed drop rate exceeding the threshold of 20 revolutions per second, combined with the resistance fluctuation amplitude being less than 0.3 Newtons, the probability of wheel slippage is further confirmed to be 75%. If the resistance value shows a sudden jump and exceeds 7.0 Newtons, while the speed fluctuation is within the normal range, it is likely that the brush is stuck, with an 80% probability; Subsequently, the classification results were correlated with environmental parameters such as the ground friction coefficient. If the friction coefficient was higher than 0.5, the probability of wheel slippage increased further to 85%. Finally, the fault category information is stored in the local cache and transmitted to the fault management module through the internal protocol for subsequent automatic adjustment of the operation strategy. The entire process does not require human intervention, forming a complete logical chain from anomaly detection to classification.

[0030] Furthermore, obtaining a preliminary abnormality confirmation result includes: Obtain fault category data, search the pre-established feature library for the data, extract the corresponding abnormal feature data through comparison, and obtain preliminary feature matching results; Based on the preliminary feature matching results, redundant or irrelevant information is removed, and the feature data is formatted through standardization to obtain a standardized feature data set; Based on the normalized feature data set, the preset logical judgment rules are called. If the key indicators in the feature data set exceed the preset threshold range, the data set is marked and the priority sequence of abnormal features is determined; Based on the priority sequence of abnormal features, the feature data is grouped and processed using a classification method to obtain the grouped feature subsets, and the correlation strength between each subset and the fault category is determined; According to the grouped feature subsets and their correlation strength, the decision tree model is used to further analyze the subsets with high correlation to obtain refined abnormal feature classification results; Based on the refined abnormal feature classification results, the performance differences of abnormal features in different time periods are analyzed through the time series comparison method, the dynamic change trend of abnormal features is obtained, and the final abnormality confirmation results are determined.

[0031] According to the final abnormality confirmation result, the corresponding adjustment instruction sequence is generated, and the relevant equipment parameters are optimized and adjusted through the internal control logic to obtain the updated operating status data.

[0032] Specifically, in one embodiment, during the operation of the sweeping robot, after the abnormality recognition module outputs the fault category, it automatically calls a pre-established fault feature library for matching to confirm the specific feature data of the abnormality and form a preliminary result; For example, assuming the module output fault category is "drive abnormality," we first extract historical data records related to drive abnormalities from the fault signature library. These records contain the motor current and torque output data for the past 500 similar abnormalities, specifically when the current value was between 2.5 amps and 3.8 amps and the torque output was less than 1.2 Nm. The currently collected motor operating data is then grouped and analyzed using a clustering algorithm. Current and torque data from the last five seconds are acquired at a sampling rate of 100 times per minute, forming a data set containing 500 data points. The mean and standard deviation are calculated, revealing a current mean of 3.2 amps and a standard deviation of 0.4, and a torque mean of 1.0 Nm and a standard deviation of 0.15. Next, these statistical results are compared with the baseline values ​​in the feature library. If the current average value exceeds the baseline value of 3.0 amps by more than 5%, and the torque average value is less than 20% of the baseline value of 1.3 Nm, the probability of motor overload is preliminarily determined to be 70%. For further verification, the ambient humidity parameter is introduced for correlation analysis. If the current humidity exceeds 75%, the probability is adjusted to 82% based on the correlation between humidity and motor overload in historical data. Finally, the matching results and analysis data are stored in the local database and transmitted to the strategy adjustment module through the internal communication interface for subsequent optimization of operating parameters. The entire process is automatically completed through algorithms and data-driven, forming a strict logical chain from category output to feature confirmation.

[0033] Furthermore, determining the execution order of the recovery actions includes: Obtain fault category data from the anomaly confirmation results. Based on the classification of wheel slippage or brush jamming, call the preset recovery action library, extract the corresponding recovery action instruction sequence through comparison, and determine the initial action allocation plan; According to the initial action allocation plan, data cleaning methods are used to remove redundant or irrelevant instruction information, and the format of the instruction sequence is adjusted through standardization processing to obtain a standardized action instruction data set; If the key parameters in the normalized action instruction data set exceed the preset threshold range, the instruction sequence is marked through the logical judgment rules to determine the priority sequence of the action instructions; According to the priority sequence of the action instructions, the instruction data set is grouped and processed to obtain the grouped instruction subsets, and the correlation strength between each subset and the fault category is determined; Based on the grouped instruction subsets and their correlation strength, the highly correlated subsets are further analyzed using a decision tree model to obtain refined action instruction classification results. Based on the refined action instruction classification results, the time series analysis method is used to compare the execution performance differences of the action instructions in different time periods, obtain the dynamic adjustment trend of the action instructions, and determine the final recovery action execution sequence; According to the final recovery action execution sequence, the corresponding equipment adjustment parameters are generated through the internal control logic to obtain the updated equipment operation status data.

[0034] Specifically, in one embodiment, when the sweeping robot identifies the fault type of wheel slippage or brush jamming through the abnormality confirmation result, the automatic recovery mechanism module is automatically activated to generate and execute the corresponding recovery action instruction; First, based on the abnormality confirmation results, action templates related to wheel slippage or brush jamming are extracted from the preset recovery action library; For example, if the result is confirmed to be wheel slip, the system searches the action library for 1,000 wheel slip cases and extracts feature data related to wheel speed and ground friction coefficient. If the wheel speed is less than 30% of the normal value of 150 revolutions per minute and the ground friction coefficient is less than 0.3, it is determined to be a slip scenario. Then, the wheel speed data within the last 3 seconds was collected at a sampling rate of 50 times per second, forming a set of 150 data points, with a mean of 100 revolutions per minute and a standard deviation of 10; Combined with the friction coefficient of 0.25 reported by the ground sensor, a decision tree algorithm analysis determined that the cause of the slip was a wet surface. The algorithm then generated a recovery action instruction: first, reduce the wheel speed to 80 revolutions per minute for 2 seconds, while increasing the suction force to a maximum of 1200 Pa to reduce the impact of the wet surface. If the anomaly is a brush stuck, 500 brush stuck cases are extracted from the action library. The characteristic data include the brush motor current exceeding 2.0 amps and the speed being less than 50 rpm; The current data collected shows that the brush motor current is 2.3 amps and the speed is 40 rpm. The logistic regression algorithm calculates the jam probability to be 85%. The recovery action is generated: first stop the brush motor for 1 second, then rotate the brush in the opposite direction for 0.5 seconds, and repeat this three times to try to clear the jam. In terms of execution order, the priority is to deal with wheel slippage to ensure mobile stability, and then deal with brush jams to restore cleaning function; All instructions are transmitted to the execution module through the internal communication interface. After the action is completed, the status of the wheels and brushes is monitored at a frequency of 10 times per second. After confirming that the speed has returned to the normal range (140-160 revolutions per minute for wheels and 80-100 revolutions per minute for brushes), the results are stored in the local database for subsequent optimization reference.

[0035] Furthermore, the determining whether the recovery action is completed includes: By recovering the action instructions, parsing the instruction content, obtaining the specific parameters for adjusting the wheel speed or reverse rotation of the brush, generating the execution unit drive signal, and determining the drive signal sequence; According to the drive signal sequence, the robot vacuum's execution unit is controlled to run, and the wheel speed or brush rotation status data is collected in real time to obtain the device operation status data set; Adopting the state feedback data analysis method, key operating parameters are extracted from the equipment operation status data set, and a standardized state parameter set is generated through standardization processing; If the key parameters in the normalized state parameter set exceed the preset threshold, the abnormal state is marked through the logical judgment rules to determine the abnormal state classification result; According to the abnormal state classification results, the decision tree model is used to classify and analyze the abnormal parameters, generate the adjusted instruction parameter set, and obtain the optimized drive signal sequence; Through the optimized drive signal sequence, the execution unit drive parameters are updated, and the updated state feedback data is collected to obtain the adjusted equipment operation state; According to the adjusted equipment operating status, the completion of the recovery action is judged by comparing with the preset threshold and the action execution result data is generated.

[0036] Specifically, in one embodiment, after the sweeping robot receives the recovery action command, it first parses the command into a specific drive signal through the internal control unit. For the wheel speed adjustment, assuming that the command requires the speed to be increased to 120 revolutions per minute, the control module sends a pulse width modulation signal to the wheel motor at a frequency of 20 times per second, gradually increasing the voltage from 3.2 volts to 4.5 volts, continuously monitoring the current change to ensure that it does not exceed 1.5 amperes to avoid overload. At the same time, the actual wheel speed data is collected through the encoder, and the average value within 5 seconds is calculated to be 118 revolutions per minute, with a deviation within 2%. It is determined that the speed adjustment is preliminarily in line with expectations; Next, if the command involves reverse rotation of the brush, a signal is transmitted to the brush motor driver, setting the reverse rotation speed to 60 revolutions per minute for 1.2 seconds. During this period, the motor torque data is recorded at a sampling rate of 30 times per second. It is found that the torque peak is 0.8 Nm, which is lower than the safety threshold of 1.0 Nm, indicating that there is no additional resistance; After the action was executed, status feedback data was obtained through sensors. The wheel speed stabilized at 119 rpm, and the brush speed reached 85 rpm after resuming forward rotation. Combined with the historical normal value range (115-125 rpm for wheels and 80-90 rpm for brushes), the weighted average algorithm was used to calculate the wheel state recovery index to be 0.92 and the brush state recovery index to be 0.88, both exceeding the threshold of 0.85. The recovery action was considered complete. If the index is lower than the threshold, the secondary adjustment mechanism is automatically triggered to increase the wheel speed by 5% or extend the reverse rotation of the brush by 0.3 seconds. The feedback data is compared with the preset model, and the potential abnormal causes, such as the accumulation of ground particles, are analyzed through the support vector machine algorithm to optimize subsequent instructions.

[0037] Furthermore, after determining the execution order of the recovery actions, the method further includes: If the status feedback data still shows abnormalities after the recovery action is completed, the obstacle distribution is re-evaluated in combination with the current environmental information to determine the path adjustment requirements; Based on the path adjustment requirements, a new cleaning path plan is generated to address the current insufficient obstacle avoidance capabilities and determine the feasibility of the alternative path; Based on the feasibility results of the alternative path, the navigation data of the sweeping robot is updated, and the mobile unit is driven to continue the cleaning task along the new path, and the real-time status of the task execution process is obtained; Through real-time status data during task execution, we continuously monitor cleaning efficiency and fault frequency, circulate the feature library, and judge the stability of operation.

[0038] Furthermore, obtaining the path adjustment requirement includes: According to the abnormal distribution range, the current environmental data is integrated and the preset decision tree model is used to re-evaluate the obstacle distribution to obtain the updated result of the obstacle distribution; Based on the updated obstacle distribution results and the path adjustment requirements, a preliminary path planning scheme is generated to determine the direction and scope of path adjustment. Based on the direction and range of path adjustment, key terrain features are extracted from environmental data, the impact of terrain features on path planning is analyzed, and path feasibility assessment data is obtained; If the path feasibility assessment data is lower than the preset threshold, the path planning scheme is iteratively optimized, integrating the abnormal feature parameters and terrain characteristics to determine the optimized path plan; Generate the corresponding control instruction sequence through the optimized path plan, and verify the instructions with the current environment data to obtain the final path execution parameters; Based on the final path execution parameters, the device operation instructions are updated, and status feedback data and environmental data are collected in real time to determine whether the path adjustment meets the expected requirements.

[0039] Specifically, in one embodiment, after the sweeping robot completes the recovery action, if the status feedback data still shows an abnormality, the environmental adaptability analysis module is automatically started to collect current environmental information through multi-sensor fusion technology. For example, an infrared sensor is used to scan surrounding obstacles at a frequency of 15 times per second, and the distance to the nearest obstacle is calculated to be 0.3 meters. At the same time, the humidity sensor detects that the ground humidity is 45%, and combines historical data to determine that the current environment may have a slippery area. Next, the built-in obstacle distribution assessment algorithm is called, and the scan data is input into a Bayesian probability model based on a grid map. The obstacle distribution probability map is calculated, and the probability of an obstacle within a 30-degree angle ahead is determined to be 75%. The route is then adjusted to avoid this area. Subsequently, based on the evaluation results, the robot generates a path adjustment requirement and replans the path using the A-star algorithm, setting a turning radius of 0.5 meters. After planning is complete, the updated path data is sent to the navigation module at a frequency of 10 times per second, ensuring that the robot moves along the new path at a speed of 0.2 meters per second. At the same time, the gyroscope is used to correct the yaw angle in real time to keep the deviation within 3 degrees. To ensure the effectiveness of path adjustments, the system dynamically analyzes environmental changes during movement. If a new obstacle is detected, the system immediately updates the grid map and recalculates the probability distribution, forming a closed-loop logic for continuous optimization. In addition, to enhance adaptability, humidity data is combined with path planning. If the humidity exceeds 50%, the movement speed is automatically reduced to 0.15 meters per second to improve stability.

[0040] Furthermore, determining the feasibility of the alternative path includes: Based on the path adjustment requirements, obstacle distribution information is extracted from the current environmental data and combined with the current status of obstacle avoidance to obtain a preliminary path adjustment direction. Based on the preliminary path adjustment direction, the path planning function is called to conduct key analysis on areas with insufficient capacity, integrate the environmental terrain characteristics, and generate a new cleaning path sketch; For the new cleaning path sketch, an alternative path screening mechanism is used, combined with an avoidance strategy, to conduct a multi-dimensional comparison of the paths and determine the feasibility score of the alternative paths; If the feasibility score of the alternative path is lower than the preset threshold, the cleaning path sketch is partially modified through the solution generation tool, and the dynamic information of obstacle avoidance is integrated to obtain the optimized path solution; Based on the optimized path plan, the verification process is activated to extract real-time feedback information from environmental data to determine whether the path plan meets the requirements of the cleaning task; Based on the verification results of the path plan, the corresponding execution instruction sequence is generated. The security of the instruction sequence is checked in combination with the avoidance strategy to determine the final path execution plan. According to the final path execution plan, the equipment operating parameters are updated, and environmental change data is collected in real time to determine whether the path execution meets the expected cleaning goals.

[0041] Specifically, in one embodiment, when the sweeping robot detects a need for path adjustment, it automatically generates a new cleaning path plan by analyzing current environmental data to cope with insufficient obstacle avoidance capabilities; First, the laser radar scans the surrounding environment at a frequency of 20 times per second, obtaining the precise location information of the nearest obstacle, which is 0.4 meters away. The ultrasonic sensor detects the height of the obstacle in front and determines it to be a low object. This data is then fed into a path optimization algorithm based on dynamic programming. Combined with the current grid map, an initial path plan is generated, which calculates that the probability of obstacles within a 45-degree angle ahead is 60%. To ensure path feasibility, an improved RRT algorithm was used for path replanning, setting the minimum turning radius to 0.6 meters to generate a smooth path around obstacles, shortening the path length to 2.5 meters; After planning is completed, the new path data is transmitted to the navigation module at a frequency of 8 times per second, controlling the robot to move along the path at a speed of 0.25 meters per second. At the same time, the acceleration sensor monitors the tilt angle in real time to keep the deviation within 2 degrees. To verify the feasibility of the alternative route, historical cleaning data was analyzed to confirm that the new route had a coverage rate of more than 95%; If a dynamic obstacle is detected, the environment map is updated immediately, and the K-means clustering algorithm is used to re-divide the obstacle area and optimize the path selection.

[0042] Furthermore, the real-time status of the task execution process is obtained including: Based on the feasibility assessment results of the alternative paths, the navigation information processing module is called to refresh the path data of the sweeping robot to obtain updated navigation information; Using the updated navigation information, the driving instruction generation process of the mobile unit is activated to generate movement control instructions for the new path and determine the execution direction of the mobile unit; According to the execution direction of the mobile unit, the cleaning task start process is triggered, the equipment is scheduled to carry out the cleaning operation along the specified path, and the initial status of the task execution is obtained at the same time; Based on the initial status of task execution, a real-time status monitoring mechanism is enabled to extract operation data from equipment sensors to determine the progress of the cleaning task; If the progress of the cleaning task deviates from the preset threshold, the path update adjustment tool is called to locally correct the navigation information based on the real-time status data to obtain the adjusted path plan; Based on the adjusted path plan, the control instructions of the mobile unit are regenerated, and the driving device continues to perform the cleaning task along the corrected path, obtaining new real-time status feedback; Through the new real-time status feedback and continuous operation status monitoring mechanism, the matching degree between equipment operation data and cleaning task objectives is analyzed to determine the completion status of task execution.

[0043] Specifically, in one embodiment, the cleaning robot automatically updates navigation data based on the feasibility result of the alternative path, drives the mobile unit to perform the cleaning task along the new path, and obtains the task status in real time; First, the embedded processor loads new path data. The path is generated by the A* algorithm and contains 12 nodes with a node spacing of 0.3 meters and a total length of 3.2 meters. The navigation module analyzes the path coordinates 10 times per second, calculates the driving instructions for the mobile unit, sets the target speed to 0.2 meters per second, and controls the steering angle error within 1.5 degrees. After receiving the command, the mobile unit drives the dual-wheel differential motor, with the left wheel speed at 0.21 meters per second and the right wheel speed at 0.19 meters per second to achieve smooth turning; At the same time, the robot's posture is monitored by a gyroscope sensor, which collects data 15 times per second to detect tilt angle deviations to ensure they do not exceed 3 degrees. If the deviation exceeds the standard, the motor power is automatically adjusted to reduce the speed to 0.15 meters per second for stable operation. In real-time status acquisition, the infrared sensor scans the cleaning area 12 times per second to detect the dust density on the ground. The current value is 0.08 grams per square meter. If it is lower than 0.05 grams per square meter, the marked area is cleaned. Environmental data is fused through the Kalman filter algorithm to generate a dynamic environmental map with a resolution of 0.1 meters and an update frequency of 5 times per second; If a new obstacle is detected, the Dijkstra algorithm is called to recalculate the path, giving priority to solutions that increase the path length by no more than 20%. After the new path is generated, it is transmitted to the navigation module at a frequency of 6 times per second; Analyze real-time cleaning data and calculate the cleaning coverage rate, which is currently 92%. If it is lower than 90%, the cleaning time will be automatically extended by 5 minutes.

[0044] Furthermore, the determination of the stability of the operation includes: Through the real-time status data obtained during the task execution, the data processing module is called to extract features of cleaning efficiency and fault occurrence to obtain a preliminary feature set; Based on the preliminary feature set, the support vector machine algorithm is used to classify the abnormal patterns of cleaning efficiency and fault occurrence and determine the category distribution of abnormal features; If the category distribution of the abnormal feature exceeds the preset threshold, the update process of the abnormal identification module is triggered, and the feature library is incrementally updated in combination with the real-time status data to obtain the updated feature library; Through the updated feature library, the stability analysis tool is called to analyze the matching degree between the operating data and the preset stability indicators to determine the stability status; Generate dynamic adjustment instructions based on the stability status, schedule the data acquisition module to optimize the frequency of real-time status data acquisition, and obtain the adjusted operating data; The random forest algorithm is used to predict the cleaning efficiency and fault occurrence trends through the adjusted operating data, and the changing trend of abnormal occurrence is obtained; Based on the changing trends, the feedback process of the monitoring mechanism is called, the rules are determined, and the long-term stability of the operation is judged.

[0045] Specifically, in one embodiment, the sweeping robot continuously monitors cleaning efficiency and frequent faults through real-time status data, and uses an embedded processor to collect sensor data at a frequency of 8 times per second, including ultrasonic sensors to detect obstacle distances. The current average distance is 0.5 meters, and if it is less than 0.3 meters, the abnormality recognition module is triggered; The anomaly recognition module uses a support vector machine algorithm to analyze a historical fault data feature library containing 500 records. Each record includes parameters such as motor current, wheel speed, and vibration frequency. The feature library is currently updated twice per minute. If the motor current is detected to be abnormally high at 1.2 amps, exceeding the normal value of 1.0 amp, it is marked as a potential fault. Vibration frequency is analyzed through Fourier transform, and a signal with a main frequency of 50 Hz is extracted. If the amplitude exceeds 0.02 mm, it is determined to be a risk of mechanical loosening and this feature is added to the feature library; During cleaning efficiency monitoring, a laser radar is used to scan the environment 10 times per second to generate a map with a resolution of 0.05 meters. The residual dust rate in the cleaning area is calculated. The current rate is 0.06 grams per square meter. If it is higher than 0.07 grams per square meter, the suction power is automatically adjusted to 80 watts and the number of adjustments is recorded. The current number is 3. Stability is determined based on a Bayesian network, integrating sensor data to calculate the probability of stable operation, currently 95%. If it falls below 90%, the movement speed will be automatically reduced to 0.1 meters per second, and the status log will be updated at a frequency of 5 times per second. The log data is compressed using the principal component analysis algorithm, retaining 90% of the information, generating a 100-byte summary, and uploading it to the cloud for long-term stability analysis. Example 2

[0046] like Figure 2 As shown, this embodiment provides a sweeping robot control system, including: The sensor and anomaly monitoring module obtains real-time status information of the sweeping robot during operation, including key parameters such as wheel speed and brush resistance. It determines whether there are abnormal signals, such as wheel speed falling below a preset threshold or brush resistance exceeding the normal range. If an abnormal signal is detected, it classifies the abnormality as wheel slippage or brush jamming and determines the specific fault type. It continuously monitors cleaning efficiency and fault frequency, cyclically updates the feature library, and determines the stability of system operation. The fault diagnosis and recovery module, based on the fault categories provided by the sensors and anomaly monitoring module, calls a pre-established fault feature library, matches the corresponding anomaly feature data, and obtains a preliminary anomaly confirmation result. Based on the anomaly confirmation result, it generates corresponding recovery action instructions and determines the execution order of the recovery actions. It drives the robot's execution unit to adjust the wheel speed or reverse the brush rotation, obtains status feedback data after execution, and determines whether the recovery action is complete. The environmental perception and path planning module, if the status feedback data still shows an abnormality after the recovery action is completed, re-evaluates the obstacle distribution based on the current environmental information and determines the path adjustment requirements. Based on the path adjustment requirements, it generates a new cleaning path plan and determines the feasibility of the alternative path. It also updates the navigation data of the sweeping robot, drives the mobile unit to continue the cleaning task along the new path, and obtains the real-time status of the task execution process. The navigation and execution control module updates navigation data based on the new path plan provided by the environmental perception and path planning module; drives the robot's mobile unit to continue cleaning along the new path; obtains real-time status data during task execution and feeds it back to the sensor and anomaly monitoring module; The system management and optimization module continuously monitors cleaning efficiency and frequent faults, and determines the stability of system operation through real-time status data provided by sensors and anomaly monitoring modules. It also establishes a feature library to optimize the performance of the fault diagnosis and recovery module. It also manages and coordinates data interaction and task execution between modules to ensure efficient operation of the system.

[0047] While the specific embodiments of the present invention have been described in detail above, they are intended only as examples, and the present invention is not limited to the specific embodiments described above. Those skilled in the art should understand that the above embodiments and descriptions are merely illustrative of the principles of the present invention, and that various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for controlling a sweeping robot, characterized in that: include: Obtain real-time status information of the sweeping robot during operation, continuously monitor key parameters, and determine whether there are any abnormal signals; If there is an abnormal signal, the abnormal type is classified to determine the specific fault category; According to the fault type, call the pre-established fault feature library, match the corresponding abnormal feature data, and obtain the preliminary abnormality confirmation result; Based on the abnormal confirmation results, the automatic recovery mechanism is activated, and the corresponding recovery action instructions are generated according to the classification of the abnormal type, and the execution order of the recovery actions is determined; According to the recovery action instruction, the sweeping robot is driven to adjust the working status of its corresponding components, and the status feedback data after execution is obtained to determine whether the recovery action is completed.

2. A sweeping robot control method according to claim 1, characterized in that: Determining whether there is an abnormal signal includes: Obtain the operating status information of the sweeping robot, record the wheel speed, brush resistance and core parameters in real time, and obtain a preliminary data stream; Based on the preliminary data stream, the wheel speed and brush resistance are compared using a preset threshold range. If a parameter is detected to be outside the preset threshold range, it is determined that an abnormal signal exists in the parameter.

3. A sweeping robot control method according to claim 1, characterized in that: Determining the specific fault category includes: Acquire the real-time data stream of wheel speed and brush resistance from the running state, perform signal processing on the real-time data stream, and obtain a standardized data sequence; If the wheel speed in the standardized data sequence is lower than the preset threshold, the wheel slip anomaly detection process is triggered. The data sequence is classified using a pre-established logistic regression model to determine the wheel slip fault category. Real-time data of brush resistance is obtained from the operating status. If the brush resistance exceeds the preset threshold, the brush stuck anomaly detection process is triggered. The data is classified through the decision tree model to determine the fault category of the brush stuck.

4. A sweeping robot control method according to claim 1, characterized in that: The preliminary abnormality confirmation results include: Obtain fault category data, search the pre-established feature library for the data, extract the corresponding abnormal feature data through comparison, and obtain preliminary feature matching results; Based on the preliminary feature matching results, redundant or irrelevant information is removed, and the feature data is formatted through standardization to obtain a standardized feature data set; Based on the normalized feature data set, the preset logical judgment rules are called. If the key indicators in the feature data set exceed the preset threshold range, the data set is marked and the priority sequence of abnormal features is determined; Based on the priority sequence of abnormal features, the feature data is grouped and processed using a classification method to obtain the grouped feature subsets, and the correlation strength between each subset and the fault category is determined; According to the grouped feature subsets and their correlation strength, the decision tree model is used to further analyze the subsets with high correlation to obtain refined abnormal feature classification results; Based on the refined abnormal feature classification results, the performance differences of abnormal features in different time periods are analyzed through the time series comparison method, the dynamic change trend of abnormal features is obtained, and the final abnormality confirmation results are determined.

5. The method for controlling a sweeping robot according to claim 1, wherein: The determining of the execution order of the recovery actions includes: Obtain fault category data from the anomaly confirmation results. Based on the classification of wheel slippage or brush jamming, call the preset recovery action library, extract the corresponding recovery action instruction sequence through comparison, and determine the initial action allocation plan; According to the initial action allocation plan, data cleaning methods are used to remove redundant or irrelevant instruction information, and the format of the instruction sequence is adjusted through standardization processing to obtain a standardized action instruction data set; If the key parameters in the normalized action instruction data set exceed the preset threshold range, the instruction sequence is marked through the logical judgment rules to determine the priority sequence of the action instructions; According to the priority sequence of the action instructions, the instruction data set is grouped and processed to obtain the grouped instruction subsets, and the correlation strength between each subset and the fault category is determined; Based on the grouped instruction subsets and their correlation strength, the highly correlated subsets are further analyzed using a decision tree model to obtain refined action instruction classification results. Based on the refined action instruction classification results, the time series analysis method is used to compare the execution performance differences of the action instructions in different time periods, obtain the dynamic adjustment trend of the action instructions, and determine the final recovery action execution sequence; According to the final recovery action execution sequence, the corresponding equipment adjustment parameters are generated through the internal control logic to obtain the updated equipment operation status data.

6. A sweeping robot control method according to claim 1, characterized in that: Determining whether the recovery action is completed includes: By recovering the action instructions, parsing the instruction content, obtaining the specific parameters for adjusting the wheel speed or reverse rotation of the brush, generating the execution unit drive signal, and determining the drive signal sequence; According to the drive signal sequence, the robot vacuum's execution unit is controlled to run, and the wheel speed or brush rotation status data is collected in real time to obtain the device operation status data set; Adopting the state feedback data analysis method, key operating parameters are extracted from the equipment operation status data set, and a standardized state parameter set is generated through standardization processing; If the key parameters in the normalized state parameter set exceed the preset threshold, the abnormal state is marked through the logical judgment rules to determine the abnormal state classification result; According to the abnormal state classification results, the decision tree model is used to classify and analyze the abnormal parameters, generate the adjusted instruction parameter set, and obtain the optimized drive signal sequence; Through the optimized drive signal sequence, the execution unit drive parameters are updated, and the updated state feedback data is collected to obtain the adjusted equipment operation state; According to the adjusted equipment operating status, the completion of the recovery action is judged by comparing with the preset threshold and the action execution result data is generated.

7. A sweeping robot control method according to claim 1, characterized in that: After determining the execution order of the recovery actions, the method further includes: If the status feedback data still shows abnormalities after the recovery action is completed, the obstacle distribution is re-evaluated in combination with the current environmental information to determine the path adjustment requirements; Based on the path adjustment requirements, a new cleaning path plan is generated to address the current insufficient obstacle avoidance capabilities and determine the feasibility of the alternative path; Based on the feasibility results of the alternative path, the navigation data of the sweeping robot is updated, and the mobile unit is driven to continue the cleaning task along the new path, and the real-time status of the task execution process is obtained; Through real-time status data during task execution, we continuously monitor cleaning efficiency and fault frequency, circulate the feature library, and judge the stability of operation.

8. A sweeping robot control method according to claim 7, characterized in that: Obtaining the path adjustment requirement includes: According to the abnormal distribution range, the current environmental data is integrated and the preset decision tree model is used to re-evaluate the obstacle distribution to obtain the updated result of the obstacle distribution; Based on the updated obstacle distribution results and the path adjustment requirements, a preliminary path planning scheme is generated to determine the direction and scope of path adjustment. Based on the direction and range of path adjustment, key terrain features are extracted from environmental data, the impact of terrain features on path planning is analyzed, and path feasibility assessment data is obtained; If the path feasibility assessment data is lower than the preset threshold, the path planning scheme is iteratively optimized, integrating the abnormal feature parameters and terrain characteristics to determine the optimized path plan; Generate the corresponding control instruction sequence through the optimized path plan, and verify the instructions with the current environment data to obtain the final path execution parameters; Based on the final path execution parameters, the device operation instructions are updated, and status feedback data and environmental data are collected in real time to determine whether the path adjustment meets the expected requirements.

9. The method for controlling a sweeping robot according to claim 7, wherein: Determining the feasibility of alternative paths includes: Based on the path adjustment requirements, obstacle distribution information is extracted from the current environmental data and combined with the current status of obstacle avoidance to obtain a preliminary path adjustment direction. Based on the preliminary path adjustment direction, the path planning function is called to conduct key analysis on areas with insufficient capacity, integrate the environmental terrain characteristics, and generate a new cleaning path sketch; For the new cleaning path sketch, an alternative path screening mechanism is used, combined with an avoidance strategy, to conduct a multi-dimensional comparison of the paths and determine the feasibility score of the alternative paths; If the feasibility score of the alternative path is lower than the preset threshold, the cleaning path sketch is partially modified through the solution generation tool, and the dynamic information of obstacle avoidance is integrated to obtain the optimized path solution; Based on the optimized path plan, the verification process is activated to extract real-time feedback information from environmental data to determine whether the path plan meets the requirements of the cleaning task; Based on the verification results of the path plan, the corresponding execution instruction sequence is generated. The security of the instruction sequence is checked in combination with the avoidance strategy to determine the final path execution plan. According to the final path execution plan, the equipment operating parameters are updated, and environmental change data is collected in real time to determine whether the path execution meets the expected cleaning goals.

10. A sweeping robot control system, used to implement a sweeping robot control method according to any one of claims 1 to 9, characterized in that: include: The sensor and anomaly monitoring module obtains real-time status information of the sweeping robot during operation, including key parameters such as wheel speed and brush resistance; Determine whether there are abnormal signals, such as wheel speed falling below a preset threshold or brush resistance exceeding the normal range; If an abnormal signal is detected, the abnormality type is classified as wheel slippage or brush jamming, and the specific fault type is determined; cleaning efficiency and fault frequency are continuously monitored, and the feature library is cyclically updated to determine the stability of system operation; The fault diagnosis and recovery module, based on the fault categories provided by the sensors and anomaly monitoring module, calls a pre-established fault feature library, matches the corresponding anomaly feature data, and obtains a preliminary anomaly confirmation result. Based on the anomaly confirmation result, it generates corresponding recovery action instructions and determines the execution order of the recovery actions. It drives the robot's execution unit to adjust the wheel speed or reverse the brush rotation, obtains status feedback data after execution, and determines whether the recovery action is complete. The environmental perception and path planning module, if the status feedback data still shows an abnormality after the recovery action is completed, re-evaluates the obstacle distribution based on the current environmental information and determines the path adjustment requirements. Based on the path adjustment requirements, it generates a new cleaning path plan and determines the feasibility of the alternative path. It also updates the navigation data of the sweeping robot, drives the mobile unit to continue the cleaning task along the new path, and obtains the real-time status of the task execution process. The navigation and execution control module updates the navigation data based on the new path plan provided by the environmental perception and path planning module, and drives the mobile unit of the sweeping robot to continue the cleaning task along the new path; Obtain real-time status data during task execution and feed it back to the sensor and anomaly monitoring module; The system management and optimization module continuously monitors cleaning efficiency and frequent faults, and determines the stability of system operation through real-time status data provided by sensors and anomaly monitoring modules. It also establishes a feature library to optimize the performance of the fault diagnosis and recovery module. It also manages and coordinates data interaction and task execution between modules to ensure efficient operation of the system.

Images