Method for monitoring and regulating operation data of unmanned sweeper

Abstract

The present application relates to the technical field of intelligent environmental sanitation data processing, and aims at the problems of untimely abnormality discovery, false alarm or misoperation in abnormal state monitoring of unmanned sweeper, proposes an unmanned sweeper operation data monitoring and regulation method, which comprises collecting multi-source monitoring data in the operation process of the unmanned sweeper, performing correlation analysis and dynamic modeling on the multi-source monitoring data, obtaining a comprehensive abnormal probability value of monitoring data of any dimension, performing hierarchical judgment on the current operation state of the unmanned sweeper based on the comprehensive abnormal probability value, and regulating operation decision according to the result of hierarchical judgment. The method can effectively improve the abnormality recognition accuracy, autonomous recovery capability and continuous stable operation capability of the unmanned sweeper in complex dynamic environment.

Description

Technical Field

[0001] This invention relates to the field of intelligent sanitation data processing technology, and in particular to a method for monitoring and controlling the operation data of unmanned sweepers. Background Technology

[0002] With the continuous development of smart cities, intelligent sanitation, and autonomous driving technologies, unmanned sweepers are increasingly being used in automated cleaning operations in industrial parks, municipal roads, and industrial areas. Typically, unmanned sweepers integrate a drive system, side brush system, roller brush system, dust collection system, water spraying system, and various environmental sensing and operational status monitoring devices. During operation, they can autonomously complete tasks such as garbage sweeping, dust collection, and ground cleaning according to a preset path.

[0003] However, because unmanned sweepers need to operate continuously for extended periods, their operational status and the external environment are constantly changing as the vehicle moves through space and over time. This results in dynamic fluctuations in their operational data. For example, the operational status of unmanned sweepers is easily affected by changes in garbage density, equipment load fluctuations, ground slippage, road slope, and other complex environmental factors, potentially leading to abnormal situations such as suction inlet blockage, tool entanglement, wheel slippage, and localized operational obstruction. Current methods for monitoring abnormal states of unmanned sweepers often rely on single-dimensional data monitoring or fixed threshold warnings, such as judging abnormalities based on motor current. These methods are ill-suited to adapting to the changing operational states of unmanned sweepers under different working conditions during dynamic operation, easily leading to delayed anomaly detection, false warnings, or malfunctions.

[0004] Therefore, it is necessary to combine various types of data collected during the operation of unmanned sweepers with historical operation data for modeling analysis and dynamic monitoring, so as to realize the identification, early warning and self-recovery regulation of abnormal states, thereby improving the unmanned sweeper's continuous autonomous operation capability, overall operational stability and adaptability to complex operating environments. Summary of the Invention

[0005] To address the problems of untimely anomaly detection, erroneous warnings, or malfunctions in the aforementioned methods for monitoring the abnormal states of unmanned sweepers, this invention provides a method for monitoring and controlling the operation data of unmanned sweepers. By dynamically modeling and analyzing multi-source monitoring data and historical operation data during the operation of the unmanned sweeper, the method enables the identification, early warning, and self-recovery control of abnormal states.

[0006] The above-mentioned objective of the present invention is achieved through the following technical solution:

[0007] The first aspect of the technical solution disclosed in this invention is: a method for monitoring and controlling the operation data of an unmanned sweeper, comprising the following steps:

[0008] Collect multi-source monitoring data during the operation of the unmanned sweeper, including environmental conditions, vehicle attitude, and the operating status of multiple on-board systems;

[0009] By performing correlation analysis and dynamic modeling on multi-source monitoring data, the comprehensive anomaly probability value of monitoring data in any dimension can be obtained;

[0010] Based on the comprehensive anomaly probability value, the current operating status of the unmanned sweeper is classified and judged, and the operation decision is adjusted according to the result of the classification judgment.

[0011] As a further optimized approach, correlation analysis and dynamic modeling are performed on multi-source monitoring data to obtain the comprehensive anomaly probability value of monitoring data in any dimension, including:

[0012] Extract the feature vectors of each time window of the multi-source monitoring data, and construct a time series of feature vectors corresponding to each dimension of the multi-source monitoring data based on the feature vectors;

[0013] Calculate the cosine distance sequence corresponding to the temporal sequence of feature vectors of arbitrary dimensions;

[0014] A first sliding window is established to obtain cosine distance segments of any two dimensions within the same time window. The Spearman correlation coefficient between the two cosine distance segments is calculated to obtain the correlation coefficient sequence. The first sliding window contains M cosine distance values, and the sliding step size is half of M.

[0015] Further analysis of the correlation coefficient sequence yields the anomaly probability value of monitoring data in any dimension.

[0016] A further preferred approach is to express the time series of the feature vectors corresponding to each dimension of the multi-source monitoring data as follows:

[0017] ;

[0018] In the formula, This represents the time series of feature vectors for monitoring data in any dimension. , , ..., These represent the feature vectors of the time series data corresponding to the 1st, 2nd, ..., tth time windows, respectively.

[0019] As a further optimized approach, the correlation coefficient sequence is further analyzed to obtain the anomaly probability value of monitoring data in any dimension, including:

[0020] A second sliding window is established to calculate the stability of the correlation between monitoring data from any two dimensions; the second sliding window contains K correlation coefficient sequences, and the sliding step size is set to half of K.

[0021] Based on the stability of the correlation relationship corresponding to any two dimensions of monitoring data, construct the corresponding correlation relationship change anomaly characterization coefficient;

[0022] By introducing historical operation data and performing probability statistics on the abnormal characterization coefficients of the correlation changes between any two dimensions of monitoring data under the latest monitoring window, the abnormal probability values ​​of any two dimensions of monitoring data are obtained.

[0023] As a preferred implementation, after obtaining the anomaly probability value, the comprehensive coupling reliability of monitoring data from any two dimensions is calculated, and the comprehensive anomaly probability value is obtained based on the anomaly probability value and the comprehensive coupling reliability.

[0024] A further preferred approach is to obtain a comprehensive anomaly probability value based on the anomaly probability value and the comprehensive coupling reliability, including:

[0025] For monitoring data of any dimension, the anomaly probability between it and monitoring data of other dimensions is fused;

[0026] Using the overall coupling reliability as the weight, the overall anomaly probability value of monitoring data in any dimension under the latest monitoring window is calculated.

[0027] As a further optimized approach, the current operating status of the unmanned sweeper is classified and judged based on the comprehensive anomaly probability value, including:

[0028] When the overall anomaly probability value is lower than the first threshold, it indicates that the unmanned sweeper is within the normal fluctuation range;

[0029] When the overall anomaly probability value exceeds the first threshold but does not reach the second threshold, it indicates that the unmanned sweeper has a potential abnormal trend.

[0030] When the overall anomaly probability value exceeds the second threshold, it indicates that the unmanned sweeper has entered a clearly abnormal state.

[0031] As a further optimization option, operational decisions are adjusted based on the results of the hierarchical assessment, including:

[0032] If the unmanned sweeping vehicle is within the normal fluctuation range, only a slight warning or record will be issued;

[0033] If the unmanned sweeping vehicle exhibits a potential abnormal trend, an early warning mechanism will be triggered.

[0034] If the unmanned sweeper enters a clearly abnormal state, the self-recovery control strategy will be activated.

[0035] A further preferred approach is that the execution process of the self-recovery control strategy includes:

[0036] Implement differentiated control measures based on the type of abnormal monitoring data and the corresponding coupling relationship;

[0037] By combining the implementation process of differentiated control measures, the comprehensive anomaly probability value is continuously updated and monitored in real time to determine whether the current control strategy is effective.

[0038] The second aspect of the technical solution disclosed in this invention is: a monitoring and control system for the operation data of an unmanned sweeper, the system being used to perform the steps of the method described above, the system comprising:

[0039] The data acquisition module is used to collect multi-source monitoring data during the operation of the unmanned sweeper. The multi-source monitoring data includes environmental conditions, vehicle attitude, and the operating status of multiple on-board systems.

[0040] The data correlation analysis module is used to perform correlation analysis and dynamic modeling on multi-source monitoring data to obtain the comprehensive anomaly probability value of monitoring data in any dimension.

[0041] The classification judgment module is used to classify and judge the current operating status of the unmanned sweeper based on the comprehensive anomaly probability value.

[0042] The control module is used to adjust operational decisions based on the results of the hierarchical judgment.

[0043] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0044] 1. This invention collects multi-source monitoring data on the environmental state, vehicle attitude, and operating status of multiple onboard systems during the operation of an unmanned sweeper in real time. It extracts feature vectors corresponding to each dimension of monitoring data using a sliding window method, thereby constructing a feature vector sequence. By calculating the cosine distance between adjacent time windows, it obtains the change sequence of monitoring data in multiple dimensions, analyzes the correlation and stability between the change sequences of monitoring data in different dimensions, calculates the corresponding correlation anomaly characterization coefficient, and constructs a multi-dimensional dynamic coupling relationship model. Based on this, it combines the stability change of correlation, the change of correlation strength, and the statistical distribution of historical operation data windows to obtain the comprehensive anomaly probability value corresponding to the monitoring data in multiple dimensions, thereby realizing dynamic identification and trend warning of abnormal states.

[0045] 2. This invention constructs a hierarchical judgment and operation decision control strategy for abnormal types. Based on the comprehensive abnormal probability value, the current operating state of the unmanned sweeper is self-recovering and controlled, realizing adaptive processing of abnormal states such as foreign object entanglement, polyp blockage, wheel slippage and operation obstruction. This effectively improves the accuracy of abnormal identification, autonomous recovery capability and continuous stable operation capability of the unmanned sweeper in complex dynamic environments. Attached Figure Description

[0046] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0047] Figure 1 This is a flowchart of a method for monitoring and controlling the operation data of an unmanned sweeper according to the present invention. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0049] Example 1:

[0050] Reference Figure 1 This invention discloses a method for monitoring and controlling the operation data of an unmanned sweeper, mainly for monitoring, analyzing, and adaptively controlling the operation status of the unmanned sweeper in complex dynamic working environments. The method includes the following steps:

[0051] S1: Collect multi-source monitoring data during the operation of the unmanned sweeper, including environmental conditions, vehicle attitude, and the operating status of multiple on-board systems;

[0052] S2: Perform correlation analysis and dynamic modeling on multi-source monitoring data to obtain the comprehensive anomaly probability value of monitoring data in any dimension;

[0053] S3: Based on the comprehensive anomaly probability value, the current operating status of the unmanned sweeper is classified and judged, and the operation decision is adjusted according to the result of the classification judgment.

[0054] Specifically, in step S1, multi-source monitoring data during the operation of the unmanned sweeper is collected in real time according to a set sampling frequency. This includes environmental status data, vehicle attitude data, and operating status data of multiple onboard systems. In this embodiment, environmental status data is acquired through visual perception devices and environmental humidity sensors installed on the unmanned sweeper, covering the distribution of garbage, obstacle areas, ground slipperiness, and local environmental changes in the current work area. Vehicle attitude data is acquired through level instruments, attitude sensors, wheel speed sensors, etc., installed on the unmanned sweeper, mainly including attitude changes, driving stability, and motion status information during vehicle operation. The operating status data of multiple onboard systems includes current, voltage, speed, and load data during the operation of the drive system, side brush system, dust collection system, and water spray system.

[0055] During the operation of unmanned sweepers, the vehicle's operating status typically changes dynamically with variations in the environment and workload. For example, when the vehicle enters an area with high garbage density, its drive load, side brush load, and suction power usually change synchronously. When there is standing water, changes in slope, or obstacles on the ground, the vehicle's wheel speed, attitude stability, and trajectory will also change accordingly. Furthermore, because unmanned sweepers need to continuously move and operate in different areas, their operating status changes constantly with spatial location and time, meaning the same parameter may correspond to different operating states under different working conditions. For example, the same increase in current could originate from normal heavy-load sweeping, or from foreign object entanglement or suction port blockage. Therefore, environmental status data, vehicle attitude data, and operating status data from multiple onboard systems typically exhibit dynamic correlation and coupling characteristics. Traditional anomaly monitoring based on fixed thresholds or single parameters is ill-suited to adapting to complex dynamic conditions, easily leading to problems such as delayed anomaly identification, false alarms, and untimely control. Therefore, it is necessary to combine multi-source monitoring data during the operation of unmanned sweepers to dynamically analyze, identify, warn of, and self-recovery control abnormal vehicle conditions, thereby improving the autonomous operation capability and operational stability of unmanned sweepers in complex environments.

[0056] In step S2, feature vectors corresponding to each time window are extracted from the multi-source monitoring data. The extraction time window for the feature vectors is set to 0.1 seconds, and the sliding step size is set to half the window length, so as to balance the ability to capture data changes and the continuity of time.

[0057] In this embodiment, for data such as current, voltage, rotational speed, drive load, attitude angle, and wheel speed, feature parameters such as mean (representing the overall level), variance (representing fluctuation characteristics), and range of change (representing the fluctuation range) are extracted from the data within each time window to characterize the operating status and fluctuation of the corresponding system under the current time window. For visual image data, continuous operation images are first acquired through the vehicle-mounted camera. The continuous operation images are preprocessed by grayscale conversion, noise reduction, and target region segmentation. Subsequently, a convolutional neural network (CNN) is used to extract features of garbage distribution, obstacle areas, and ground slipperiness in the images. Furthermore, the YOLO target detection algorithm is used to detect garbage distribution areas, obstacle areas, and abnormal occlusion areas in real time, and extract the number of targets, target area ratio, and spatial distribution features. At the same time, the motion change features between continuous images are extracted by combining optical flow method to characterize environmental changes and local dynamic disturbances during vehicle operation.

[0058] Subsequently, based on the feature vectors extracted under each time window, a time series of feature vectors corresponding to each dimension of the multi-source monitoring data is constructed, which is represented as follows:

[0059] ;

[0060] In the formula, This represents the time series of feature vectors for monitoring data in any dimension. , , ..., These represent the feature vectors of the time series data corresponding to the 1st, 2nd, ..., tth time windows, respectively.

[0061] Calculate the cosine distance between adjacent time windows of the feature vector time series of monitoring data in any dimension, and construct the corresponding cosine distance sequence based on the cosine distance to characterize the change characteristics of the monitoring data in that dimension during the time evolution process.

[0062] Since the unmanned sweeping vehicle operates with multiple components such as the drive system, sweeping system, and vacuuming system working in coordination, the monitoring data from different dimensions often exhibit dynamic correlations during their changes. Therefore, a first sliding window is established to obtain cosine distance segments between any two dimensions within the same time window. The first sliding window contains M cosine distance values, with a sliding step size of half M. Then, the Spearman correlation coefficient between the two cosine distance segments is calculated to characterize the degree of correlation between the changing trends of the monitoring data in the two dimensions within the current window. This allows for the acquisition of a correlation coefficient sequence between the cosine distance sequences corresponding to any two dimensions of monitoring data during the time evolution process.

[0063] During the normal operation of an unmanned sweeper, the drive system, side brush system, vacuuming system, and attitude control system typically operate in a coordinated manner. Therefore, the changes in monitoring data from different dimensions exhibit a certain degree of synchronicity and stability. For example, when the vehicle enters an area with high garbage density, the drive load, side brush load, and vacuuming power usually increase synchronously; similarly, when ground resistance increases, the relationship between changes in vehicle drive torque and vehicle speed remains relatively stable.

[0064] However, when an unmanned sweeper experiences localized anomalies, this stable dynamic correlation can change. For example, when foreign objects become entangled in the side brush system, the current in the side brush system may continuously increase, but the changes in vehicle driving status and suction load are not significant. In this case, the correlation between the side brush system dimension and other dimensions may weaken. Similarly, when a localized blockage occurs at the suction inlet, the suction negative pressure and fan load may abnormally increase, while the overall vehicle operating status changes little, causing a shift in the original multi-dimensional collaborative relationship. Therefore, compared to directly analyzing the magnitude of a single parameter, analyzing the stability of the correlation between monitoring data from different dimensions is more beneficial for identifying abnormal changes under complex operating conditions.

[0065] In this embodiment, a second sliding window is further established to calculate the stability of the correlation between monitoring data from any two dimensions; the second sliding window contains K correlation coefficient sequences, and the sliding step size is set to half of K; the expression for the correlation stability is as follows:

[0066] ;

[0067] In the formula, This indicates the stability of the correlation between the monitoring data of the i-th dimension and the monitoring data of the j-th dimension under the t-th monitoring window; This represents the linear normalization function; K represents the number of correlation coefficient values ​​contained within the window. , These represent the r-th and (r+1)-th correlation coefficient values ​​of the monitoring data of the i-th dimension and the monitoring data of the j-th dimension within the t-th window, respectively. This represents the average absolute value of the difference between adjacent correlation coefficients within the current window. It is used to characterize the degree of fluctuation of the correlation over time. The smaller the value, the more stable the correlation between the two dimensions of monitored data. Conversely, if the value increases, it indicates that the synergistic change relationship between the two may be abnormal.

[0068] Based on the stability of the correlation between monitoring data from any two dimensions, an anomaly characterization coefficient for the change in the correlation between monitoring data from any two dimensions under the t-th window is constructed, and its expression is as follows:

[0069] ;

[0070] In the formula, The coefficient representing the abnormal change in the correlation between the monitoring data of the i-th dimension and the monitoring data of the j-th dimension under the t-th window; e represents the natural constant; , Let represent the stability of the corresponding association relationships under the t-th and t-1-th windows, respectively; , Let represent the mean values ​​of the correlation coefficients under the t-th and t-1-th windows, respectively; This value is used to characterize changes in the stability of association relationships. When the stability of association relationships decreases in the current window, this value will increase, indicating that the volatility of the association relationships is increasing. Used to characterize changes in the strength of the correlation. When the mean of the correlation in the current window decreases, it indicates that the dynamic coupling between the two dimensions of monitored data is weakening, which may indicate an abnormal state.

[0071] By combining the stability and strength of the correlation relationship, the abnormal changes in the collaborative relationship of multi-dimensional monitoring data during the operation of the unmanned sweeper are comprehensively characterized. Therefore, the abnormal characterization coefficient of the correlation relationship change of any two dimensions of monitoring data under each historical monitoring window is obtained, so as to characterize the degree of abnormal evolution of the dynamic coupling relationship between monitoring data of different dimensions.

[0072] Building upon this, to avoid the instability caused by relying solely on the anomaly representation coefficients of a single window, historical operational data is introduced. The anomaly representation level of the current monitoring window is mapped to a probability position in the corresponding historical distribution, thus achieving a probabilistic description of the anomaly representation. Therefore, under the latest monitoring window, the anomaly representation coefficients of the correlation changes between any two dimensions of monitoring data are represented using probabilistic statistics to obtain the anomaly probability values ​​for any two dimensions of monitoring data. The expression is as follows:

[0073] ;

[0074] In the formula, This represents the probability value of an anomaly in the correlation between the monitoring data of the i-th dimension and the monitoring data of the j-th dimension under the latest monitoring window n; This indicates the total number of historical windows preceding this window; This indicates the number of historical windows whose abnormal relationship change index is less than the corresponding value of the current window.

[0075] By monitoring the anomaly probability values ​​of data from any two dimensions, the degree of anomaly can be transformed from a single numerical value into a historical relative position probability, thus obtaining an adaptive relative anomaly situation. After obtaining the anomaly probability values, the comprehensive coupling reliability of the monitoring data from any two dimensions is further introduced to measure the stable and coordinated ability of the two dimensions of monitoring data during historical operation. Its expression is as follows:

[0076] ;

[0077] In the formula, This represents the overall coupling reliability between the i-th and j-th dimensions; This represents the absolute value of the mean of the correlation coefficients between the i-th and j-th dimensions in the historical operation data corresponding to the u-th historical window before the latest monitoring window n. The larger this value is, the stronger the local correlation between the two dimensions is, and the more reliable the cross-reference analysis is. This indicates the stability of the correlation between the i-th and j-th dimensions in the historical data corresponding to the u-th historical window before the latest monitoring window n. The larger this value, the more reliable the relationship. Characterizes the local integrated coupling reliability of the two; Characterizes the overall historical reliability of both, and is used to measure the credibility of monitoring data in anomaly detection.

[0078] After obtaining the anomaly probability values ​​and corresponding comprehensive coupling reliability of monitoring data from any two dimensions, a comprehensive evaluation of the anomaly state of single-dimensional monitoring data is further conducted from the perspective of multidimensional consistency. Specifically, for any dimension monitoring data i, its anomaly probability with other dimension monitoring data is fused, and the comprehensive coupling reliability is used as the weight to calculate the comprehensive anomaly probability value of any dimension monitoring data under the latest monitoring window n. The expression is as follows:

[0079] ;

[0080] In the formula, This represents the overall anomaly probability value of the monitoring data in the i-th dimension; The probability value of anomalies in the correlation between the monitoring data of the i-th dimension and the monitoring data of the j-th dimension under the latest monitoring window n; denoted by , indicating the overall coupling reliability between the i-th and j-th dimensions; z represents the total number of dimensions involved in the analysis; This represents the weighted average of the anomaly probability obtained by combining the latest monitoring window n of the i-th dimension monitoring data with the comprehensive coupling reliability of the monitoring data of other dimensions. The larger this value is, the more abnormal the comprehensive probability of the monitoring data of this dimension is in the latest monitoring window n.

[0081] Through the above steps, the comprehensive anomaly probability value of monitoring data in any dimension under the latest monitoring window n can be obtained. In step S3, based on the comprehensive anomaly probability value, the current operating status of the unmanned sweeper is classified and the operation decision is adjusted. The specific process is as follows:

[0082] First, based on the magnitude of the comprehensive anomaly probability value, the current operating status is divided into different anomaly levels. In this embodiment, the anomaly level is judged by setting a first threshold and a second threshold. When the comprehensive anomaly probability value is lower than the first threshold, it indicates that the unmanned sweeper is within the normal fluctuation range. When the comprehensive anomaly probability value exceeds the first threshold but does not reach the second threshold, it indicates that the unmanned sweeper has a potential abnormal trend. When the comprehensive anomaly probability value exceeds the second threshold, it indicates that the unmanned sweeper has entered a significant abnormal state.

[0083] Accordingly, operational decisions are adjusted based on the results of the tiered assessment. If the unmanned sweeper is within the normal fluctuation range, only a mild warning or record is issued; if the unmanned sweeper shows a potential abnormal trend, an early warning mechanism is triggered; if the unmanned sweeper enters a clearly abnormal state, a self-recovery control strategy is initiated.

[0084] Furthermore, the execution process of the self-recovery control strategy is as follows: based on the type of abnormal monitoring data and the corresponding coupling relationship, differentiated control measures are implemented. For example, when the abnormality mainly originates from the drive system, the system load fluctuation can be reduced by decreasing the driving speed, adjusting the drive torque distribution, and optimizing the path planning method. When the abnormality mainly originates from the side brush or sweeping system, the possible entanglement or blockage can be eliminated by adjusting the side brush speed, starting and stopping the sweeping components, or performing reverse cleaning actions. When the abnormality mainly originates from the vacuuming system, the normal airflow channel can be restored by increasing the instantaneous suction power, performing intermittent backflushing operations, or adjusting the working mode of the suction inlet. When the abnormality originates from local environmental influences (such as slippery ground or dense obstacles), the path replanning mechanism can be combined to detour or perform local repeated cleaning, thereby reducing the impact of the abnormal area on the overall operation.

[0085] During the implementation of differentiated control measures, the comprehensive anomaly probability value is continuously updated and monitored in real time to determine the effectiveness of the current control strategy. If the comprehensive anomaly probability value gradually decreases after control, it indicates that the system status is returning to normal. If the comprehensive anomaly probability value continues to rise or shows no significant improvement after control, the anomaly level of the warning type is further increased, and a safety protection mechanism is triggered when necessary. For example, if foreign objects become entangled in the side brush system or roller brush system, and the comprehensive anomaly probability value of the corresponding dimension continues to rise after self-recovery measures such as reverse cleaning and speed adjustment, it indicates that the entanglement fault has not been resolved. In this case, the vehicle operating speed can be reduced or the cleaning operation can be suspended. If the suction inlet or dust collection pipe is blocked, and the abnormal state persists after measures such as suction adjustment and backflushing cleaning, resulting in a significant decrease in dust collection efficiency, the dust collection system can be stopped and a maintenance warning can be issued to prevent the abnormal state from further expanding.

[0086] By combining the above processing steps, hierarchical early warning and adaptive recovery control of the abnormal probability of monitoring data from multiple dimensions are realized, which significantly improves the continuous and stable operation capability and fault self-handling capability of unmanned sweepers in complex dynamic environments.

[0087] The beneficial effects of the above embodiment 1 are as follows: This invention collects multi-source monitoring data on the environmental state, vehicle attitude, and operating status of multiple on-board systems during the operation of the unmanned sweeper in real time. It extracts feature vectors corresponding to each dimension of monitoring data using a sliding window method, thereby constructing a feature vector sequence. By calculating the cosine distance between adjacent time windows, it obtains the change sequence of monitoring data in multiple dimensions. It analyzes the correlation and stability between the change sequences of monitoring data in different dimensions, calculates the corresponding correlation anomaly characterization coefficient, and constructs a multi-dimensional dynamic coupling relationship model. Based on this, it combines the stability change of correlation, the change of correlation strength, and the statistical distribution of historical operation data windows to obtain the comprehensive anomaly probability value corresponding to the monitoring data in multiple dimensions. Based on the comprehensive anomaly probability value, it performs hierarchical judgment and operation decision control of the unmanned sweeper's operating status, realizes dynamic identification and trend warning of abnormal states, and effectively improves the accuracy of anomaly identification, autonomous recovery capability, and continuous and stable operation capability of the unmanned sweeper in complex dynamic environments.

[0088] Example 2:

[0089] This invention also provides a monitoring and control system for the operation data of an unmanned sweeper vehicle. The system is used to execute the steps described above. The system includes a data acquisition module, a data correlation analysis module, a hierarchical judgment module, and a control module, wherein:

[0090] The data acquisition module is used to collect multi-source monitoring data during the operation of the unmanned sweeper. The multi-source monitoring data includes environmental conditions, vehicle attitude, and the operating status of multiple on-board systems.

[0091] The data correlation analysis module is used to perform correlation analysis and dynamic modeling on multi-source monitoring data to obtain the comprehensive anomaly probability value of monitoring data in any dimension.

[0092] The classification judgment module is used to classify and judge the current operating status of the unmanned sweeper based on the comprehensive anomaly probability value.

[0093] The control module is used to adjust operational decisions based on the results of the hierarchical judgment.

[0094] In addition, embodiments of the present invention also provide an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus.

[0095] Memory, used to store computer programs;

[0096] When the processor executes the program stored in the memory, it implements the steps of the above-mentioned method for monitoring and controlling the operation data of an unmanned sweeper.

[0097] The communication bus mentioned in the above electronic devices can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of representation, only one thick line is used in the diagram, but this does not indicate that there is only one bus or one type of bus.

[0098] The communication interface is used for communication between the aforementioned electronic devices and other devices.

[0099] The memory may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0100] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0101] The present invention also provides a computer-readable storage medium. A computer program is stored in the computer-readable storage medium, and when executed by a processor, the computer program implements the steps of the above-described method for monitoring and controlling the operation data of an unmanned sweeper.

[0102] Optionally, the computer-readable storage medium may be non-volatile memory (NVM), such as at least one disk storage device.

[0103] Optionally, the computer-readable storage medium may also be at least one storage device located remotely from the aforementioned processor.

[0104] In another embodiment of the present invention, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to execute the steps of the above-described method for monitoring and controlling the operation data of an unmanned sweeper.

[0105] It should be noted that, for the embodiments of the device / electronic device / storage medium / computer program, since they are basically similar to the method embodiments, the description is relatively simple. For relevant parts, please refer to the description of the method embodiments. All embodiments of the above-mentioned unmanned sweeper operation data monitoring and control method are applicable to the device, electronic device and storage medium, and can achieve the same or similar beneficial effects.

[0106] It should be noted that the terms "first," "second," etc., are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of apparatuses and methods consistent with some aspects of the invention.

[0107] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such modifications and substitutions should be considered within the scope of protection of the present invention.

Claims

1. A method for monitoring and controlling the operation data of an unmanned sweeper, characterized in that, Includes the following steps: Collect multi-source monitoring data during the operation of the unmanned sweeper, including environmental conditions, vehicle attitude, and the operating status of multiple on-board systems; By performing correlation analysis and dynamic modeling on multi-source monitoring data, the comprehensive anomaly probability value of monitoring data in any dimension can be obtained; Based on the comprehensive anomaly probability value, the current operating status of the unmanned sweeper is classified and judged, and the operation decision is adjusted according to the result of the classification judgment.

2. The method for monitoring and controlling the operation data of an unmanned sweeper according to claim 1, characterized in that, By performing correlation analysis and dynamic modeling on multi-source monitoring data, the comprehensive anomaly probability value of monitoring data in any dimension is obtained, including: Extract the feature vectors of each time window of the multi-source monitoring data, and construct a time series of feature vectors corresponding to each dimension of the multi-source monitoring data based on the feature vectors; Calculate the cosine distance sequence corresponding to the temporal sequence of feature vectors of arbitrary dimensions; A first sliding window is established to obtain cosine distance segments of any two dimensions within the same time window. The Spearman correlation coefficient between the two cosine distance segments is calculated to obtain the correlation coefficient sequence. The first sliding window contains M cosine distance values, and the sliding step size is half of M. Further analysis of the correlation coefficient sequence yields the anomaly probability value of monitoring data in any dimension.

3. The method for monitoring and controlling the operation data of an unmanned sweeper according to claim 2, characterized in that, The expressions for the time series of feature vectors corresponding to each dimension of multi-source monitoring data are as follows: ； In the formula, This represents the time series of feature vectors for monitoring data in any dimension. , , ..., These represent the feature vectors of the time series data corresponding to the 1st, 2nd, ..., tth time windows, respectively.

4. The method for monitoring and controlling the operation data of an unmanned sweeper according to claim 2, characterized in that, Further analysis of the correlation coefficient sequence yields the anomaly probability values ​​for monitoring data in any dimension, including: A second sliding window is established to calculate the stability of the correlation between monitoring data from any two dimensions; the second sliding window contains K correlation coefficient sequences, and the sliding step size is set to half of K; Based on the stability of the correlation relationship corresponding to any two dimensions of monitoring data, construct the corresponding correlation relationship change anomaly characterization coefficient; By introducing historical operation data and performing probability statistics on the abnormal characterization coefficients of the correlation changes between monitoring data of any two dimensions under the latest monitoring window, the abnormal probability values ​​of monitoring data of any two dimensions are obtained.

5. The method for monitoring and controlling the operation data of an unmanned sweeper according to claim 4, characterized in that, After obtaining the anomaly probability value, calculate the comprehensive coupling reliability of monitoring data from any two dimensions, and obtain the comprehensive anomaly probability value based on the anomaly probability value and the comprehensive coupling reliability.

6. The method for monitoring and controlling the operation data of an unmanned sweeper according to claim 5, characterized in that, The comprehensive anomaly probability value is obtained based on the anomaly probability value and the comprehensive coupling reliability, including: For monitoring data of any dimension, the anomaly probability between it and monitoring data of other dimensions is fused; Using the overall coupling reliability as the weight, the overall anomaly probability value of monitoring data in any dimension under the latest monitoring window is calculated.

7. The method for monitoring and controlling the operation data of an unmanned sweeper according to claim 6, characterized in that, Based on the comprehensive anomaly probability value, the current operating status of the unmanned sweeper is classified and judged, including: When the overall anomaly probability value is lower than the first threshold, it indicates that the unmanned sweeper is within the normal fluctuation range; When the overall anomaly probability value exceeds the first threshold but does not reach the second threshold, it indicates that the unmanned sweeper has a potential abnormal trend. When the overall anomaly probability value exceeds the second threshold, it indicates that the unmanned sweeper has entered a clearly abnormal state.

8. The method for monitoring and controlling the operation data of an unmanned sweeper according to claim 7, characterized in that, Adjusting operational decisions based on the results of the hierarchical assessment, including: If the unmanned sweeping vehicle is within the normal fluctuation range, only a slight warning or record will be issued; If the unmanned sweeping vehicle exhibits a potential abnormal trend, an early warning mechanism will be triggered. If the unmanned sweeper enters a clearly abnormal state, the self-recovery control strategy will be activated.

9. The method for monitoring and controlling the operation data of an unmanned sweeper according to claim 8, characterized in that, The execution process of the self-recovery control strategy includes: Implement differentiated control measures based on the type of abnormal monitoring data and the corresponding coupling relationship; By combining the implementation process of differentiated control measures, the comprehensive anomaly probability value is continuously updated and monitored in real time to determine whether the current control strategy is effective.

Images