Fault prediction method and device applied to AMR and electronic equipment

Abstract

The embodiment of the invention discloses a fault prediction method applied to AMR, and the method comprises the following steps: carrying out the preprocessing of collected AMR operation data, and enabling the data to serve as a training data set of a large model training module; inputting the training data set, and training the large model training module; and calling real-time operation data of the AMR, inputting the real-time operation data into the large model training module for analysis, and outputting a fault prediction result of the AMR. According to the fault prediction method provided by the embodiment of the invention, the accuracy, reliability and generalization ability of AMR fault prediction can be effectively improved by relying on the strong feature extraction and mode recognition ability of the large model. The embodiment of the invention further provides a fault prediction device, electronic equipment and a storage medium.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of intelligent logistics. Specifically, it relates to a fault prediction method, device, and electronic device applied to AMR. Background Art

[0002] With the accelerating advancement of Industry 4.0 and the development of intelligent technologies, AMR (Autonomous Mobile Robot) has been widely used in fields such as logistics warehousing and intelligent manufacturing due to its advantages such as autonomous navigation and intelligent decision-making. It can flexibly complete tasks such as material handling and equipment inspection in complex industrial scenarios according to preset instructions, effectively improving production efficiency and the refinement level of logistics management.

[0003] However, the operating environment of AMR is complex, with situations such as high temperature, dust, and electromagnetic interference, and it itself integrates complex systems such as multi-sensor fusion, motion control, and energy management. Under long-term continuous operation, it will inevitably face various failure risks. Once an AMR suddenly fails, it will cause the production line to stop and the logistics distribution chain to break, leading to a series of chain reactions such as production delays and order backlogs, causing relatively serious economic losses and reputation damage to the enterprise.

[0004] Currently, AMR fault detection and diagnosis mainly adopt traditional technical solutions. For example, the diagnosis method based on a rule library relies on expert experience to build a rule system and is difficult to quickly adapt to new fault modes; the diagnosis method based on a physical model has strict requirements for model accuracy and is prone to deviation in complex actual working conditions. Although artificial intelligence technology has been continuously expanding in the field of fault diagnosis, and the fault prediction method based on machine learning has achieved certain results, there are also problems such as high cost of obtaining labeled data and weak model generalization ability. Especially when facing rare fault types or sudden changes in the operating environment, the prediction accuracy will be greatly reduced. Summary of the Invention

[0005] In view of the problems existing in the above-mentioned prior art, the embodiments of the present application provide a fault prediction method, device, electronic device, and storage medium, which rely on the powerful feature extraction and pattern recognition capabilities of the large model and can effectively improve the accuracy, reliability, and generalization ability of AMR fault prediction.

[0006] In a first aspect, the embodiments of the present application provide a fault prediction method applied to AMR, including the following steps: Preprocess the collected operation data of AMR as the training data set of the large model training module; Input the training data set and train the large model training module; and Retrieve the real-time operation data of the AMR, input it into the large model training module for analysis, and output the fault prediction result of the AMR.

[0007] Further, the input of the training data set to train the large model training module includes: Divide the training data set into a training set, a validation set, and a test set according to a preset ratio; Input the training set, the validation set, and the test set to train the large model training module, where the training set is used to drive the large model module to learn the data feature patterns in the normal operation and fault states of the AMR, the validation set is used to evaluate the performance of the large model training module in real time during the training process to prevent overfitting of the large model training module, and the test set is used to evaluate the core indicators of the large model training module after the training is completed.

[0008] Further, the large model training module selects a deep learning model based on the diffusion architecture.

[0009] Further, the preprocessing of the collected operation data of the AMR as the training data set of the large model training module includes: Clean, denoise, and / or normalize the collected operation data of the AMR as the training data set of the large model training module.

[0010] Further, before the preprocessing of the collected operation data of the AMR as the training data set of the large model training module, it further includes: Real-time collect the operation data of the AMR, where the operation data includes the sensor data, motor operation data, and task-related data of the AMR.

[0011] Further, after retrieving the real-time operation data of the AMR, inputting it into the large model training module for analysis, and outputting the fault prediction result of the AMR, it further includes: Use a fault tracing algorithm to locate the cause of the fault and generate a fault warning message.

[0012] Further, after retrieving the real-time operation data of the AMR, inputting it into the large model training module for analysis, and outputting the fault prediction result of the AMR, it further includes: Retrain and evaluate the large model training model according to the environmental dynamic changes and task types of the AMR.

[0013] In a second aspect, an embodiment of the present application provides a fault prediction device applied to an AMR, including: A data processing module, configured to preprocess the operation data of the collected AMR as the training data set of the large model training module; A model training module, configured to input the training data set and train the large model training module; and A fault prediction module, configured to retrieve the real-time operation data of the AMR, input it into the large model training module for analysis, and output the fault prediction result of the AMR.

[0014] In a third aspect, an embodiment of the present application further provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. The processor is configured to implement the fault prediction method according to the first aspect described above when executing the program.

[0015] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored. The computer program is configured to implement the fault prediction method according to the first aspect described above.

[0016] The embodiments of the present application bring the following beneficial effects: In the fault prediction method applied to AMR provided by the embodiment of the present application, first, the operation data of the collected AMR is preprocessed as the training data set of the large model training module, then the training data set is input to train the large model training module, and finally, the real-time operation data of the AMR is retrieved and input into the large model training module for analysis to output the fault prediction result of the AMR. The fault prediction method applied to AMR provided by the embodiment of the present application relies on the powerful feature extraction and pattern recognition capabilities of the large model, and can effectively improve the accuracy, reliability, and generalization ability of AMR fault prediction. Description of the Drawings

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on the structures shown in these drawings without creative efforts.

[0018] Figure 1 It is a flowchart of a fault prediction method applied to AMR provided by an embodiment of the present application; Figure 2 It is a structural block diagram of a fault prediction device applied to AMR provided by an embodiment of the present application; Figure 3 It is a structural schematic diagram of an electronic device provided by an embodiment of the present application.

[0019] The realization, functional features and advantages of the present application will be further described in conjunction with embodiments with reference to the accompanying drawings. Detailed implementation manners

[0020] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without creative efforts shall fall within the protection scope of the present application.

[0021] In the specification, claims and above-mentioned accompanying drawings of the present application, the terms "first" and "second" are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present application, unless otherwise specified, the meaning of "a plurality" is two or more. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to specific circumstances.

[0022] The following describes a fault prediction method, device, electronic device and storage medium according to an embodiment of the present application in conjunction with the accompanying drawings.

[0023] Figure 1 is a flowchart of a fault prediction method applied to an AMR according to an embodiment of the present application. As Figure 1 shown, the fault prediction method of the embodiment of the present application includes the following steps: S101: Preprocess the collected operation data of the AMR as the training data set of the large model training module; First, it is necessary to perform preprocessing operations on the collected operation data of the AMR. The AMR will generate a large amount of data during actual operation, but these original data may have problems such as noise, missing values, and inconsistent formats, and direct use will affect the effect of subsequent model training. Therefore, through preprocessing means such as data cleaning, normalization, and feature extraction, the original data is converted into a high-quality data set suitable for model training, laying a foundation for the subsequent steps. The preprocessed data will be used as the training data set of the large model training module.

[0024] S102: Input the training data set and train the large model training module; and Next, the preprocessed training data set is input into the large model training module to train the module. The large model has strong learning and generalization capabilities and can mine the potential relationship between AMR operation data and faults from a large amount of training data. During the training process, the model parameters are continuously adjusted so that the model can fit the training data as accurately as possible, thereby learning effective fault prediction patterns.

[0025] S103: Retrieve the real-time operation data of the AMR, input it into the large model training module for analysis, and output the fault prediction result of the AMR.

[0026] Finally, in actual application scenarios, the real-time operation data of the AMR is retrieved and input into the trained large model training module for analysis. The model will process and judge the real-time data based on the learned knowledge, and finally output the fault prediction results of the AMR. In this way, possible faults of the AMR can be discovered in advance so that timely measures can be taken for repair and maintenance to avoid losses caused by faults.

[0027] Therefore, in the fault prediction method applied to AMR provided in the embodiment of the present application, the collected operation data of the AMR is first preprocessed as a training data set for the large model training module, and then the training data set is input to train the large model training module. Finally, the real-time operation data of the AMR is retrieved and input into the large model training module for analysis, and the fault prediction result of the AMR is output. The fault prediction method applied to AMR provided in the embodiment of the present application relies on the powerful feature extraction and pattern recognition capabilities of the large model, which can effectively improve the accuracy, reliability and generalization ability of AMR fault prediction.

[0028] The fault prediction method for AMR provided in the embodiment of the present application, relying on the feature extraction and pattern recognition capabilities of the large model, can deeply explore the complex features in the AMR operation data and accurately distinguish the data patterns of normal operation and fault status. Compared with traditional methods and simple machine learning algorithms, the fault prediction method provided in the embodiment of the present application can significantly reduce the false alarm rate and missed alarm rate in the prediction of known fault types. For new and rare fault types, the fault prediction method provided in the embodiment of the present application also shows good prediction capabilities, effectively improving the accuracy and reliability of fault prediction, and providing a solid guarantee for the stable operation of AMR.

[0029] In addition, the fault prediction method provided in the embodiment of the present application enables the large model to have good generalization performance through massive data training and diversified scenario simulation. Whether facing AMRs of different brands and models, differentiated operating environments, or changing operating tasks, it can adapt quickly without redeveloping models or formulating rules for specific scenarios. This greatly improves the versatility and adaptability of the system and reduces the cost of use and the difficulty of maintenance.

[0030] Moreover, in terms of early fault warning, the fault prediction method provided by the embodiment of the present application can timely detect potential risks at the embryonic stage of faults and issue early warning signals by virtue of its in-depth analysis of data and its ability to capture abnormal features. This reserves sufficient time for maintenance personnel to troubleshoot and repair faults, effectively avoids the expansion of faults, reduces equipment downtime, and thus reduces economic losses caused by production interruptions, thus ensuring the continuity and stability of production.

[0031] Further, in some embodiments of the present application, inputting the training data set to train the large model training module includes: Dividing the training data set into a training set, a validation set and a test set according to a preset ratio; The training set, the validation set and the test set are input to train the large model training module, wherein the training set is used to drive the large model module to learn the data feature patterns under the normal operation and fault conditions of the AMR, the validation set is used to evaluate the performance of the large model training module in real time during the training process to prevent the large model training module from overfitting, and the test set is used to evaluate the core indicators of the large model training module after the training is completed.

[0032] Specifically, during the training phase, the preprocessed data set is scientifically divided into a training set, a validation set, and a test set according to preset ratios, such as 70%, 15%, and 15%. The training set is used to drive the model to learn the data feature patterns under normal operation and fault conditions of the AMR; the validation set evaluates the model performance in real time during the training process, and prevents the model from overfitting by dynamically adjusting hyperparameters such as the learning rate, number of network layers, and number of hidden units; the test set comprehensively evaluates the core indicators such as accuracy, recall rate, and F1 value after the model training is completed. During the training process, it is necessary to use the back propagation algorithm and the stochastic gradient descent optimization algorithm to continuously iterate and update the model parameters until the model reaches the optimal prediction performance. Therefore, the fault prediction method provided in the embodiment of the present application splits the prepared training data set according to a preset ratio to obtain a training set, a validation set, and a test set. This division method lays the foundation for subsequent model training and evaluation, ensuring the orderliness of the training process and the reliability of the evaluation results.

[0033] Furthermore, in some embodiments of the present application, the large model training module selects a deep learning model based on the diffusion architecture (Transformer).

[0034] Specifically, the AMR operation data is high-dimensional, multi-modal, and has complex associations. The self-attention mechanism of the diffusion model can capture long-distance dependencies in the data, mine spatio-temporal associations of different features, and extract valuable information even if the data is noisy or incomplete. Moreover, it has strong scalability. As the data scale and complexity increase, the model parameters can be increased or the structure can be adjusted to adapt to new requirements without large-scale reconstruction.

[0035] And the diffusion model can improve the prediction accuracy. Its self-attention mechanism can globally perceive features, accurately distinguish operating states, reduce the false alarm and missed detection rates, and can also learn complex patterns to predict rare faults. At the same time, training with massive data and simulating diverse scenarios ensure strong generalization ability, enabling it to quickly adapt to different AMRs, environments, and tasks. In addition, its parallel computing advantage can make full use of computing devices, shorten the training and inference time, achieve real-time fault prediction, and reduce equipment downtime and economic losses.

[0036] In an embodiment of the present application, the large model training module selects a pre-trained model based on the Transformer architecture as the basic framework, and divides the dataset in the data storage module into a training set, a validation set, and a test set according to the ratio of 70%, 15%, and 15%. During the training process, the Adam optimization algorithm is adopted, the initial learning rate is set to 0.001, and the performance of the validation set is evaluated once every 10 training batches. When the loss function value on the validation set no longer decreases continuously for 5 iterations and the accuracy fluctuates, the hyperparameter adjustment mechanism is triggered, such as reducing the number of network layers or the number of hidden units, and then training is restarted. After multiple rounds of iterative training, the test set is used to evaluate the final model. If the model accuracy reaches, for example, [X]%, the recall rate reaches, for example, [X]%, and the F1 value reaches [X], it is determined that the model training is completed and has practical application value.

[0037] Furthermore, in some embodiments of the present application, the preprocessing of the collected operation data of the AMR as the training dataset of the large model training module includes: Cleaning, denoising, and / or normalizing the collected operation data of the AMR as the training dataset of the large model training module.

[0038] Specifically, the data preprocessing module adopts a multi-level processing mechanism. First, it uses an outlier detection algorithm (such as the 3σ principle) to clean the original data and eliminate the error data caused by sensor failures and communication interferences. Then, it uses algorithms such as Kalman filtering and mean filtering to denoise the data, effectively suppressing data fluctuations and noise interferences. Finally, through normalization operations, data of different types and dimensions are uniformly mapped to a standard interval, improving the consistency and usability of the data. The processed data is stored in the data storage module in an orderly manner, forming a high-quality data set to meet the requirements of model training and fault prediction.

[0039] Furthermore, in some embodiments of the present application, before preprocessing the collected operation data of the AMR and using it as the training data set of the large model training module, it further includes: Collecting the operation data of the AMR in real time, where the operation data includes the sensor data, motor operation data, and task-related data of the AMR.

[0040] The fault prediction method provided by the embodiments of the present application constructs an all-round and multi-dimensional data acquisition system, and through wired (such as CAN bus, Ethernet) and wireless (such as Wi-Fi, Bluetooth) communication technologies, it collects multivariate data during the operation of the AMR in real time. Among them, the sensor data includes the environmental perception data of the lidar, the image information of the vision sensor, and the attitude data of the gyroscope and accelerometer; the motor operation data covers key parameters such as current, voltage, and rotation speed; the battery status data records indicators such as power, voltage, and charge-discharge cycle times; and the task-related data includes information such as task type, execution duration, and operation path. The collected data is transmitted to the data preprocessing module for processing in real time after being standardized and encapsulated, providing basic data support for subsequent analysis.

[0041] In an actual industrial application scenario, the data acquisition module establishes a communication connection with the motor driver and battery management system of the AMR through the CAN bus, and takes, for example, 100 milliseconds as the sampling period to obtain data such as motor current, voltage, rotation speed, and battery power and voltage in real time; it is connected to the lidar and vision camera through Ethernet. The lidar collects environmental distance information at a frequency of, for example, [X] times per second, and the vision camera captures image data at a rate of [Y] frames per second. After all the collected data is marked with a timestamp and preliminarily verified locally, it is transmitted to the data preprocessing module through the industrial Ethernet.

[0042] After the data preprocessing module receives the original data, first, for lidar data, by setting a distance threshold and comparing it with the neighborhood data algorithm, abnormal distance values are identified and removed; for motor current data, the Kalman filter algorithm is used for denoising processing to effectively smooth the current fluctuation curve. Subsequently, the minimum-maximum normalization method is adopted to map the motor current data to the interval [0, 1], and the battery power data is normalized to the range of [0, 100]%. The processed data is uniformly stored in the MySQL database to construct a standardized data set.

[0043] Further, in some embodiments of the present application, after retrieving the real-time operation data of the AMR, inputting it into the large model training module for analysis, and outputting the fault prediction result of the AMR, it further includes: Using the fault tracing algorithm to locate the cause of the fault and generate a fault warning message.

[0044] Specifically, when the large model training module completes training and passes the performance verification, it will retrieve the real-time data of the AMR operation from the above data storage module in real time and input it into the trained large model training module for analysis. Based on the learned feature patterns, the model intelligently judges the current operation state of the AMR and outputs the fault prediction result, including information such as whether there is a fault, the type of fault, and the probability of the fault occurring. If a fault risk is detected, this module will further combine historical data with current operation parameters, use the fault tracing algorithm to accurately locate the cause of the fault, and generate a fault warning message containing the type of fault, risk level, fault cause analysis, and maintenance suggestions. The warning message is displayed to the on-site operators through a visual interface, and at the same time sent to professional maintenance personnel in multiple ways such as text messages, emails, and mobile APP push notifications to ensure that the fault can be processed in a timely manner.

[0045] For example, during the operation of the AMR, the latest operation data, including position coordinates, motor state parameters, battery power, etc., is obtained at a frequency of once per second and input into the trained large model for analysis. For example, when the AMR is performing a logistics distribution task, if the model predicts that a certain motor has a fault risk, the fault prediction module will retrieve the operation data of this motor in the past 1 hour, combine it with the historical fault case library, and analyze through the fault diagnosis algorithm that it may be due to excessive motor load resulting in bearing wear. Immediately generate a fault warning message containing the type of fault (abnormal motor bearing), occurrence probability (85%), cause of the fault (long-term high-load operation), and handling suggestions (stop the machine to check the bearing wear condition and replace the bearing if necessary), display it to the operator through the AMR vehicle-mounted display screen, and simultaneously push it to the maintenance personnel's mobile APP to remind them to handle it in a timely manner.

[0046] Further, in some embodiments of the present application, after retrieving the real-time operation data of the AMR, inputting it into the large model training module for analysis, and outputting the fault prediction result of the AMR, it further includes: Retrain and evaluate the large model training model according to the dynamic changes in the environment and task types of the AMR.

[0047] Specifically, considering the dynamic changes in the AMR operating environment (such as temperature, humidity, job scenario adjustment) and the diversity of task types (such as new handling tasks, complex path planning tasks), an intelligent update mechanism needs to be established to support regular (such as once a week) update checks, and at the same time automatically trigger the update process when major changes occur in the AMR operating environment or tasks. The update process includes re-collecting operation data, merging historical data to form a new data set, re-partitioning the training set and test set, and retraining and evaluating the model. If the performance metrics of the new trained model on the test set (such as the accuracy improvement exceeding a preset ratio [X]%) are better than the original model, the new model will be put into actual application; if the performance does not meet expectations, the data quality and model training process will be analyzed in depth, and targeted improvements will be made through data augmentation, optimization of algorithm parameters, etc., to ensure that the large model always maintains strong fault prediction performance.

[0048] For example, the automatic update check is executed at 2:00 am every Sunday by default. When the temperature of the AMR working environment suddenly changes by more than, for example, 10°C, the humidity changes by more than, for example, 30%, or a new type of handling task (such as heavy-load long-distance handling) starts to be executed, the model update process will be immediately triggered. First, collect the operation data of the AMR for, for example, 48 consecutive hours in the new environment or new task, merge it with the historical data, and then re-partition it into a training set, a validation set, and a test set, and retrain the large model. If the accuracy of the new trained model on the test set is improved by more than [X]% compared with the original model, the new model will be deployed to the production environment to replace the original model; if the improvement is less than [X]%, the outlier analysis and data augmentation processing will be performed on the newly collected data, and at the same time, the optimization algorithm parameters (such as the learning rate is adjusted to 0.0005) will be adjusted, and the training and evaluation will be performed again until the performance of the new model meets the actual application requirements.

[0049] Figure 2 It is the structural block diagram of the fault prediction device 200 for AMR according to the embodiments of the present application. As Figure 2 shown, the fault prediction device 200 according to the embodiments of the present application includes a data processing module 210, a model training module 220, and a fault prediction module 230, where: The data processing module 210 is used to preprocess the collected operation data of the AMR as the training data set of the large model training module; The model training module 220 is configured to input the training data set and train the large model training module; and The fault prediction module 230 is configured to retrieve the real-time operation data of the AMR, input the data into the large model training module for analysis, and output the fault prediction result of the AMR.

[0050] In the fault prediction device provided in the embodiment of the present application, first, the operation data of the collected AMR is preprocessed as the training data set of the large model training module, then the training data set is input to train the large model training module, and finally, the real-time operation data of the AMR is retrieved and input into the large model training module for analysis, and the fault prediction result of the AMR is output. The fault prediction method for AMR provided in the embodiment of the present application relies on the powerful feature extraction and pattern recognition capabilities of the large model, and can effectively improve the accuracy, reliability, and generalization ability of AMR fault prediction.

[0051] It should be noted that the specific implementation manner of the file compression device in the embodiment of the present application is similar to the specific implementation manner of the file compression method in the embodiment of the present application. For details, please refer to the description in the method part, and will not be elaborated here.

[0052] Figure 3 It is a schematic structural diagram of the electronic device 300 according to an embodiment of the present application.

[0053] As Figure 3 shown, the electronic device 300 includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 302 or the program loaded from the storage part 302 into the random access memory (RAM) 303. In the RAM 303, various programs and data required for the operation of the electronic device 300 are also stored. The CPU 301, ROM 302, and RAM 303 are connected to each other via a bus 304. The input / output (I / O) interface 305 is also connected to the bus 304.

[0054] The following components are connected to the I / O interface 305: an input part 306 including a keyboard, a mouse, etc.; an output part 307 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc. and a speaker; a storage part 308 including a hard disk, etc.; and a communication part 309 including a network interface card such as a LAN card, a modem, etc. The communication part 309 performs communication processing via a network such as the Internet. The drive 310 is also connected to the I / O interface 305 as required. A removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 310 as required, so that the computer program read from it can be installed into the storage part 308 as required.

[0055] In particular, according to an embodiment of the present application, the processes described above with reference to the flowchart can be implemented as computer software programs. For example, an embodiment of the present application includes a computer program product that includes a computer program carried on a machine-readable medium, and the computer program includes program code for performing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 309, and / or installed from the removable medium 311. When the computer program is executed by the central processing unit (CPU) 301, the above functions defined in the electronic device of the present application are executed.

[0056] It should be noted that the computer-readable medium shown in the present application can be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two. The computer-readable storage medium can be, for example, but not limited to, an electronic device, apparatus, or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples of the computer-readable storage medium can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0057] In the present application, the computer-readable storage medium can be any tangible medium that contains or stores a program, and the program can be used by or in combination with an instruction-executing electronic device, apparatus, or device. In the present application, the computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, in which the computer-readable program code is carried. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer-readable signal medium can also be any computer-readable medium other than the computer-readable storage medium, and the computer-readable medium can send, propagate, or transmit a program for use by or in combination with an instruction-executing electronic device, apparatus, or device. The program code contained on the computer-readable medium can be transmitted by any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0058] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of a processing receiving device, method, and computer program product according to various embodiments of the present application. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a portion of code that contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions denoted in the blocks may occur in a different order than that denoted in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based electronic device that performs the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0059] The units or modules involved in the embodiments described in the present application can be implemented in software or in hardware. The described units or modules can also be provided in a processor, and the processor is used to implement the fault prediction method of the present application when executing the program: Preprocess the collected operation data of the AMR as the training data set of the large model training module; Input the training data set to train the large model training module; and Retrieve the real-time operation data of the AMR, input it into the large model training module for analysis, and output the fault prediction result of the AMR.

[0060] As another aspect, the present application also provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiments; or may exist alone without being assembled into the electronic device. The above computer-readable storage medium stores one or more programs, and when the above-mentioned programs are used by one or more processors to execute the fault prediction method described in the present application: Preprocess the collected operation data of the AMR as the training data set of the large model training module; Input the training data set to train the large model training module; and Retrieve the real-time operation data of the AMR, input it into the large model training module for analysis, and output the fault prediction result of the AMR.

[0061] As another aspect, the present application also provides a computer program product, which may be included in the electronic device described in the above embodiments; or may exist alone without being assembled into the electronic device. The above computer program product stores one or more programs, and when the foregoing program is executed by one or more processors to perform the fault prediction method described in the present application: Preprocess the collected operation data of the AMR as the training data set for the large model training module; Input the training data set to train the large model training module; and Retrieve the real-time operation data of the AMR, input it into the large model training module for analysis, and output the fault prediction result of the AMR.

[0062] The above are only the preferred embodiments of the present application, and do not limit the patent scope of the present application. Any equivalent structural transformation made by using the content of the specification and drawings of the present application under the application concept of the present application, or direct / indirect application in other related technical fields, is included in the patent protection scope of the present application.

Claims

1. A fault prediction method applied to AMR, characterized in that, Including the following steps: Preprocess the collected operation data of the AMR as the training data set for the large model training module; Input the training data set and train the large model training module; And Retrieve the real-time operation data of the AMR, input it into the large model training module for analysis, and output the fault prediction result of the AMR.

2. The fault prediction method according to claim 1, wherein The step of inputting the training data set and training the large model training module includes: Divide the training data set into a training set, a validation set, and a test set according to a preset ratio; Input the training set, the validation set, and the test set to train the large model training module, where the training set is used to drive the large model training module to learn the data feature patterns in the normal operation and fault states of the AMR, the validation set is used to evaluate the performance of the large model training module in real time during the training process to prevent overfitting of the large model training module, and the test set is used to evaluate the core indicators of the large model training module after the training is completed.

3. The fault prediction method according to claim 2, wherein The large model training module selects a deep learning model based on the diffusion architecture.

4. The fault prediction method according to claim 1, wherein The step of preprocessing the collected operation data of the AMR as the training data set for the large model training module includes: Clean, denoise, and / or normalize the collected operation data of the AMR as the training data set for the large model training module.

5. The fault prediction method according to claim 1, characterized in that Before the step of preprocessing the collected operation data of the AMR as the training data set for the large model training module, it further includes: Collect the real-time operation data of the AMR, and the operation data includes the sensor data, motor operation data, and task-related data of the AMR.

6. The fault prediction method according to claim 1, wherein After the step of retrieving the real-time operation data of the AMR, inputting it into the large model training module for analysis, and outputting the fault prediction result of the AMR, it further includes: Use a fault tracing algorithm to locate the fault cause and generate a fault warning message.

7. The fault prediction method according to claim 1, wherein After the step of retrieving the real-time operation data of the AMR, inputting it into the large model training module for analysis, and outputting the fault prediction result of the AMR, it further includes: Retrain and evaluate the large model training model according to the dynamic changes of the AMR's environment and task types.

8. A fault prediction device applied to AMR, characterized in that, Including: A data processing module for preprocessing the collected operation data of the AMR as the training data set for the large model training module; A model training module for inputting the training data set and training the large model training module; And A fault prediction module for retrieving the real-time operation data of the AMR, inputting it into the large model training module for analysis, and outputting the fault prediction result of the AMR.

9. An electronic device, characterized in that, Including a memory, a processor, and a computer program stored on the memory and executable on the processor, and the processor is used to implement the fault prediction method according to any one of claims 1-7 when executing the program.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, The computer program is used to implement the fault prediction method according to any one of claims 1-7.

Images