Fault early warning method and device for automatic teller machine and storage medium

Abstract

The invention discloses a fault early warning method and device for an automatic teller machine and a storage medium, and relates to the field of financial science and technology. The method comprises the following steps: acquiring multi-source data of a target automatic teller machine through acquisition equipment; constructing a digital twinborn model according to the physical structure information and the multi-source data of the target automatic teller machine; determining fault information of the target automatic teller machine according to the digital twin model; early warning information for the target automatic teller machine is generated according to the fault information of the target automatic teller machine and a threshold value set, the threshold value set comprises a plurality of threshold values, the threshold value set is dynamically adjusted based on real-time information of the target automatic teller machine, and each piece of fault information corresponds to one threshold value. The technical problem of insufficient fault diagnosis precision in the existing bank intelligent automatic teller operation and maintenance system is solved.

Description

Technical Field

[0001] This application relates to the field of financial technology, and more specifically, to a fault warning method, device, and storage medium for ATMs. Background Technology

[0002] In the current field of bank ATM (Automated Teller Machine) operation and maintenance, traditional fault diagnosis systems face numerous challenges, particularly in terms of diagnostic accuracy. Existing technologies primarily rely on data collection from single or limited sensors, such as temperature sensors and banknote counters. While this data provides a macroscopic view of the equipment's status, it falls short in fault localization. For example, the system can only identify macroscopic faults like "banknote dispensing failure," but cannot precisely pinpoint whether the cause is a micro-crack in the transmission gears, capacitor aging, or software lag. This vague fault localization significantly increases diagnostic time and reduces operational efficiency.

[0003] Furthermore, the existing system's static threshold alarm mechanism treats both new and old devices equally. This leads to frequent false alarms on new devices due to overly strict threshold settings, while older devices miss early warnings due to lenient thresholds, further exacerbating the inaccuracy of fault diagnosis. In addition, the lack of user behavior analysis limits the system's ability to identify malicious operations, affecting the comprehensiveness and accuracy of fault diagnosis.

[0004] There is currently no effective solution to the above problems. Summary of the Invention

[0005] This application provides a fault early warning method, device, and storage medium for ATMs, to at least solve the technical problem of insufficient fault diagnosis accuracy in existing bank intelligent ATM operation and maintenance systems.

[0006] According to one aspect of the embodiments of this application, a fault early warning method for an ATM is provided, comprising: collecting multi-source data of a target ATM through a data acquisition device, wherein the multi-source data includes at least mechanical vibration information, temperature distribution information, circuit load information, paper status information during cash dispensing, and behavioral information of a target user of the target ATM; constructing a digital twin model based on the physical structure information of the target ATM and the multi-source data, wherein the digital twin model is used to virtually represent the three-dimensional entity information of the target ATM; determining fault information of the target ATM based on the digital twin model, wherein the fault information includes at least the fault location, fault type, and fault occurrence time of the target ATM; and generating early warning information for the target ATM based on the fault information of the target ATM and a threshold set, wherein the threshold set includes multiple thresholds, the threshold set is dynamically adjusted based on real-time information of the target ATM, and each fault information corresponds to a threshold.

[0007] Optionally, determining the fault information of the target ATM based on the digital twin model includes: fusing the physical structure information of the target ATM with multi-source data to obtain target information of the target ATM; extracting at least one fault feature from the target information to obtain a feature set, wherein the fault feature is used to characterize the feature whose impact on the operating state of the target ATM is greater than a preset threshold; and determining the fault information in the target information based on the digital twin model, according to the target mode and the feature set, wherein the target mode is used to capture the correlation between different time scales in the data and to identify information that deviates from the historical normal operating mode based on the correlation.

[0008] Optionally, after determining the fault information of the target ATM based on the digital twin model, the method further includes: generating a risk distribution heatmap based on the fault information, wherein the risk distribution heatmap includes at least the fault information of the target ATM and the predicted risk information within a preset time window; and determining a maintenance strategy based on the risk distribution heatmap and external resource information, wherein the external resource information includes at least maintenance personnel information, equipment inventory status, and traffic status information.

[0009] Optionally, the method further includes: encrypting the fault information into a homomorphic encryption vector and transmitting the homomorphic encryption vector to a target server via a blockchain, wherein the target server is used to manage the information of each ATM in the blockchain; after receiving the homomorphic encryption vector, the target server verifies the homomorphic encryption vector via the blockchain, wherein the verification is used to verify the source and authenticity of the homomorphic encryption vector; after the homomorphic encryption vector passes verification, updating the target model based on the homomorphic encryption vector, wherein the target model is used to predict risk information based on the fault information and generate a defense strategy based on the risk information; and updating the maintenance strategy corresponding to each ATM in the blockchain based on the updated target model.

[0010] Optionally, the threshold set is obtained through the following steps: obtaining the association information of the target ATM, wherein the association information includes the operating status, operating age information, environmental information and historical fault information of each device in the target ATM; and dynamically adjusting the threshold corresponding to each fault information in the threshold set based on the association information.

[0011] Optionally, after generating warning information for the target ATM based on the fault information and threshold set of the target ATM, the method further includes: receiving feedback information obtained by maintenance personnel through augmented reality to verify the fault information determined by the digital twin model and the actual operating information of the target ATM, wherein the feedback information is used to characterize the degree of deviation between the fault information and the actual operating information; and adjusting the digital twin model based on the feedback information.

[0012] Optionally, the method further includes: determining the risk level of the fault information based on the risk distribution heatmap and external resource information; determining the maintenance strategy corresponding to the risk level based on the risk level and the fault information; and generating early warning information for the target ATM based on the threshold corresponding to the risk level and the fault information.

[0013] According to another aspect of the embodiments of this application, a fault warning device for an ATM is also provided, comprising: a data acquisition unit, configured to acquire multi-source data of a target ATM through a data acquisition device, wherein the multi-source data includes at least mechanical vibration information, temperature distribution information, circuit load information, paper status information during the cash dispensing process, and behavioral information of a target user of the target ATM; a construction unit, configured to construct a digital twin model based on the physical structure information of the target ATM and the multi-source data, wherein the digital twin model is used to virtually represent the three-dimensional entity information of the target ATM; a determination unit, configured to determine fault information of the target ATM based on the digital twin model, wherein the fault information includes at least the fault location, fault type, and fault occurrence time of the target ATM; and a generation unit, configured to generate warning information for the target ATM based on the fault information of the target ATM and a threshold set, wherein the threshold set includes multiple thresholds, the threshold set is dynamically adjusted based on the real-time information of the target ATM, and each fault information corresponds to a threshold.

[0014] According to another aspect of this application, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored executable program, wherein, when the executable program is running, it controls the device where the computer-readable storage medium is located to perform the above-described fault warning method for ATMs.

[0015] According to another aspect of this application, an electronic device is also provided, including one or more processors and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to perform the above-described fault warning method for ATMs.

[0016] According to another aspect of the embodiments of this application, a computer program product is also provided, including computer instructions, which, when executed by a processor, implement the steps of the above-described fault warning method for ATMs.

[0017] In this application, firstly, multi-source data of the target ATM is collected using a data acquisition device. This multi-source data includes at least the mechanical vibration information, temperature distribution information, circuit load information, paper status information during the cash dispensing process, and the target user's behavioral information of the target ATM. Next, a digital twin model is constructed based on the physical structure information of the target ATM and the multi-source data. This digital twin model is used to virtually represent the three-dimensional entity information of the target ATM. Then, fault information of the target ATM is determined based on the digital twin model. This fault information includes at least the fault location, fault type, and fault occurrence time of the target ATM. Finally, warning information for the target ATM is generated based on the fault information and a threshold set. This threshold set includes multiple thresholds, which are dynamically adjusted based on the real-time information of the target ATM, with each fault information corresponding to one threshold. By integrating multi-source data with digital twin models, the system achieves precise fault location and intelligent threshold adjustment, thereby improving both fault diagnosis accuracy and response efficiency. This solves the technical problem of insufficient fault diagnosis accuracy in existing bank intelligent automated teller systems. Attached Figure Description

[0018] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0019] Figure 1 This is a flowchart of an optional fault warning method for ATMs according to an embodiment of this application;

[0020] Figure 2 This is a schematic diagram of an optional fault warning device for an ATM according to an embodiment of this application. Detailed Implementation

[0021] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0022] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application 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 this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0023] It should be noted that the information collected in this application (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for display, data used for analysis, etc.) are information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of this data all comply with relevant laws, regulations, and standards, necessary confidentiality measures have been taken, and they do not violate public order and good morals. Corresponding access points are provided for users to choose to authorize or refuse. For example, interfaces are set up between this system and relevant users or organizations, providing users with corresponding access points to choose to agree to or refuse automated decision-making results; if the user chooses to refuse, the process proceeds to the expert decision-making stage.

[0024] According to an embodiment of this application, a method embodiment for fault warning of an ATM is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0025] It should be noted that an intelligent early warning system can serve as the executing entity for the fault early warning method for ATMs in this application embodiment. It is understood that the fault early warning method for ATMs provided in this application embodiment can also be executed by other systems or devices, and this application embodiment does not specifically limit this.

[0026] Figure 1 This is a flowchart of an optional fault warning method for ATMs according to an embodiment of this application, such as... Figure 1 As shown, the method includes the following steps:

[0027] Step S101: Collect multi-source data from the target ATM using the data acquisition device.

[0028] In step S101, the multi-source data includes at least the mechanical vibration information, temperature distribution information, circuit load information, paper status information during the cash dispensing process, and the behavior information of the target ATM.

[0029] Optionally, the data acquisition equipment refers to various sensors and data acquisition devices deployed on the target ATM, including but not limited to vibration sensors, thermal imaging cameras, current monitors, paper thickness sensors, and user behavior monitoring devices (such as operation logs and video surveillance). These devices work together to monitor and acquire the operating status of the target ATM in real time.

[0030] Optionally, multi-source data refers to various types of information collected by the acquisition device, covering mechanical vibration information, temperature distribution information, circuit load information, paper status information during the banknote dispensing process, and target user behavior information. The collection of multi-source data is the foundation for achieving in-depth fault analysis.

[0031] Optionally, the intelligent early warning system uses operation logs and camera data, combined with an NLP intent recognition engine, to identify user behavior information, such as normal operations (e.g., continuous withdrawals) and malicious behavior (e.g., repeatedly tapping the screen).

[0032] Optionally, the intelligent early warning system collects data in real time through multiple sensors on the target ATM. Multi-source data acquisition ensures the comprehensiveness and accuracy of the fault diagnosis system, enabling it to detect anomalies that are easily overlooked by a single sensor, such as abnormal user operations or subtle changes during banknote transfer.

[0033] Optionally, the collected multi-source data can be encrypted by edge nodes and then uploaded to the blockchain, stored in a Merkle tree structure on a private chain to ensure that the original data such as vibration spectrum and heat map cannot be tampered with, providing a credible chain of evidence for subsequent fault liability tracing.

[0034] Step S102: Construct a digital twin model based on the physical structure information of the target ATM and multi-source data.

[0035] In step S102, the digital twin model is used to virtually represent the three-dimensional entity information of the target ATM.

[0036] Optionally, physical structure information refers to data such as the hardware configuration, internal structure, and working principle of the target ATM, including the location, connection method, and functional description of each component.

[0037] Optionally, a digital twin model is a virtual model corresponding to the target ATM physical entity. It not only replicates the ATM's three-dimensional structure but also reflects the device's operational status in real time. When constructing this model, physical structure information needs to be combined with the aforementioned multi-source data so that the model can accurately simulate and predict the behavior of the real device.

[0038] Optionally, the digital twin model, as the central hub for fault diagnosis, can visualize the relationship between multi-source data and physical entities, providing a high-fidelity fault analysis environment that makes fault prediction and location more accurate.

[0039] Optionally, the intelligent early warning system establishes a 3D digital twin corresponding to the physical device in a 1:1 ratio based on the physical structure information and multi-source data of the target ATM. It maps the core modules such as mechanical transmission structure, circuit topology, and banknote path in real time, and dynamically binds the sensor data stream to the twin components (such as specific spectrum vibration corresponding to gear wear, and local temperature rise corresponding to motor overload).

[0040] Step S103: Determine the fault information of the target ATM based on the digital twin model.

[0041] In step S103, the fault information includes at least the fault location, fault type, and fault occurrence time of the target ATM.

[0042] Optionally, fault information refers to specific fault details analyzed and identified in the digital twin model, including the precise location of the fault (e.g., a specific gear or motor), the type of fault (e.g., mechanical wear or circuit overload), and the time of the fault. This step is achieved by comparing the differences between multi-source data and normal operating modes, and by utilizing a pre-trained fault classification model.

[0043] Step S104: Generate early warning information for the target ATM based on the fault information and threshold set of the target ATM.

[0044] In step S104, the threshold set includes multiple thresholds, which are dynamically adjusted based on the real-time information of the target ATM, with each fault information corresponding to a threshold.

[0045] Optionally, the threshold set is a series of judgment criteria that are dynamically adjusted according to the equipment status and environmental conditions to distinguish between normal operation and fault status. The threshold set contains multiple thresholds, each corresponding to different types of data (such as vibration intensity, temperature level, etc.), and can be automatically adjusted based on real-time information to adapt to the dynamic changing environment of the ATM.

[0046] Optionally, when a fault triggers a corresponding threshold, the system will automatically generate a warning notification. This notification includes a preliminary assessment of the fault, the possible risk level, and suggested response measures. The generation of warning information means that the system can promptly identify potential problems and take action to prevent escalation.

[0047] Optionally, the dynamically adjusted threshold set and early warning mechanism improve the system's sensitivity to faults, reduce false alarms and missed alarms, and ensure the timeliness and effectiveness of fault response. Furthermore, by matching thresholds with specific fault information, the system can more intelligently determine when to issue an early warning and the severity of the warning, enhancing the predictability and controllability of fault handling.

[0048] Optionally, the above method can reduce the detection time of latent faults such as gear microcracks, and the positioning accuracy can reach the component level (such as accurately identifying wear of the third drive shaft).

[0049] As can be seen from steps S101 to S104, in this application, firstly, multi-source data of the target ATM is collected by a data acquisition device. This multi-source data includes at least the mechanical vibration information, temperature distribution information, circuit load information, paper status information during the cash dispensing process, and the target user's behavioral information of the target ATM. Next, a digital twin model is constructed based on the physical structure information of the target ATM and the multi-source data. This digital twin model is used to virtually represent the three-dimensional entity information of the target ATM. Then, fault information of the target ATM is determined based on the digital twin model. This fault information includes at least the fault location, fault type, and fault occurrence time of the target ATM. Finally, warning information for the target ATM is generated based on the fault information and a threshold set. This threshold set includes multiple thresholds, which are dynamically adjusted based on the real-time information of the target ATM, with each fault information corresponding to a threshold. By integrating multi-source data with digital twin models, the system achieves precise fault location and intelligent threshold adjustment, thereby improving both fault diagnosis accuracy and response efficiency. This solves the technical problem of insufficient fault diagnosis accuracy in existing bank intelligent automated teller systems.

[0050] In one optional embodiment, the intelligent early warning system fuses the physical structure information of the target ATM with multi-source data to obtain target information of the target ATM. Then, it extracts at least one fault feature from the target information to obtain a feature set. The fault feature is used to characterize the feature whose impact on the operating state of the target ATM is greater than a preset threshold. Then, based on a digital twin model, the system determines the fault information in the target information according to the target mode and the feature set. The target mode is used to capture the correlation between different time scales in the data and to identify information that deviates from the historical normal operation mode based on the correlation.

[0051] Optionally, the intelligent early warning system deeply integrates physical structure information with multi-source data to form target information. This process involves complex data cleaning, format conversion, and correlation matching to ensure that the system can understand the meaning and position of each data point within the overall ATM operating framework. For example, the system might bind the vibration data of a particular gear to its specific location and function, giving the data spatial attributes.

[0052] Optionally, the system identifies all possible fault characteristics through in-depth analysis of target information, such as abnormally high circuit loads, temperature distributions exceeding normal ranges, and unexpected mechanical vibration patterns. These characteristics are screened and confirmed to form a feature set. Then, the intelligent early warning system uses digital twin models and target-oriented methods (such as time series analysis) to deeply mine and compare the data in the feature set to determine which features truly represent the fault, as well as the specific location, type, and time of occurrence of the fault. This process may involve complex statistical analysis, pattern recognition, and machine learning algorithms to ensure the accuracy and reliability of the fault information.

[0053] As described above, the implementation methods not only deepen the extraction and analysis of fault features but also further optimize the fault information determination process by introducing a target-oriented approach. Specifically, by integrating physical structure information and multi-source data, the system can construct a rich and comprehensive target information database. Based on this, through precise fault feature extraction, the system can efficiently screen potential fault points. Finally, by using a digital twin model and a target-oriented approach to conduct in-depth analysis of these features, the system can accurately locate faults, predict their impact, and thus generate effective early warning information. This seamless process greatly improves the speed and accuracy of fault diagnosis, enabling dynamic monitoring and intelligent early warning of the target ATM's operational status.

[0054] In one optional embodiment, the intelligent early warning system generates a risk distribution heat map based on fault information. The risk distribution heat map includes at least fault information of the target ATM and predicted risk information within a preset time window. Then, a maintenance strategy is determined based on the risk distribution heat map and external resource information, wherein the external resource information includes at least maintenance personnel information, equipment inventory status, and traffic status information.

[0055] Alternatively, maintenance strategies may include: low-risk faults (such as software lag) → remote reset; medium-risk faults (such as module aging) → module replacement; high-risk faults (such as core component failure) → complete machine replacement.

[0056] Optionally, the intelligent early warning system first inputs fault information, including fault location, fault type, and fault occurrence time, into a pre-trained risk prediction model. This model utilizes deep learning algorithms, considering various variables such as historical fault records, operating environment, seasonal factors, and equipment age, to predict the probability and impact of each fault recurring or a new fault occurring within a preset time window. Subsequently, the system merges this predicted risk information with current fault information to generate a risk distribution heatmap covering the entire service area. Each heatmap point indicates the maintenance and downtime risk faced by ATM equipment within a certain range, with darker colors indicating higher risks.

[0057] Optionally, the intelligent early warning system, based on high-risk areas revealed by the risk distribution heatmap and combined with current external resource information, such as the location and skills of maintenance personnel, the quantity and type of spare parts in stock, and real-time traffic congestion, uses resource scheduling algorithms (such as dynamic scheduling algorithms based on DRL) to calculate the optimal maintenance path and time window. The system automatically matches the nearest qualified maintenance personnel, plans a reasonable spare parts supply route, and takes into account the impact of traffic conditions to minimize maintenance response time and cost. Simultaneously, the system adjusts inventory based on predicted spare parts demand to ensure that maintenance personnel obtain the necessary spare parts as quickly as possible.

[0058] Optionally, the system can also dynamically adjust spare parts inventory based on the predicted failure rate (e.g., if the gear failure rate increases by 15% next month, the system will automatically increase the purchase).

[0059] Optionally, maintenance personnel can query the device status via natural language commands (such as "query the lifespan of card reader for machine A03"). The system will parse the commands and provide voice feedback with predicted results and maintenance suggestions.

[0060] As can be seen from the above, the aforementioned implementation methods effectively elevate fault early warning to the level of resource scheduling and maintenance strategy optimization. The generation of the risk distribution heatmap not only reveals the spatiotemporal distribution characteristics of ATM equipment faults but also, combined with predictive analysis, provides direction for proactive maintenance and emergency response. Furthermore, the intelligent formulation of maintenance strategies, based on a comprehensive consideration of internal and external factors, achieves efficient utilization of maintenance resources, reduces operation and maintenance costs, and simultaneously improves customer service continuity and satisfaction.

[0061] In one optional embodiment, the intelligent early warning system encrypts fault information into a homomorphic encryption vector and transmits the homomorphic encryption vector to a target server via a blockchain. The target server manages the information of each ATM in the blockchain. After receiving the homomorphic encryption vector, the target server verifies the homomorphic encryption vector via the blockchain. The verification verifies the source and authenticity of the homomorphic encryption vector. After the homomorphic encryption vector passes verification, the target model is updated based on the homomorphic encryption vector. The target model is used to predict risk information based on fault information and generate defense strategies based on risk information. Then, based on the updated target model, the maintenance strategy corresponding to each ATM in the blockchain is updated.

[0062] Optionally, a homomorphic encryption vector is a special form of encrypted data designed to allow certain types of computations to be performed without decryption, thus protecting the privacy of the original data while still enabling analysis. In this embodiment, the intelligent early warning system converts the identified fault information into a homomorphic encryption vector, so that even if the data is intercepted during transmission, its content cannot be deciphered.

[0063] Optionally, the intelligent early warning system converts fault information into homomorphic encrypted vectors. The encrypted vectors are then transmitted via the blockchain network to a centralized target server, which manages information for every ATM in the entire blockchain. During the receipt of the homomorphic encrypted vectors, the target server initiates a smart contract to automatically verify the vector's origin and authenticity, ensuring data security and accuracy.

[0064] Optionally, once the homomorphic encryption vectors are verified, the target server decrypts and analyzes these vectors, integrates them into the target model, and updates the model in real time. The updated model can more accurately predict the risk information that similar faults may occur in the future, thereby generating a more comprehensive defense strategy. Subsequently, the target server again distributes the updated model parameters and corresponding maintenance strategies to every ATM in the network through a blockchain smart contract, realizing intelligent operation and maintenance upgrades across the entire network.

[0065] Optionally, when each branch detects attack characteristics, such as abnormal thermal imaging textures caused by glue blockage, it will immediately convert such sensitive information into homomorphic encryption vectors to ensure the privacy and security of data during transmission. Through distributed transmission via the blockchain network, triple protection is achieved: tamper-proof: each data packet generates a unique digital fingerprint (hash value), and any modification will cause the fingerprint to become invalid; integrity: the data is split into multiple fragments and stored on different nodes, and the original information cannot be restored if any fragment is missing; traceability: the entire transmission path is recorded on the blockchain, which can accurately track the source and flow of data.

[0066] Once the attack signature is detected and converted into a homomorphic encryption vector, a pre-defined smart contract is automatically triggered, initiating the data sharing process. The smart contract performs three core checks:

[0067] 1) Data signature authentication: Verify the submitter's identity (via device digital certificate);

[0068] 2) Encryption compliance check: Confirm that the homomorphic encryption format complies with security standards (e.g., vector length ≥ 256 bits);

[0069] 3) Source trust verification: Verify the historical data submission records of the device (no abnormalities in on-chain evidence storage).

[0070] Verified data will generate a blockchain transaction request and enter the distributed consensus network. If the data is non-compliant, the source device will be immediately isolated and tagged to prevent potential security threats.

[0071] The central server (target server) aggregates homomorphic encrypted vectors from across the network and trains an updated defense model using a federated learning mechanism. During this process, the blockchain simultaneously verifies the authenticity of the data, ensuring that all training inputs are rigorously screened and rejecting any unverified features.

[0072] Simultaneously, real-world attack and defense scenarios are simulated in a digital twin environment, and strategy parameters are optimized through deep reinforcement learning (DRL). The smart contract automatically adjusts sensitive operation thresholds, increasing defense strength in the event of new attacks. Each optimization result is recorded on the blockchain for evidence, forming a clear evolutionary trajectory for easy auditing and monitoring. Once the model update is complete, the smart contract automatically triggers a strategy synchronization procedure, ensuring that the updated defense strategy is encrypted and pushed to every terminal in the network within 24 hours. New devices also automatically receive the latest version of the smart contract library upon first connection, quickly integrating into the intelligent early warning system's defense ecosystem.

[0073] Optionally, through the above methods, the system can automatically generate a defense strategy library and pass the regulatory sandbox verification, thus meeting the proactive defense audit requirements of the financial industry.

[0074] As described above, by integrating homomorphic encryption, blockchain, and intelligent model update technologies, a secure, intelligent, and dynamic ATM fault information management and operation and maintenance guidance system has been constructed. This innovative mechanism not only ensures the secure transmission of fault information, avoiding the risks of data leakage and tampering, but also enhances the accuracy of fault prediction and the effectiveness of defense strategies through continuous model optimization. This significantly reduces operational interruptions and customer inconvenience caused by ATM faults, while also saving substantial human and material resources and improving the overall operational efficiency and security of the ATM network. Furthermore, the system facilitates rapid knowledge sharing and learning across branches, accelerating fault handling processes and providing strong technical support for the continuity of banking services and service quality.

[0075] In one optional embodiment, the intelligent early warning system acquires the associated information of the target ATM, wherein the associated information includes the operating status, operating age information, environmental information and historical fault information of each device in the target ATM, and then dynamically adjusts the threshold corresponding to each fault information in the threshold set based on the associated information.

[0076] Optionally, the system continuously collects the operating status of each device through a sensor network deployed on the ATM, such as the working time of the cash dispensing module and changes in the brightness of the display screen; calculates the actual operating age of each component based on the device's activation date and the current date, using this as an important indicator to assess the degree of aging; collects environmental parameters such as temperature and humidity at the location of the target ATM through built-in or external environmental monitoring devices, taking into account the impact of environmental factors on device performance; and integrates the ATM's historical maintenance records to analyze common failure modes and seasonal failure rates in order to predict potential future failures.

[0077] Optionally, big data analytics and machine learning algorithms, such as regression analysis and neural networks, can be used to deeply process the related information and extract the key factors affecting the accuracy of fault warnings. Based on the analysis results, the intelligent warning system adjusts the thresholds in the threshold set for each fault information in a customized manner, taking into account the operating age of different equipment, current environmental conditions, and historical fault tendencies. For example, the thresholds are relaxed for new equipment and tightened for older equipment.

[0078] As described above, through the aforementioned implementation methods, the intelligent early warning system can dynamically adjust the threshold set based on the associated information of the target ATM. This process significantly enhances the accuracy and applicability of fault early warning, reducing false alarms and missed alarms, especially when dealing with fault early warnings of aging equipment and complex environments. The personalized threshold setting also allows the system to respond more flexibly to different equipment states and environmental changes, improving the timeliness and effectiveness of early warnings, effectively reducing unplanned downtime caused by faults, and improving ATM operation and maintenance efficiency and customer satisfaction. More importantly, the introduction of the dynamic adjustment mechanism enables the intelligent early warning system to have self-learning and improvement capabilities. Over time, the system can gradually master more accurate early warning strategies, which has long-term significance for improving the overall equipment management level of the bank. Through continuous monitoring and intelligent analysis, this embodiment not only solves the problem of insufficient fault diagnosis accuracy in existing bank intelligent ATM operation and maintenance systems, but also opens up new avenues for the long-term operation and maintenance of ATMs.

[0079] In one optional embodiment, the intelligent early warning system receives feedback information obtained by maintenance personnel through augmented reality, which verifies the fault information determined by the digital twin model and the actual operating information of the target ATM. The feedback information is used to characterize the degree of deviation between the fault information and the actual operating information, and then the digital twin model is adjusted based on the feedback information.

[0080] Optionally, when the intelligent early warning system identifies a fault and generates fault information, it sends this information to the AR device of the personnel responsible for maintenance. Upon arriving on-site, maintenance personnel use the AR device to view the fault prediction provided by the digital twin model, including the specific location, type, and time of the fault. During maintenance, personnel compare the actual observed equipment status with the predicted information, inputting feedback through the interactive interface on the AR device, including the actual location and type of the fault, whether it matches the prediction, and any additional observations. This feedback accurately describes the deviation between the model's prediction and reality, providing data support for subsequent model improvements.

[0081] Optionally, upon receiving feedback from maintenance personnel, the intelligent early warning system automatically analyzes the degree of deviation and identifies shortcomings in the model's predictions. This may involve adjusting certain model parameters or optimizing algorithms for identifying certain fault characteristics. The system compares this feedback with existing multi-source data and trains the model using machine learning algorithms (such as deep neural networks or reinforcement learning) to gradually reduce prediction errors and improve the accuracy of fault location. Each adjustment is recorded in the system with a timestamp, forming a trajectory of model evolution for subsequent review and optimization.

[0082] As described above, by introducing AR technology into the fault verification process through the above implementation methods, combined with the professional knowledge and on-site observation of maintenance personnel, high-quality feedback is provided to the digital twin model, promoting continuous model optimization. Specifically, through AR verification, maintenance personnel can quickly and accurately compare predictions with actual results and promptly report deviation information. The intelligent early warning system then automatically adjusts model parameters based on this feedback, improving the accuracy and stability of fault prediction. This closed-loop model optimization mechanism not only enhances the reliability of fault diagnosis but also accelerates the system's response and learning capabilities to new types of faults, playing a crucial role in improving overall operational efficiency and reducing false alarm rates.

[0083] In one optional embodiment, the intelligent early warning system determines the risk level of the fault information based on the risk distribution heat map and external resource information, then determines the maintenance strategy corresponding to the risk level based on the risk level and the fault information, and finally generates early warning information for the target ATM based on the threshold corresponding to the risk level and the fault information.

[0084] Optionally, after the intelligent early warning system generates a risk distribution heatmap, it will further combine real-time external resource information, such as the current location and skill level of maintenance personnel, the type and quantity of spare parts in stock, real-time traffic conditions, and current weather conditions, to comprehensively analyze the risk level of each fault. This analysis process may involve multi-factor assessment models, such as rule-based expert systems or machine learning models, considering the impact of different factors on fault handling, and ultimately assigning a risk level, such as low risk, medium risk, or high risk, to each fault.

[0085] Optionally, the intelligent early warning system will automatically formulate maintenance strategies based on the risk level of each fault, combined with the specific type of fault and external resource information. For example, for low-risk faults, remote diagnostics and software reset may be prioritized; medium-risk faults may require on-site repair, and the system will dispatch the nearest maintenance personnel while ensuring sufficient spare parts; high-risk faults may involve the failure of core components, requiring immediate equipment replacement to minimize service interruption time. The system will also consider traffic conditions and weather factors to plan the optimal maintenance route and time window to minimize maintenance costs and time consumption.

[0086] Optionally, after comprehensively analyzing risk levels and fault information, the intelligent early warning system will check whether this information has reached a preset early warning threshold. The early warning threshold is set based on the equipment's operating status, historical fault records, and the current risk environment to determine whether fault information warrants an early warning and the urgency of the warning. For fault information that reaches the threshold, the system will immediately generate an early warning, including the fault type, location, risk level, and suggested initial response measures. The early warning information will be sent to relevant maintenance personnel and bank management personnel so that they can respond quickly and take appropriate measures.

[0087] As described above, the implementation steps, through risk assessment, maintenance strategy formulation, and early warning information generation by the intelligent early warning system, achieve refined management and efficient response to ATM malfunctions. The risk distribution heatmap provides an intuitive view of the malfunction risk distribution, comprehensive analysis of external information ensures accurate risk level assessment, and maintenance strategies based on risk levels and malfunction information achieve optimal resource allocation. The generation of early warning information not only improves the proactivity of maintenance work but also ensures the timeliness and effectiveness of malfunction handling.

[0088] This application also provides a fault warning device for ATMs. It should be noted that the fault warning device for ATMs in this application can be used to execute the fault warning method for ATMs provided in this application. The fault warning device for ATMs provided in this application will be described below.

[0089] According to an embodiment of this application, an apparatus for implementing the above-described fault warning method for ATMs is also provided. Figure 2 This is a schematic diagram of an optional fault warning device for an ATM according to an embodiment of this application, such as... Figure 2As shown, the device includes: a data acquisition unit 201, used to acquire multi-source data of the target ATM through acquisition equipment, wherein the multi-source data includes at least the mechanical vibration information, temperature distribution information, circuit load information, paper status information during the cash dispensing process, and target user behavior information of the target ATM; a construction unit 202, used to construct a digital twin model based on the physical structure information of the target ATM and the multi-source data, wherein the digital twin model is used to virtually represent the three-dimensional entity information of the target ATM; a determination unit 203, used to determine the fault information of the target ATM based on the digital twin model, wherein the fault information includes at least the fault location, fault type, and fault occurrence time of the target ATM; and a generation unit 204, used to generate early warning information for the target ATM based on the fault information of the target ATM and a threshold set, wherein the threshold set includes multiple thresholds, the threshold set is dynamically adjusted based on the real-time information of the target ATM, and each fault information corresponds to a threshold.

[0090] Optionally, the determining unit 203 includes: a first fusion subunit, a first extraction subunit, and a first determining subunit. The first fusion subunit is used to fuse the physical structure information of the target ATM with multi-source data to obtain target information of the target ATM; the first extraction subunit is used to extract at least one fault feature from the target information to obtain a feature set, wherein the fault feature is used to characterize features whose impact on the operating state of the target ATM is greater than a preset threshold; the first determining subunit is used to determine fault information in the target information based on a digital twin model, according to the target mode and the feature set, wherein the target mode is used to capture the correlation between different time scales in the data and identify information that deviates from the historical normal operating mode based on the correlation.

[0091] Optionally, the fault warning device for ATMs further includes a first generation unit and a first determination unit. The first generation unit is used to generate a risk distribution heatmap based on fault information, wherein the risk distribution heatmap includes at least fault information of the target ATM and predicted risk information within a preset time window; the first determination unit is used to determine a maintenance strategy based on the risk distribution heatmap and external resource information, wherein the external resource information includes at least maintenance personnel information, equipment inventory status, and traffic status information.

[0092] Optionally, the fault warning device for ATMs further includes: a first encryption unit, a first processing unit, a first update unit, and a second update unit. The first encryption unit encrypts fault information into a homomorphic encryption vector and transmits the homomorphic encryption vector to a target server via a blockchain, where the target server manages the information of each ATM in the blockchain. The first processing unit verifies the homomorphic encryption vector via the blockchain after the target server receives it, verifying its source and authenticity. The first update unit updates the target model based on the homomorphic encryption vector after verification, where the target model predicts risk information based on the fault information and generates defense strategies based on the risk information. The second update unit updates the maintenance strategy corresponding to each ATM in the blockchain based on the updated target model.

[0093] Optionally, the fault warning device for ATMs further includes a first acquisition unit and a first adjustment unit. The first acquisition unit is used to acquire associated information about the target ATM, including the operating status, operating age, environmental information, and historical fault information of each device in the target ATM. The first adjustment unit is used to dynamically adjust the threshold corresponding to each fault information in the threshold set based on the associated information.

[0094] Optionally, the fault warning device for ATMs further includes: a first receiving unit and a second adjustment unit. The first receiving unit receives feedback information obtained by maintenance personnel through augmented reality, verifying the fault information determined by the digital twin model against the actual operating information of the target ATM. The feedback information characterizes the degree of deviation between the fault information and the actual operating information. The second adjustment unit adjusts the digital twin model based on the feedback information.

[0095] Optionally, the fault warning device for ATMs further includes: a second determining unit, a third determining unit, and a second generating unit. The second determining unit is used to determine the risk level of the fault information based on a risk distribution heatmap and external resource information; the third determining unit is used to determine the maintenance strategy corresponding to the risk level based on the risk level and the fault information; and the second generating unit is used to generate warning information for the target ATM based on the threshold corresponding to the risk level and the fault information.

[0096] According to another aspect of this application, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored executable program, wherein, when the executable program is running, it controls the device where the computer-readable storage medium is located to perform the above-described fault warning method for ATMs.

[0097] According to another aspect of this application, an electronic device is also provided, including one or more processors and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to perform the above-described fault warning method for ATMs.

[0098] According to another aspect of the embodiments of this application, a computer program product is also provided, including computer instructions, which, when executed by a processor, implement the steps of the above-described fault warning method for ATMs.

[0099] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0100] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0101] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0102] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0103] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0104] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0105] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A fault early warning method for automated teller machines, characterized in that, include: Multi-source data of the target ATM is collected by the acquisition device. The multi-source data includes at least the mechanical vibration information, temperature distribution information, circuit load information, paper status information during the cash dispensing process, and the behavior information of the target user of the target ATM. A digital twin model is constructed based on the physical structure information of the target ATM and the multi-source data, wherein the digital twin model is used to virtually represent the three-dimensional entity information of the target ATM; The fault information of the target ATM is determined based on the digital twin model, wherein the fault information includes at least the fault location, fault type, and fault occurrence time of the target ATM; Early warning information for the target ATM is generated based on the fault information and threshold set of the target ATM. The threshold set includes multiple thresholds and is dynamically adjusted based on the real-time information of the target ATM. Each fault information corresponds to a threshold.

2. The fault early warning method for ATMs according to claim 1, characterized in that, The fault information of the target ATM is determined based on the digital twin model, including: The physical structure information of the target ATM and the multi-source data are fused to obtain the target information of the target ATM; At least one fault feature is extracted from the target information to obtain a feature set, wherein the fault feature is used to characterize the feature whose impact on the operating status of the target ATM is greater than a preset threshold. Based on the digital twin model, fault information in the target information is determined according to the target method and the feature set, wherein the target method is used to capture the correlation between different time scales in the data, and to identify information that deviates from the historical normal operation mode based on the correlation.

3. The fault early warning method for ATMs according to claim 2, characterized in that, After determining the fault information of the target ATM based on the digital twin model, the method further includes: A risk distribution heatmap is generated based on the fault information, wherein the risk distribution heatmap includes at least the fault information of the target ATM and the predicted risk information within a preset time window; The maintenance strategy is determined based on the risk distribution heatmap and external resource information, wherein the external resource information includes at least maintenance personnel information, equipment inventory status, and traffic status information.

4. The fault early warning method for ATMs according to claim 3, characterized in that, The method further includes: The fault information is encrypted into a homomorphic encryption vector, and the homomorphic encryption vector is transmitted to the target server through the blockchain. The target server is used to manage the information of each ATM in the blockchain. After the target server receives the homomorphic encryption vector, it verifies the homomorphic encryption vector through the blockchain, wherein the verification is used to verify the source and authenticity of the homomorphic encryption vector; After the homomorphic encryption vector passes the verification, the target model is updated based on the homomorphic encryption vector, wherein the target model is used to predict risk information based on the fault information and generate a defense strategy based on the risk information; Based on the updated target model, the maintenance strategy corresponding to each ATM in the blockchain is updated.

5. The fault early warning method for ATMs according to claim 1, characterized in that, The threshold set is obtained through the following steps: Obtain the association information of the target ATM, wherein the association information includes the operating status, operating age information, environmental information, and historical fault information of each device in the target ATM; The threshold corresponding to each fault information in the threshold set is dynamically adjusted based on the associated information.

6. The fault early warning method for ATMs according to claim 1, characterized in that, After generating warning information for the target ATM based on its fault information and threshold set, the method further includes: The system receives feedback information obtained by maintenance personnel through augmented reality to verify the fault information determined by the digital twin model and the actual operating information of the target ATM. The feedback information is used to characterize the degree of deviation between the fault information and the actual operating information. The digital twin model is adjusted based on the feedback information.

7. The fault early warning method for ATMs according to claim 3, characterized in that, The method further includes: The risk level of the fault information is determined based on the risk distribution heatmap and the external resource information; Determine the maintenance strategy corresponding to the risk level based on the risk level and the fault information; Early warning information is generated for the target ATM based on the risk level and the threshold corresponding to the fault information.

8. A fault early warning device for automated teller machines, characterized in that, include: The acquisition unit is used to acquire multi-source data of the target ATM through the acquisition device. The multi-source data includes at least the mechanical vibration information, temperature distribution information, circuit load information, paper status information during the cash dispensing process, and the behavior information of the target user of the target ATM. A construction unit is used to construct a digital twin model based on the physical structure information of the target ATM and the multi-source data, wherein the digital twin model is used to virtually represent the three-dimensional entity information of the target ATM; The determining unit is configured to determine the fault information of the target ATM based on the digital twin model, wherein the fault information includes at least the fault location, fault type, and fault occurrence time of the target ATM; The generation unit is used to generate early warning information for the target ATM based on the fault information and threshold set of the target ATM. The threshold set includes multiple thresholds and is dynamically adjusted based on the real-time information of the target ATM. Each fault information corresponds to a threshold.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, wherein when the computer program is executed, the device in which the computer-readable storage medium is located performs the fault warning method for ATMs as described in any one of claims 1 to 7.

10. An electronic device, characterized in that, It includes one or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to perform the fault warning method for ATMs according to any one of claims 1 to 7.

11. A computer program product comprising computer instructions, characterized in that, When the computer instructions are executed by the processor, they implement the steps of the fault warning method for ATMs as described in any one of claims 1 to 7.

Images