Platform and Method for Full-Cycle Monitoring and Compliance Verification of Digestive Endoscopy Disinfection Procedures

Abstract

The application discloses a digestive endoscope disinfection process whole-cycle monitoring and compliance verification platform and method, comprising a data acquisition and preprocessing module, which is used for acquiring real-time state data and environmental data and preprocessing to generate a standardized data set; a path optimization and decision module, which is used for simulating multiple operation paths of the disinfection process by using a Monte Carlo tree search algorithm and selecting an optimal disinfection path; a compliance reasoning and verification module, which is used for reasoning symbolic rules in the disinfection process; a blockchain storage module, which is used for writing data blocks into a blockchain ledger to form a disinfection data chain; an intelligent contract processing module, which is used for real-time judgment, automatic triggering of alarms and recording of non-compliant operations; a feedback and query module, which is used for querying, retrieving and auditing disinfection process whole-process information, alarm records and compliance events. The application realizes intelligent optimization of the disinfection process, real-time compliance verification and on-chain storage, and comprehensively improves the safety supervision efficiency.

Description

Technical Field

[0001] This invention relates to the field of medical device disinfection management and information security technology, and in particular to a platform and method for full-cycle monitoring and compliance verification of the disinfection process of digestive endoscopes. Background Technology

[0002] In the current healthcare industry, the disinfection management of high-risk medical devices such as digestive endoscopes is a crucial aspect of infection control and medical safety. With the increasing frequency of digestive endoscope use and the growing complexity of disinfection procedures, hospitals and medical institutions widely adopt automated disinfection equipment and information systems for process management. Existing technologies typically rely on local data recording by the disinfection equipment or centralized database-based information management to archive equipment operating status, disinfection parameters, and process results. While these management solutions improve process automation, the lack of tamper-proof data storage makes it difficult to achieve reliable data storage and transparent multi-department supervision throughout the entire process. Furthermore, their ability to assess compliance in real time is limited, leading to risks such as data loss, falsification, or inability to promptly trace responsibility during the disinfection process.

[0003] Furthermore, traditional disinfection process monitoring systems primarily rely on preset static rules for compliance assessments, lacking the ability to intelligently perceive and dynamically optimize changes in the disinfection environment, fluctuations in equipment parameters, and actual operational paths. This makes it difficult for the system to perform targeted risk assessments and operational optimizations when faced with complex operating conditions, operational differences, or multi-source data anomalies during actual disinfection processes. It also fails to provide intelligent, real-time, and multi-dimensional assessments of disinfection compliance. Particularly in multi-point, multi-stage collaborative scenarios, traditional systems struggle to efficiently address the needs for process optimization, compliance traceability, and cross-agency supervision, making it difficult to balance management efficiency and data security.

[0004] Existing technologies also generally suffer from information silos and inconsistent standard interfaces. Data between different disinfection equipment and information systems lacks interoperability and sharing, making it difficult to provide consistent and complete disinfection process data to multiple entities such as external regulatory platforms and medical quality control departments. At the same time, existing monitoring systems lack the ability to provide real-time early warnings and automatic accountability for abnormal disinfection behaviors, usually relying on manual intervention or post-event audits, which cannot effectively achieve closed-loop risk management throughout the entire process.

[0005] In summary, existing technologies for monitoring and verifying the disinfection process of digestive endoscopy are insufficient to meet the comprehensive requirements of the modern medical industry for full-cycle monitoring, intelligent compliance verification, full-process traceability, and trusted collaboration among multiple institutions.

[0006] Therefore, how to provide a platform for full-cycle monitoring and compliance verification of the disinfection process in digestive endoscopy is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0007] One objective of this invention is to propose a full-cycle monitoring and compliance verification platform for the disinfection process of digestive endoscopy. This invention fully integrates Monte Carlo tree search algorithm, neural symbolic reasoning technology, and blockchain evidence storage mechanism, and details the implementation methods for multi-path intelligent optimization of the disinfection process, real-time compliance judgment, and tamper-proof recording of data throughout the entire process. Through the full-process collection and intelligent analysis of real-time status data, operation paths, operation parameters, and compliance verification results of disinfection equipment, combined with smart contracts to achieve automated compliance early warning and traceability, the platform possesses advantages such as high intelligence, strong data security, accurate compliance judgment, support for multi-entity trusted collaboration, and full-process traceability, effectively improving the information management and risk control level of medical disinfection processes.

[0008] The digestive endoscope disinfection process full-cycle monitoring and compliance verification platform according to embodiments of the present invention includes:

[0009] The data acquisition and preprocessing module is used to collect real-time status data and environmental data, preprocess them, and generate standardized datasets.

[0010] The path optimization and decision module is used to simulate multiple operation paths of the disinfection process based on a standardized dataset using the Monte Carlo tree search algorithm, and to select the optimal disinfection path.

[0011] The compliance reasoning and verification module is used to reason about the symbol rules in the disinfection process based on the optimal disinfection path, and output the operation parameters and compliance verification results.

[0012] The blockchain evidence storage module is used to write the real-time status data, operation path, operation parameters and compliance verification results of the disinfection equipment into the blockchain ledger in blocks, forming an immutable disinfection data chain.

[0013] The smart contract processing module is used to make real-time judgments on the disinfection data chain, automatically trigger alarms, and record non-compliant operations;

[0014] The feedback and query module is used to query, retrieve, and audit information on the entire disinfection process, alarm records, and compliance events.

[0015] Optionally, modules can be integrated using the following methods:

[0016] S1. Collect real-time status data and environmental data of the disinfection equipment, and preprocess them to generate a standardized dataset;

[0017] S2. Based on a standardized dataset, the Monte Carlo tree search algorithm is used to simulate multiple operation paths of the disinfection process, and the optimal disinfection path is selected according to multiple variables in the disinfection process.

[0018] S3. Based on the optimal disinfection path, use neural symbolic reasoning to reason about the symbolic rules in the disinfection process, and combine relevant laws and regulations to verify the compliance of the disinfection process, and dynamically adjust the operating parameters in the disinfection process.

[0019] S4. Use blockchain technology to record the real-time status data of disinfection equipment, the operation path during the disinfection process, the operation parameters in the disinfection process, and the compliance verification results to generate an immutable disinfection data chain.

[0020] S5. Real-time verification of the disinfection data chain is performed through smart contracts, which automatically triggers alarms and records non-compliant operations. When the disinfection time, equipment temperature or other parameters do not meet the predetermined standards, the smart contract will automatically issue an alarm and record relevant information.

[0021] S6. Based on real-time verification results, generate detailed disinfection reports and promptly transmit rectification suggestions for non-compliant operations to operators through a feedback mechanism.

[0022] Optionally, the real-time status data of the disinfection equipment specifically includes the equipment's operating status, temperature, humidity, pressure, disinfection time, operator information, and equipment working status. The environmental data specifically includes the temperature, humidity, air quality, and lighting intensity in the disinfection environment.

[0023] Optionally, the preprocessing specifically includes cleaning, standardizing, filling missing values, detecting outliers, and converting data formats for the real-time status data and environmental data of the disinfection equipment.

[0024] Optionally, S2 specifically includes:

[0025] S21. Based on the generated standardized dataset, the disinfection process is divided into multiple independent stages, and each stage is defined as a hierarchical state space. This includes pretreatment, primary disinfection, drying, testing, and storage stages;

[0026] S22, in each hierarchical state space Internally, based on the operation nodes, state transitions, and operation parameters of each stage in the disinfection process, a Monte Carlo tree search subtree is independently constructed for each layer to simulate and optimize the path.

[0027] S23. During the node expansion process of each layered subtree, based on the pre-set disinfection process compliance standards, including disinfection equipment operation specifications, disinfection time thresholds, temperature and pressure, and safety rules, each expandable action is judged according to rules. Only actions that meet all standards are expanded to form a constraint-aware adaptive tree structure. The safety rules include equipment operation safety parameters and operator identity verification.

[0028] S24. During the expansion process at each node, risk assessment factors for the disinfection process are dynamically generated through real-time monitoring of disinfection status data. Adjust path rewards accordingly;

[0029] S25. Introduce an adaptive reward mechanism based on real-time environmental feedback. This mechanism adjusts the path reward function in real-time using monitoring data, dynamically adjusting the priority of path selection and the reward function accordingly. ;

[0030] S26. During the simulation, combining real-time status data of the disinfection equipment, environmental data, and risk assessment of the current disinfection path, an adaptive policy network dynamically guides the selection of the simulation path. The adaptive policy network includes:

[0031] The input layer receives real-time status data from the disinfection equipment and environmental data.

[0032] The hidden layer uses a multilayer sensor to extract features and process information from the real-time status data and environmental data of the disinfection equipment.

[0033] The output layer uses the Softmax function to generate the selection probability of each operation path and determine the priority of the path.

[0034] The adaptive strategy network is trained based on the historical operating status of disinfection equipment, disinfection time, disinfection effect and monitoring data. It continuously adjusts the ratio of exploration and utilization. By setting thresholds for compliance, risk and efficiency, it prioritizes operation sequences that meet preset compliance standards, risk control requirements and efficiency higher than the set thresholds, and optimizes path selection in the disinfection process. The adaptive strategy network responds flexibly and adjusts decision-making strategies according to real-time changes in the disinfection environment.

[0035] S27. During the path expansion and simulation process, based on the disinfection process compliance standards and adaptive reward mechanism, prune paths that do not meet the compliance standards, violate the timing or safety requirements, or whose path rewards do not reach the set threshold, and terminate the path expansion and simulation in advance.

[0036] S28. By using parallel computing technology, different hierarchical state spaces are... Simulation tasks with different operation paths can be processed in a multi-threaded or distributed manner.

[0037] S29. In each simulation and backtracking process, the reward value of each layer path is accumulated, and the optimal decision parameters of the current layer are passed to the state space of the next layer to realize the local optimal decision of each layer and the global optimal path selection of the entire disinfection process.

[0038] S210. Repeat steps S21 to S29 until the preset number of simulations or computational resource threshold is reached. Finally, combine the optimal paths at all levels to output the optimal disinfection path for the entire disinfection process, which serves as the basis for optimizing the disinfection process.

[0039] Optionally, S3 specifically includes:

[0040] S31. Based on the generated optimal disinfection path, extract the operation parameters, disinfection time, temperature, and humidity information during the disinfection process, and combine them with relevant regulations and standards. To jointly build a symbolic rule base Where n is the total number of symbol rules, The number of relevant laws and regulations;

[0041] S32. Construct a neural symbolic reasoning model, specifically including:

[0042] Input layer, receiving symbolic rule base This includes relevant regulations and standards. Real-time status data and operation sequence of disinfection equipment ;

[0043] The neural network layer performs feature extraction and high-dimensional representation learning on the data received from the input layer, and outputs a high-dimensional feature representation.

[0044] The symbol rule encoding module encodes the symbol rules and regulations in the input layer. Perform vectorized encoding to obtain rule embeddings;

[0045] The inference mechanism integration module embeds the high-dimensional feature representation output from the neural network layer and the rules output from the symbolic rule encoding module into the symbolic inference engine. Through joint inference using logical rules, it outputs a local compliance judgment value. ;

[0046] Output layer, outputs a vector of local compliance judgment values. , This indicates the local compliance judgment value. ;

[0047] S33, Based on the obtained local compliance judgment value vector By combining the real-time status data of the disinfection equipment received by the input layer, the high-dimensional feature representation output by the neural network layer, the rule embedding output by the symbolic rule encoding module, and the reasoning results of the reasoning mechanism integration module, each local compliance judgment value is compared with relevant laws and regulations. Compare and combine with the disinfection operation sequence. Based on time-series dependencies, risk assessments, and real-time environmental feedback data, a time-series risk-weighted global compliance score is calculated. ;

[0048] S34. For non-compliant or high-risk paths, based on the local compliance judgment value output by the inference mechanism integration module, a set of corrective action suggestions is automatically generated, and the operation parameters in the disinfection path are adjusted. Each corrective action suggestion corresponds to a specific correction action.

[0049] S35. Feed back the set of corrective action suggestions, update the disinfection path operation parameters, and modify the disinfection path operation parameters according to the corrective action suggestions based on the current disinfection path parameters.

[0050] S36. For paths deemed compliant, continuously call the integrated module of the neural network layer and inference mechanism to optimize and adjust the disinfection parameters through the policy function. Output the optimal parameter set ;

[0051] S37. During the path optimization process, the symbolic rule encoding module and the inference mechanism integration module are used to automatically update the symbolic rule base based on the real-time status data of the disinfection equipment and environmental feedback, generating an updated symbolic rule base. ;

[0052] S38. Train the neural symbolic reasoning model through reinforcement learning mechanism to optimize disinfection path selection and symbolic rule generation, and apply a compliance loss function. To minimize compliance losses;

[0053] S39. Input the updated symbolic rule base and operation parameters to re-simulate the path, ensuring that the disinfection process is always compliant, efficient and safe. Through a cyclical feedback mechanism, continuously monitor and optimize the disinfection path.

[0054] Optionally, S4 specifically includes:

[0055] S41. Collect real-time status data of disinfection equipment, operation paths during the disinfection process, dynamically adjusted disinfection process operation parameters, and global compliance score. and the updated symbolic rule base Organized into a dataset ;

[0056] S42. Organize the dataset according to the time sequence of the disinfection process and the unique identifier of the equipment. The data is divided into blocks, each containing a timestamp, device number, operation path, disinfection process operation parameters, and global compliance score. and the updated symbolic rule base ;

[0057] S43. Perform a hash operation on each data block to obtain the block hash value, where the hash input is the content of the current data block and the hash value of the previous block, and the hash algorithm used is consistent with the blockchain ledger.

[0058] S44 will include hash value, timestamp, operation path, disinfection process operation parameters, and global compliance score. and the updated symbolic rule base Data blocks are written sequentially to the blockchain ledger in block form, forming a disinfection data chain. ;

[0059] S45. Authentication of each data block is performed using a digital signature. The digital signature corresponds one-to-one with the node identity to ensure that data writing is non-repudiable and operations are traceable. The signature content includes the block hash value and the node identifier.

[0060] S46. Smart contracts based on blockchain ledgers automatically detect the global compliance score in each data block. With relevant laws and regulations The comparison relationship shows that if the score is lower than the set compliance threshold, it will be automatically marked as a non-compliant event and recorded on the chain.

[0061] S47. For all compliant and non-compliant events, record the disinfection process data and parameter adjustment information in detail to the blockchain ledger to achieve tamper-proof evidence storage and full-process traceability auditing.

[0062] S48. When the blockchain ledger receives a data query, access, or anomaly alarm request, it retrieves disinfection path parameters, equipment status, operating parameters, and global compliance score based on the ledger content. Symbolic rule base And compliance incidents, supporting full-process monitoring and compliance verification of data transparency and traceability;

[0063] S49. Periodically perform integrity checks on the blockchain ledger and verify the consistency of blocks through the hash chain structure.

[0064] Optionally, S5 specifically includes:

[0065] S51. The disinfection time, equipment temperature and other key operating parameters are compared with the preset standards in real time, and the judgment is made according to the compliance rules set by the smart contract.

[0066] S52. When the smart contract detects that the disinfection time, equipment temperature or other key parameters do not meet the predetermined standards, it will immediately and automatically determine that it is a non-compliant operation.

[0067] S53. For each non-compliant operation, the smart contract automatically triggers the early warning mechanism, generates alarm information, and sends it to the relevant responsible personnel or management terminal in real time.

[0068] S54. Automatically record and write all non-compliant operation details, including operation time, device number, abnormal parameters, and specific values ​​deviating from the standard, into the blockchain ledger.

[0069] S55. Supports querying, retrieving, and auditing alarm information and non-compliant operations recorded in the blockchain ledger, facilitating the tracing of abnormal situations in the disinfection process and promoting rectification.

[0070] The beneficial effects of this invention are:

[0071] This invention deeply integrates the Monte Carlo tree search algorithm with neural symbolic reasoning to achieve intelligent simulation, optimization, and dynamic adjustment of multi-path disinfection processes. It can automatically generate optimal operation paths and parameters for different disinfection conditions and equipment states, significantly improving the adaptability and intelligent decision-making level of the disinfection process. By using neural symbolic reasoning combined with regulatory standards to perform real-time compliance judgments on each step of the operation, it achieves multi-dimensional, end-to-end intelligent compliance verification, effectively avoiding the problem of untimely identification of anomalies and edge cases under traditional static rule systems. Simultaneously, this invention uses blockchain technology to write the real-time status data of disinfection equipment, operation paths, operation parameters, and compliance verification results into the blockchain ledger throughout the entire process, ensuring the immutability of disinfection process data, full lifecycle traceability, and trusted collaboration among multiple institutions, greatly improving data security and accountability.

[0072] This invention also utilizes a smart contract mechanism for real-time automated compliance assessment and alerts. It can trigger warnings and detailed recordings immediately upon the occurrence of non-compliant behavior in the disinfection process, significantly improving the efficiency and standardization of abnormal risk response. The platform supports dynamic querying and auditing of data throughout the entire disinfection process, facilitating quality control, supervision, and multi-departmental collaboration. It achieves an integrated and innovative solution encompassing disinfection process optimization, intelligent compliance assessment, secure data storage throughout the entire process, and cross-institutional collaborative management. Compared to existing technologies, this invention significantly enhances the intelligence, standardization, and information security of medical disinfection process management, providing more reliable technical support for infection control and quality management in medical institutions. Attached Figure Description

[0073] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0074] Figure 1 This is a flowchart of the method for full-cycle monitoring and compliance verification of the disinfection process of digestive endoscopes proposed in this invention;

[0075] Figure 2 This is a flowchart illustrating the blockchain data chain generation and automatic storage of compliance events for the whole-cycle monitoring and compliance verification method for the disinfection process of digestive endoscopy proposed in this invention. Detailed Implementation

[0076] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0077] refer to Figure 1 and Figure 2 A platform for full-cycle monitoring and compliance verification of the disinfection process in digestive endoscopy, including:

[0078] The data acquisition and preprocessing module is used to collect real-time status data and environmental data, preprocess them, and generate standardized datasets.

[0079] The path optimization and decision module is used to simulate multiple operation paths of the disinfection process based on a standardized dataset using the Monte Carlo tree search algorithm, and to select the optimal disinfection path.

[0080] The compliance reasoning and verification module is used to reason about the symbol rules in the disinfection process based on the optimal disinfection path, and output the operation parameters and compliance verification results.

[0081] The blockchain evidence storage module is used to write the real-time status data, operation path, operation parameters and compliance verification results of the disinfection equipment into the blockchain ledger in blocks, forming an immutable disinfection data chain.

[0082] The smart contract processing module is used to make real-time judgments on the disinfection data chain, automatically trigger alarms, and record non-compliant operations;

[0083] The feedback and query module is used to query, retrieve, and audit information on the entire disinfection process, alarm records, and compliance events.

[0084] In this embodiment, the modules are interconnected using the following method:

[0085] S1. Collect real-time status data and environmental data of the disinfection equipment, and preprocess them to generate a standardized dataset;

[0086] S2. Based on a standardized dataset, the Monte Carlo tree search algorithm is used to simulate multiple operation paths of the disinfection process, and the optimal disinfection path is selected according to multiple variables in the disinfection process.

[0087] S3. Based on the optimal disinfection path, use neural symbolic reasoning to reason about the symbolic rules in the disinfection process, and combine relevant laws and regulations to verify the compliance of the disinfection process, and dynamically adjust the operating parameters in the disinfection process.

[0088] S4. Use blockchain technology to record the real-time status data of disinfection equipment, the operation path during the disinfection process, the operation parameters in the disinfection process, and the compliance verification results to generate an immutable disinfection data chain.

[0089] S5. Real-time verification of the disinfection data chain is performed through smart contracts, which automatically triggers alarms and records non-compliant operations. When the disinfection time, equipment temperature or other parameters do not meet the predetermined standards, the smart contract will automatically issue an alarm and record relevant information.

[0090] S6. Based on real-time verification results, generate detailed disinfection reports and promptly transmit rectification suggestions for non-compliant operations to operators through a feedback mechanism.

[0091] This invention achieves high-quality and consistent input data for the disinfection process by comprehensively collecting and standardizing real-time status data of disinfection equipment and environmental data, thereby improving the reliability of system decision-making. Utilizing the Monte Carlo tree search algorithm to simulate and optimize the multivariate disinfection process significantly enhances the intelligence and adaptability of process optimization. Combining neural symbolic reasoning and regulatory standards, intelligent compliance reasoning and dynamic parameter adjustment are performed on disinfection operations, enabling the disinfection process to flexibly adapt to actual conditions and making compliance judgment more accurate and efficient. Blockchain technology is used to record key data and compliance results throughout the entire process, ensuring the immutability and full traceability of the disinfection data chain, improving data security and transparency of responsibility attribution. Smart contracts enable automatic real-time verification of compliance and anomaly alerts, greatly improving the speed and standardization of risk response. Non-compliant operations can be identified promptly and rectification suggestions can be generated, effectively supporting operators in closed-loop process optimization. Overall, this invention significantly improves the informatization, intelligence, and compliance management capabilities of medical disinfection processes, enhancing medical safety and regulatory efficiency.

[0092] In this embodiment, the real-time status data of the disinfection equipment specifically includes the equipment's operating status, temperature, humidity, pressure, disinfection time, operator information, and equipment working status. The environmental data specifically includes the temperature, humidity, air quality, and lighting intensity in the disinfection environment.

[0093] In this embodiment, the preprocessing specifically includes cleaning, standardizing, filling in missing values, detecting outliers, and converting the data format of the real-time status data of the disinfection equipment and the environmental data. Specifically, cleaning, standardizing, filling in missing values, detecting outliers, and converting the data format refer to cleaning the collected real-time status data of the disinfection equipment and the environmental data, removing duplicate, invalid, or logically contradictory data entries; then, standardization is performed to normalize various types of data according to a unified dimension, facilitating subsequent model processing; for missing data, mean imputation, interpolation, or model prediction methods are used to fill in missing values; for outlier detection, Z-score, box plot, or sliding window methods are used to identify and mark data that significantly deviates from the normal range; finally, the processed data is uniformly converted into a structured format acceptable to the model, including time series reconstruction, field renaming, and encoding format standardization.

[0094] In this embodiment, S2 specifically includes:

[0095] S21. Based on the generated standardized dataset, the disinfection process is divided into multiple independent stages, and each stage is defined as a hierarchical state space. This includes pretreatment, primary disinfection, drying, testing, and storage stages;

[0096] S22, in each hierarchical state space Within the disinfection process, based on the operation nodes, state transitions, and operation parameters of each stage, a Monte Carlo tree search subtree is independently constructed for each level. Path simulation and optimization are then performed. Specifically, this involves constructing independent Monte Carlo tree search subtrees for each stage of the disinfection process. Within each subtree, state transition paths between different operation nodes are simulated, and possible operation sequences are progressively expanded and evaluated based on the operation parameters of the current stage. By executing the simulation process in each subtree, the performance of each path in terms of efficiency, risk, resource consumption, and compliance is predicted. Simultaneously, the search strategy is continuously adjusted according to a set reward mechanism to select the path with the optimal overall benefit. Finally, the optimal operation decision sequence for each stage is output.

[0097] S23. During the node expansion process of each hierarchical subtree, based on the pre-set disinfection process compliance standards, including disinfection equipment operation specifications, disinfection time thresholds, temperature and pressure, and safety rules, each expandable action is judged according to rules. Only actions that meet all standards are expanded, forming a constraint-aware adaptive tree structure. The safety rules include equipment operation safety parameters and operator identity verification. Specifically, the rule judgment for each expandable action refers to verifying each operation to be executed item by item according to the disinfection process compliance standards during the Monte Carlo tree search expansion process. This includes verifying whether the action complies with the equipment operation specifications, whether it is within the specified time, temperature, and pressure range, whether it meets the equipment operation safety parameter requirements, and whether it is initiated by an operator with legitimate authority. Only when all rules are met is the action allowed to be expanded into the search tree.

[0098] S24. During the expansion process at each node, risk assessment factors for the disinfection process are dynamically generated through real-time monitoring of disinfection status data. The process involves path reward correction. Specifically, during the expansion of each node in the Monte Carlo tree search, based on real-time collected disinfection status data, the potential risk factors of the current operation path are assessed, and this risk level is quantified as a risk assessment factor. Subsequently, this factor is introduced into the path's reward function to adjust the original simulation reward, reducing the reward value of high-risk paths and increasing the priority of low-risk paths. This guides the search algorithm to prioritize disinfection paths with high safety and stability, thereby optimizing the overall process decision quality.

[0099] S25. Introduce an adaptive reward mechanism based on real-time environmental feedback. This mechanism adjusts the path reward function in real-time using monitoring data, dynamically adjusting the priority of path selection and the reward function accordingly. :

[0100] ;

[0101] in, For simulation rewards, Rewards are based on risk assessments generated from real-time monitoring data. As a feedback reward based on the disinfection status, , , As a weighting factor;

[0102] reward function The practical significance of the formula lies in incorporating the dynamic environment and real-time risk feedback during the disinfection process optimization into the path reward function, achieving intelligent adaptive path selection. Traditional Monte Carlo tree search algorithms, when optimizing paths, primarily rely on preset rewards or historical data for decision-making, making it difficult to cope with changes in the operating environment of disinfection equipment, sudden situations, or external risk factors. This invention, however, introduces risk assessment rewards and disinfection status feedback rewards generated based on real-time monitoring data, assigning different weight factors to them, enabling the reward function to reflect equipment status, environmental changes, and risk levels in real time. This mechanism not only improves the flexibility of path simulation but also prioritizes disinfection paths with low risk, high compliance, and high efficiency during the optimization process, thereby enhancing the overall safety and adaptability of the process. Through dynamic adjustment of the reward function, the system can continuously integrate environmental changes and equipment feedback during disinfection process simulation and decision-making, forming a closed-loop optimization decision and ultimately outputting the optimal disinfection path that best meets the requirements of actual working conditions. This innovative mechanism greatly enhances the intelligence and robustness of disinfection process optimization, ensuring the standardization of disinfection operations and medical safety.

[0103] S26. During the simulation, combining real-time status data of the disinfection equipment, environmental data, and risk assessment of the current disinfection path, an adaptive policy network dynamically guides the selection of the simulation path. The adaptive policy network includes:

[0104] The input layer receives real-time status data from the disinfection equipment and environmental data.

[0105] The hidden layer uses a multilayer perceptron to extract features and process information from the real-time status data and environmental data of the disinfection equipment. Specifically, this feature extraction and processing involves using the multilayer perceptron structure to perform nonlinear mapping and dimensionality reduction on the input real-time status data and environmental data of the disinfection equipment, automatically mining potential high-order correlation features. This process includes weighted calculation of the original input, activation function transformation, and multilayer information fusion to extract key feature representations that reflect disinfection effectiveness, safety risks, or operational anomalies.

[0106] The output layer uses the Softmax function to generate the selection probability of each operation path and determine the priority of the path.

[0107] The adaptive strategy network is trained based on the historical operating status of disinfection equipment, disinfection time, disinfection effect and monitoring data. It continuously adjusts the ratio of exploration and utilization. By setting thresholds for compliance, risk and efficiency, it prioritizes operation sequences that meet preset compliance standards, risk control requirements and efficiency higher than the set thresholds, and optimizes path selection in the disinfection process. The adaptive strategy network responds flexibly and adjusts decision-making strategies according to real-time changes in the disinfection environment.

[0108] S27. During the path expansion and simulation process, based on the disinfection process compliance standards and adaptive reward mechanism, prune paths that do not meet the compliance standards, violate the timing or safety requirements, or whose path rewards do not reach the set threshold, and terminate the path expansion and simulation in advance.

[0109] S28. By using parallel computing technology, different hierarchical state spaces are... Simulation tasks with different operation paths are processed in a multi-threaded or distributed manner. Specifically, when executing the disinfection process path simulation task, a parallel computing framework is used to distribute the simulation subtasks corresponding to each hierarchical state space and different operation paths under the same state space to multiple computing threads or computing nodes for parallel execution, which significantly improves the overall simulation efficiency. Specifically, multi-threaded processing enables multiple processing threads to run multiple path simulation tasks synchronously within a single computing node; while distributed processing divides the tasks and assigns them to multiple physical nodes or virtual computing resources, and uses a task scheduling mechanism to coordinate concurrent execution, ensuring that while maintaining the integrity and consistency of the simulation, the path search and optimization time is shortened, meeting the real-time requirements of large-scale disinfection path calculation.

[0110] S29. In each simulation and backtracking process, the reward value of each layer path is accumulated, and the optimal decision parameters of the current layer are passed to the state space of the next layer to realize the local optimal decision of each layer and the global optimal path selection of the entire disinfection process.

[0111] S210. Repeat steps S21 to S29 until the preset number of simulations or computational resource threshold is reached. Finally, combine the optimal paths at all levels to output the optimal disinfection path for the entire disinfection process, which serves as the basis for optimizing the disinfection process.

[0112] This invention divides the disinfection process into multiple independent stages and employs a hierarchical state space and hierarchical Monte Carlo tree search strategy to achieve refined modeling and intelligent path optimization of the entire disinfection process. Each stage independently constructs a Monte Carlo tree subtree, combining operation nodes, parameters, and state transition mechanisms to make path simulation and optimization more flexible and controllable. By introducing compliance standards, operational safety, and personnel identity verification into node expansion, compliance and safety at each step are effectively ensured. Real-time monitoring data and environmental feedback dynamically participate in path rewards and risk assessments, forming an adaptive reward mechanism that not only improves the rationality of path selection but also allows for agile responses to changes in the environment and operating conditions. An adaptive strategy network is used to continuously optimize the simulation strategy based on real-time equipment status and historical operating data, automatically adjusting the exploration and utilization ratio to ensure priority is given to compliant, risk-controllable, and efficient operation paths. The introduction of pruning and parallel simulation techniques effectively filters out non-compliant or low-priority paths and improves computational efficiency, promoting the efficient output of the globally optimal disinfection path. Overall, this invention significantly improves the intelligence, real-time nature, and globality of disinfection process optimization, enhancing the safety assurance and quality control capabilities of the disinfection process.

[0113] In this embodiment, S3 specifically includes:

[0114] S31. Based on the generated optimal disinfection path, extract the operation parameters, disinfection time, temperature, and humidity information during the disinfection process, and combine them with relevant regulations and standards. To jointly build a symbolic rule base Where n is the total number of symbol rules, The number of relevant laws and regulations;

[0115] S32. Construct a neural symbolic reasoning model, specifically including:

[0116] Input layer, receiving symbolic rule base This includes relevant regulations and standards. Real-time status data and operation sequence of disinfection equipment ;

[0117] The neural network layer performs feature extraction and high-dimensional representation learning on the data received from the input layer, outputting a high-dimensional feature representation. Specifically, the feature extraction and high-dimensional representation learning refer to the neural network layer using a multi-layered neuron structure to perform nonlinear mapping and pattern recognition on the real-time status data, operation sequences, symbol rules, and other information received from the input layer of the disinfection equipment. It extracts representative key features and converts them into vector representations in a high-dimensional space. This process includes convolution operations, activation function transformations, normalization processing, and inter-layer weight connections. Through multi-layer perception and training optimization, the original input data possesses stronger separability and expressive power in the high-dimensional space.

[0118] The symbol rule encoding module encodes the symbol rules and regulations in the input layer. Vectorization encoding is performed to obtain rule embeddings. Specifically, the symbolic rule encoding module converts the symbolic rules in the input layer and the structured or semi-structured text information in the regulations and standards into numerical vectors through a preset symbolic embedding method. This process includes operations such as symbolic word segmentation, semantic feature extraction, keyword tagging, and rule structure analysis. Combined with encoding techniques such as One-Hot encoding, word vectors, and structural embedding, symbolic semantics are mapped into fixed-dimensional vector representations, thereby realizing the numerical expression of semantic relationships and logical structures between rules, forming a "rule embedding" that can be fused with neural network outputs for reasoning.

[0119] The inference mechanism integration module embeds the high-dimensional feature representation output from the neural network layer and the rules output from the symbolic rule encoding module into the symbolic inference engine. Through joint inference using logical rules, it outputs a local compliance judgment value. The joint reasoning through logical rules specifically refers to the reasoning mechanism integration module taking the high-dimensional feature representation output by the neural network layer and the rule embedding vector generated by the symbolic rule encoding module as joint inputs and feeding them into the symbolic reasoning engine. Based on the preset formal logical rules, semantic matching and constraint verification are performed to judge the compliance of each operation step or state node. This reasoning process not only relies on the data features learned by the deep model, but also on the symbolic rules explicitly expressed in the regulations and standards to achieve the integration of "data-driven" and "knowledge-driven". The reasoning engine will compare the identified state with the rule conditions, and combine the operation sequence, parameter range, and state transition to evaluate whether the operation meets the compliance requirements one by one, and finally output a local compliance judgment value.

[0120] Output layer, outputs a vector of local compliance judgment values. , This indicates the local compliance judgment value. ;

[0121] S33, Based on the obtained local compliance judgment value vector By combining the real-time status data of the disinfection equipment received by the input layer, the high-dimensional feature representation output by the neural network layer, the rule embedding output by the symbolic rule encoding module, and the reasoning results of the reasoning mechanism integration module, each local compliance judgment value is compared with relevant laws and regulations. Compare and combine with the disinfection operation sequence. Based on time-series dependencies, risk assessments, and real-time environmental feedback data, a time-series risk-weighted global compliance score is calculated. :

[0122] ;

[0123] in, For local decision items, For the first Step risk factor, For environmental feedback weighting coefficients, This is a correction value based on environmental feedback. To provide real-time data feedback, These are time-dependent weighting coefficients. The result represents the temporal dependency determination of the operation sequence, where H is the total number of local compliance determination items. For the matching indicator function of rules and regulations, if Failed to meet relevant regulatory standards Corresponding compliance threshold If it does not, it is deemed non-compliant; otherwise, it is compliant and a correction mechanism is triggered.

[0124] Global compliance score The practical significance of the formula lies in its ability to comprehensively assess the compliance of each step in the disinfection process under different risk environments through a global compliance scoring mechanism weighted by time-series risk. This formula not only considers the compliance judgment results of each local operation but also introduces multi-dimensional dynamic factors such as risk coefficients, environmental feedback, and time-series dependence, incorporating the real-time status of the disinfection operation, the external environment, and regulatory standards into a unified evaluation system. By assigning weights to each operational node and combining them with actual risks, the system can accurately reflect the importance of certain high-risk or critical links in the overall process. Simultaneously, through environmental feedback correction values ​​and time-series dependence weighting, it effectively adapts to changes in operating conditions and process complexity during disinfection. The introduction of a matching indicator function ensures that only operations that truly comply with relevant regulatory standards are recognized and included in the total compliance score. Ultimately, the global compliance score not only serves as the basis for determining whether the process is compliant but also provides quantitative support for subsequent corrective action recommendations and path optimization. This mechanism significantly improves the scientific rigor and sensitivity of compliance judgments, providing a solid decision-making foundation for intelligent supervision, risk warning, and continuous optimization of medical disinfection processes.

[0125] S34. For non-compliant or high-risk paths, based on the local compliance judgment value output by the inference mechanism integration module, a set of corrective action suggestions is automatically generated, and the operation parameters in the disinfection path are adjusted. Each corrective action suggestion corresponds to a specific correction action.

[0126] S35. Feed back the set of corrective action suggestions, update the disinfection path operation parameters, and modify the disinfection path operation parameters according to the corrective action suggestions based on the current disinfection path parameters.

[0127] S36. For paths deemed compliant, continuously call the integrated module of the neural network layer and inference mechanism to optimize and adjust the disinfection parameters through the policy function. Output the optimal parameter set :

[0128] ;

[0129] in, For path The cost of disinfection This is a set of operational parameters for candidate disinfection paths. This represents the parameter corresponding to the j-th operation step in the disinfection path. This indicates whether the parameters of the j-th operation step meet the compliance requirements; a value of 1 indicates compliance, and a value of 0 indicates non-compliance. Let P represent the set of parameters that minimizes the total cost, and X represent the number of disinfection steps.

[0130] Optimal parameter set The practical significance of this formula lies in providing a scientific and quantitative basis for parameter setting and path optimization in disinfection processes, ensuring optimal resource utilization while meeting compliance requirements. The formula uses all possible combinations of operational parameters as a candidate set, combines the compliance judgment results of each step's parameters, filters out all compliant parameter configurations, and performs a weighted cumulative calculation of disinfection costs for these compliant paths. Through the objective of "minimizing total cost," the system can automatically select the optimal parameter combination that both complies with regulatory standards and reduces disinfection time, energy consumption, or resource consumption. A compliance indicator function ensures that no non-compliant solutions are encountered during the optimization process, improving the standardization and reliability of decision-making. This mechanism not only adapts to the dynamic needs of different disinfection equipment and environments but can also be flexibly adjusted based on actual operating costs, thereby achieving an organic combination of process optimization and compliance supervision. Overall, the formula effectively improves the economy, scientific rigor, and intelligence of disinfection processes, promoting efficient operation and quality improvement in medical institutions while ensuring safety and compliance.

[0131] S37. During the path optimization process, the symbolic rule encoding module and the inference mechanism integration module are used to automatically update the symbolic rule base based on the real-time status data of the disinfection equipment and environmental feedback, generating an updated symbolic rule base. ;

[0132] S38. Train the neural symbolic reasoning model through reinforcement learning mechanism to optimize disinfection path selection and symbolic rule generation, and apply a compliance loss function. Minimize compliance losses:

[0133] ;

[0134] in, For compliance penalties, The total number of training samples, Indicates the first The global compliance score corresponding to each training sample. The preset compliance threshold, For indicator functions, when the first The compliance score of each sample is below the compliance threshold. When the value is 1, it indicates that the sample is non-compliant; otherwise, it is 0.

[0135] Compliance loss function The practical significance of this formula lies in introducing a compliance-oriented automatic learning mechanism into the training and disinfection path optimization process of the neural symbolic reasoning model. This enables the model to continuously improve its ability to identify and avoid compliance risks in the disinfection process through continuous training iterations. The loss function automatically identifies non-compliant samples by comparing the global compliance score of all training samples with a preset threshold and penalizes the degree of violation, thereby guiding the model to continuously reduce the probability of non-compliant paths occurring during the learning process. The introduction of the penalty mechanism prompts the model to pay more attention to disinfection paths that are prone to risks or violations of regulations, improving the model's sensitivity and corrective ability to compliance issues under complex operating conditions. This mechanism not only enhances the model's generalization ability and adaptability but also achieves a deep integration of optimized path selection and strengthened compliance management. By minimizing compliance losses, the system can continuously output disinfection path solutions that meet actual regulatory requirements and safety standards under dynamic environments and changing rules, providing solid data and model support for intelligent management and risk prevention of medical disinfection processes.

[0136] S39. Input the updated symbolic rule base and operation parameters to re-simulate the path, ensuring that the disinfection process is always compliant, efficient and safe. Through a cyclical feedback mechanism, continuously monitor and optimize the disinfection path.

[0137] This invention constructs a symbolic rule base based on regulations, standards, and operational parameters, combined with a neural symbolic reasoning model, to achieve intelligent compliance determination and dynamic optimization of each operational step in the disinfection process. Through real-time status data, operational sequences, and regulations collected at the input layer, the model can perform deep feature extraction and encoding of multi-dimensional information, and output compliance determination results for operational steps through logical reasoning. Under a time-series risk weighting mechanism, the system comprehensively considers time-series dependencies, risk factors, and environmental feedback, achieving global compliance scoring and risk quantification throughout the disinfection process, accurately identifying non-compliant or high-risk operations. Corrective suggestions are automatically generated and fed back for non-compliant paths, prompting dynamic adjustments to process parameters, achieving real-time closed-loop operation and proactive risk intervention. For compliant paths, the system continuously optimizes disinfection parameters, further improving resource utilization efficiency through cost minimization strategies. A reinforcement learning mechanism endows the model with adaptive evolution capabilities, continuously improving the generalization level of path optimization and compliance judgment. The symbolic rule base can be dynamically updated with the environment and equipment status, ensuring that the decision-making system is always highly matched to the actual scenario. Overall, this invention significantly improves the intelligent compliance judgment capability, process optimization efficiency, and risk control level of the disinfection process, providing precise, efficient, and sustainable technical support for medical disinfection management.

[0138] In this embodiment, S4 specifically includes:

[0139] S41. Collect real-time status data of disinfection equipment, operation paths during the disinfection process, dynamically adjusted disinfection process operation parameters, and global compliance score. and the updated symbolic rule base Organized into a dataset ;

[0140] S42. Organize the dataset according to the time sequence of the disinfection process and the unique identifier of the equipment. The data is divided into blocks, each containing a timestamp, device number, operation path, disinfection process operation parameters, and global compliance score. and the updated symbolic rule base ;

[0141] S43. Perform a hash operation on each data block to obtain the block hash value, where the hash input is the content of the current data block and the hash value of the previous block, and the hash algorithm used is consistent with the blockchain ledger.

[0142] S44 will include hash value, timestamp, operation path, disinfection process operation parameters, and global compliance score. and the updated symbolic rule base Data blocks are written sequentially to the blockchain ledger in block form, forming a disinfection data chain. ;

[0143] S45. Authentication of each data block is performed using a digital signature. The digital signature corresponds one-to-one with the node identity to ensure that data writing is non-repudiable and operations are traceable. The signature content includes the block hash value and the node identifier.

[0144] S46. Smart contracts based on blockchain ledgers automatically detect the global compliance score in each data block. With relevant laws and regulations The comparison relationship shows that if the score is lower than the set compliance threshold, it will be automatically marked as a non-compliant event and recorded on the chain.

[0145] S47. For all compliant and non-compliant events, record the disinfection process data and parameter adjustment information in detail to the blockchain ledger to achieve tamper-proof evidence storage and full-process traceability auditing.

[0146] S48. When the blockchain ledger receives a data query, access, or anomaly alarm request, it retrieves disinfection path parameters, equipment status, operating parameters, and global compliance score based on the ledger content. Symbolic rule base And compliance incidents, supporting full-process monitoring and compliance verification of data transparency and traceability;

[0147] S49. Periodically perform integrity checks on the blockchain ledger and verify the consistency of blocks through the hash chain structure.

[0148] This invention unifies and structures real-time status data of disinfection equipment, disinfection process operation paths, dynamically adjusted operation parameters, global compliance scores, and updated symbolic rule bases to form a high-quality dataset, providing a fundamental guarantee for subsequent full-process information storage. Through block storage and hash operations, each data block possesses a unique identifier and tamper-proof characteristics, ensuring the integrity and security of the data chain. The blockchain ledger employs sequential writing and digital signature authentication, achieving non-repudiation and traceability of key data in the disinfection process. An automatic detection mechanism based on smart contracts can compare global compliance scores with regulatory standards in real time. When non-compliance events are detected, they are immediately marked and recorded on the chain, improving risk warning and automatic supervision capabilities. All process data, regardless of compliance, is meticulously stored, greatly enhancing the transparency, compliance, and traceability of disinfection operations. The ledger supports dynamic querying and anomaly alarm responses, providing convenient and efficient technical support for regulatory departments, quality control management, and accountability analysis. Regular integrity checks further ensure the continuous reliability of the blockchain ledger, comprehensively improving the level of trusted storage, risk management, and intelligent supervision of medical disinfection data.

[0149] In this embodiment, S5 specifically includes:

[0150] S51. The system compares the disinfection time, equipment temperature, and other key operating parameters with preset standards in real time, and makes judgments based on the compliance rules set in the smart contract. Specifically, this real-time comparison with preset standards means that during the disinfection operation, the system continuously collects real-time data on disinfection time, equipment temperature, and other key operating parameters, and compares them item by item with the compliance standards stored in the smart contract. The comparison includes, but is not limited to: whether the current disinfection time meets the minimum duration requirement, whether the current equipment temperature is within the set temperature range, and whether parameters such as humidity or pressure remain within safe ranges. The system uses threshold judgment, range verification, or logical rule matching to achieve rapid comparison, ensuring that abnormal states are identified as soon as parameters deviate from the standards, providing a basis for subsequent compliance judgments and early warnings.

[0151] S52. When the smart contract detects that the disinfection time, equipment temperature or other key parameters do not meet the predetermined standards, it will immediately and automatically determine that it is a non-compliant operation.

[0152] S53. For each non-compliant operation, the smart contract automatically triggers an early warning mechanism, generates alarm information, and sends it to the relevant responsible personnel or management terminal in real time. The alarm information specifically refers to a set of structured data automatically generated by the smart contract after detecting a non-compliant operation. The content includes the timestamp of the anomaly, the corresponding disinfection equipment number, the operator's identification, the specific non-compliant parameter items and their deviation from the standard, the non-compliance level assessment result, and the suggested preliminary handling measures. This information is pushed to the responsible personnel's mobile terminal, management backend, or alarm control platform in real time through the system interface for rapid response and handling of abnormal events.

[0153] S54. Automatically record and write all non-compliant operation details, including operation time, device number, abnormal parameters, and specific values ​​deviating from the standard, into the blockchain ledger.

[0154] S55. Supports querying, retrieving, and auditing alarm information and non-compliant operations recorded in the blockchain ledger, facilitating the tracing of anomalies in the disinfection process and promoting rectification. Specifically, the querying, retrieving, and auditing refer to the ability of users to quickly filter and locate data in the blockchain ledger by time range, device number, anomaly type, etc., through interface tools, and to call historical alarm records and details of non-compliant operations. The system supports the generation of operation logs and event chains to assist managers in reconstructing abnormal processes, verifying the source of parameter deviations, and providing visual audit reports, comprehensively supporting the tracing, responsibility allocation, and rectification analysis of anomalies in the disinfection process.

[0155] This invention uses smart contracts to compare key operational parameters such as disinfection time and equipment temperature with preset standards in real time, enabling the timely detection of any non-compliant operations during the disinfection process. The smart contracts possess automatic judgment and response capabilities; once an anomaly is detected, it is immediately and automatically determined to be non-compliant, triggering an early warning mechanism in real time to quickly send alarm information to relevant responsible persons or management terminals, significantly improving the timeliness of risk response and operational safety. Detailed information on each non-compliant operation, including operation time, equipment number, and specific abnormal parameters, is automatically written into the blockchain ledger, achieving tamper-proof and transparent record-keeping throughout the entire process. The system also supports querying, retrieving, and auditing all alarms and non-compliant events, facilitating efficient tracing of the causes of anomalies by managers and timely promotion of process rectification and risk closure. Overall, this invention significantly improves the automation, intelligence, and compliance level of disinfection process monitoring, providing strong data support and risk prevention capabilities for infection control and quality management in medical institutions.

[0156] In this embodiment, S6 specifically includes:

[0157] Based on equipment status data, operation paths, parameter anomaly records, and compliance scoring results collected during real-time verification, a structured and detailed disinfection report is automatically generated. This report includes operation parameter values ​​for each stage, compliance analysis results compared to standards, judgment criteria, anomaly logs, and historical data comparison results, ensuring traceability, quantification, and verification. Simultaneously, through a built-in feedback mechanism and corrective suggestions from the neural symbolic reasoning module, the system automatically generates a list of rectification recommendations for non-compliant operations. This list is then communicated to operators immediately via digital terminals or prompt interfaces, helping them quickly pinpoint problematic areas, guide parameter adjustments and process corrections, and achieve closed-loop management and continuous optimization of the disinfection process.

[0158] Example 1:

[0159] To verify the feasibility of this invention in practice, it was applied to the digestive endoscopy center of a tertiary hospital, which performs approximately 24,000 endoscopic examinations and treatments annually. Previously, the center used a traditional method of manual inspection combined with decentralized information system recording. The compliance rate of the disinfection process was affected by factors such as equipment fluctuations, environmental interference, and human error, making data traceability difficult and abnormal disinfection operations difficult to detect in a timely manner and manage in a closed loop. In May 2024, the hospital deployed the "Full-cycle Monitoring and Compliance Verification Platform for Digestive Endoscopy Disinfection Process" proposed in this invention on all four endoscopic disinfection devices and their associated drying cabinets and testing areas.

[0160] The platform first uses automated data acquisition devices and environmental sensors to acquire real-time key data throughout the entire process, including equipment operating status, operational parameters (such as temperature, humidity, disinfectant concentration, and disinfection duration), ambient humidity, and operator identity. This data is preprocessed to generate a standardized dataset, which serves as input for the platform's intelligent optimization and compliance assessment. The platform's core algorithm uses Monte Carlo tree search to simulate the entire process path across multiple operational stages, including cleaning, primary disinfection, drying, and testing, and intelligently selects the optimal path based on the real-time operating conditions of each disinfection task. Subsequently, the neural symbolic reasoning module intelligently assesses the compliance of each operational step based on national standards, hospital disinfection SOPs, and real-time collected process parameters, automatically generating rectification suggestions for non-compliant or risky paths and supporting dynamic adjustment of disinfection parameters.

[0161] All key data and compliance scores in the disinfection process are automatically verified by smart contracts and written to the blockchain ledger in real time. Each record includes the device number, operator ID, timestamp, and digital signature, ensuring the entire process is tamper-proof and traceable. If the system detects insufficient disinfection time, temperature below standard, or abnormal operating parameters, it automatically alarms within 15 seconds and notifies the responsible nurse and infection control personnel via multiple terminals such as mobile phones and computers. All abnormal data, rectification processes, and feedback results are synchronously stored in the blockchain ledger, facilitating subsequent accountability, quality control, and auditing by regulatory departments. The platform automatically generates disinfection reports for each batch, providing reliable data for infection control management, internal self-inspection, and external supervision.

[0162] From May to July 2024, a tertiary hospital compared its disinfection process monitoring data and compliance management effectiveness before and after the implementation of this invention's system, achieving significant improvements. After the platform went live, the compliance rate of disinfection processes increased, the response time to anomalies was significantly shortened, the rate of manual intervention and underreporting decreased substantially, the risk of hospital-acquired infections was effectively controlled, and no cross-infection incidents caused by non-compliance with disinfection regulations occurred. The following is a comparison of the main data at the center from February to April 2024 (before the system went live) and from May to July 2024 (after the system went live).

[0163] Table 1. Comparison of the effects of the intelligent monitoring platform for disinfection procedures in the digestive endoscopy center before and after application.

[0164]

[0165] As clearly shown in Table 1 above, the implementation of this invention effectively improves the compliance and safety of the disinfection process. Firstly, the average compliance rate of the disinfection process increased from 95.6% before the system went live to 99.7%, indicating that the platform achieved intelligent control over the entire disinfection operation process, greatly reducing the risks caused by human error and non-standard operations. The average timeout for abnormal disinfection alarms was shortened from 2.2 hours to 18 seconds, and the time for correcting abnormal equipment parameters decreased from 61 minutes to 8 minutes, demonstrating the system's extremely high automation and real-time performance in anomaly detection, response, and problem handling, significantly improving risk warning and processing efficiency. The number of non-compliant disinfection batches decreased significantly, from 43 batches to 5 batches, greatly reducing the risk of hospital-acquired infections and management pressure.

[0166] Furthermore, the accuracy rate of data traceability improved from 91.8% to 100%, demonstrating that blockchain-based evidence storage and end-to-end digital management enable accurate and complete traceability of all disinfection data, completely eliminating data loss and tampering. Since the platform's launch, a total of 317 batches of disinfection reports have been automatically generated, providing efficient and standardized data support for managers and regulatory departments. More notably, the number of disinfection-related hospital-acquired infections has decreased from 1 to 0, essentially eliminating safety hazards. The proportion of manual recording and intervention has also decreased from 100% to 14.2%, significantly reducing the burden on medical staff and improving work efficiency. Overall, the platform's application has not only achieved intelligent, standardized, and data-driven disinfection processes but has also brought about a qualitative leap in the hospital's medical safety and management level.

[0167] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for monitoring and verifying compliance of the entire cycle of the disinfection process of digestive endoscopes, characterized in that, The method comprises the following steps: S1, collecting real-time state data and environmental data of the disinfection equipment and preprocessing to generate a standardized data set; S2, based on the standardized data set, simulating multiple operation paths of the disinfection process using the Monte Carlo tree search algorithm, selecting the optimal disinfection path according to multiple variables in the disinfection process; S3, based on the optimal disinfection path, using neural-symbolic reasoning to reason the symbolic rules in the disinfection process, and combining relevant regulatory standards to verify the compliance of the disinfection process, and dynamically adjusting the operation parameters in the disinfection process; S4, using blockchain technology to record the real-time state data of the disinfection equipment, the operation path in the disinfection process, the operation parameters in the disinfection process and the compliance verification results, and generate an unalterable disinfection data chain; S5, real-time verification of the disinfection data chain through the smart contract, automatic triggering of the alarm and recording of the non-compliant operation, when the disinfection time, equipment temperature or other parameters do not meet the predetermined standards, the smart contract will automatically issue an alarm and record the relevant information; S6, based on the real-time verification result, generating a detailed disinfection report, and through the feedback mechanism, the rectification suggestion of the non-compliant operation is timely delivered to the operator; The real-time state data of the disinfection equipment specifically includes the running state, temperature, humidity, pressure, disinfection time, operator information and working state of the equipment, and the environmental data specifically includes the temperature, humidity, air quality and illumination intensity in the disinfection environment; S2 specifically includes: S21, based on the generated standardized dataset, dividing the disinfection process into a plurality of independent stages, each stage defined as a hierarchical state space including a pre-treatment, main disinfection, drying, detection and storage stage S22, in each hierarchical state space S22, in each hierarchical state space S22, in each hierarchical state space S22, in each hierarchical state space S22, in each hierarchical state space S22, in each hierarchical state space S22, in each hierarchical state space S22, in each hierarchical state space S22, in each hierarchical state space <000002 S23, during the node expansion of each hierarchical sub-tree, according to the pre-set disinfection process compliance standards, including disinfection equipment operation specifications, disinfection time threshold, temperature and pressure and safety rules, rule judgment is carried out for each extendable action, only the action meeting all the standards is expanded, forming a constraint-aware adaptive tree structure, and the safety rules include equipment running safety parameters and operator identity verification; S24, in the process of expanding each node, through the monitoring of real-time disinfection state data, dynamically generating risk assessment factors of disinfection process , path reward correction is performed; The path reward correction specifically refers to that in each node expansion process of the Monte Carlo tree search, based on the real-time collected disinfection state data, the risk factors possibly existing in the current operation path are evaluated, and the risk level is quantified as a risk evaluation factor; S25, introduce an adaptive reward mechanism based on real-time environmental feedback, adjust the path reward function in real time through monitoring data, dynamically adjust the priority of path selection, adjust the reward function ; ； wherein, is a simulation reward, is a risk assessment reward generated based on real-time monitoring data, is a feedback reward based on disinfection status, , , is a weight factor; S26, in the simulation process, combining the real-time state data of the disinfection equipment, the environmental data and the risk evaluation of the current disinfection path, the simulation path selection is dynamically guided through the adaptive strategy network, and the adaptive strategy network includes: The input layer receives the real-time state data and environmental data of the disinfection equipment; The hidden layer extracts features and processes information of the real-time state data and environmental data of the disinfection equipment through a multilayer perceptron; The output layer generates selection probability of each operation path through the Softmax function to determine the priority of the path; The adaptive strategy network is trained based on historical operation state, disinfection time, disinfection effect and monitoring data of the disinfection equipment, continuously adjusts the ratio of exploration and utilization, sets thresholds of compliance, risk and efficiency, preferentially selects an operation sequence meeting preset compliance standards, meeting risk control requirements and being higher than the set threshold in efficiency, optimizes path selection in the disinfection process, and flexibly responds to and adjusts the decision strategy according to real-time disinfection environment changes; S27. In the path expansion and simulation process, paths that do not meet the compliance standards, violate the timing or safety requirements, or the path reward does not reach the set threshold are pruned according to the disinfection process compliance standards and the adaptive reward mechanism, and the expansion and simulation of the paths are terminated in advance; S28, by parallel computing technology, the simulation tasks of different hierarchical state spaces and different operation paths are processed in multi-thread or distributed manner; S29. In each simulation and backtracking process, the return value of each hierarchical path is accumulated, and the optimal decision parameter of the current layer is passed to the next layer state space, realizing local optimal decision of each layer and global optimal path selection of the entire disinfection process; S210. Steps S21 to S29 are repeatedly executed until a preset simulation number or a calculation resource threshold is reached, and finally the optimal disinfection path of the entire disinfection process is output as the basis for the optimization decision of the disinfection process; The S3 specifically comprises: S31, based on the generated optimal disinfection path, the operation parameters, disinfection time, temperature and humidity information in the disinfection process are extracted, and combined with relevant regulatory standards , and jointly construct a symbolization rule base , wherein n is the total number of symbol rules, is the number of relevant regulatory standards; S32. Building a neural-symbolic reasoning model, specifically comprising: Input layer, receiving a library of encoding rules containing relevant regulatory standards real-time status data and operational sequences of sterilization equipment ; a neural network layer, which performs feature extraction and high-dimensional representation learning on the data received by the input layer, and outputs high-dimensional feature representation; a symbol rule coding module, coding the symbol rules and the regulatory standards in the input layer vectorization coding is performed to obtain rule embedding; The reasoning mechanism integration module inputs the high-dimensional feature representation output by the neural network layer and the rule embedding output by the symbolic rule encoding module into a symbolic reasoning engine, performs joint reasoning through logical rules, and outputs a local compliance judgment value ; an output layer outputting a local compliance decision value vector , representing a local compliance decision value, ; S33, based on the obtained local compliance judgment value vector , combined with the real-time state data of the disinfection equipment received by the input layer, the high-dimensional feature representation output by the neural network layer, the rule embedding output by the symbolic rule coding module and the inference result of the inference mechanism integration module, each local compliance judgment value is compared with the related regulation standard , and combined with the time sequence dependence relationship of the disinfection operation sequence , risk assessment and real-time environmental feedback data, the time sequence risk weighted global compliance score is calculated ; ； wherein, is a local compliance decision item weight, is a first step risk coefficient, is an environmental feedback weighting coefficient, is an environmental feedback correction value, is real-time feedback data, is a timing dependency weighting coefficient, is a timing dependency decision result of the operation sequence, H is the total number of local compliance decision items, is a matching indication function of the rule and the regulatory standard, if the relevant regulatory standard is not met , the decision is non-compliant, otherwise, it is compliant, and a correction mechanism is triggered. S34. For non-compliant or high-risk paths, based on the local compliance judgment value output by the reasoning mechanism integration module, a set of automatic correction operation suggestions is generated, and the operation parameters in the disinfection path are adjusted, each correction operation suggestion corresponding to a specific correction action; S35. The set of correction operation suggestions is fed back to update the disinfection path operation parameters, and the disinfection path operation parameters are modified by increasing or decreasing based on the correction suggestions; S36, for the path determined to be compliant, continue to call the neural network layer and integrate the inference mechanism module, optimize and adjust the disinfection parameters, pass through the strategy function output the optimal parameter set ; ； wherein, is the disinfection cost of the path , is the candidate disinfection path operation parameter set, represents the parameter corresponding to the jth operation step in the disinfection path, represents whether the parameter of the jth operation step meets the compliance requirement, and the compliance value is 1, and the non-compliance value is 0, represents the parameter set P that minimizes the total cost, and X represents the number of disinfection operation steps; S37, in the path optimization process, the symbol rule coding module is integrated with the reasoning mechanism set module, the symbol rule library is automatically updated according to the real-time state data of the disinfection equipment and environmental feedback, and an updated symbol rule library is generated ; S38, training the neural-symbolic reasoning model through a reinforcement learning mechanism, optimizing the disinfection path selection and the symbolic rule generation, through a compliance loss function minimizing the compliance loss; ； wherein, is a compliance penalty, is the total number of training samples, denotes the global compliance score corresponding to the th training sample, is a pre-defined compliance threshold, is an indicator function that takes the value 1 if the compliance score of the th sample is below the compliance threshold , indicating that the sample is non-compliant, and 0 otherwise. S39. The updated symbolic rule base and operation parameters are input to re-simulate the path, ensuring that the disinfection process is always compliant, efficient and safe, and continuously monitoring and optimizing the disinfection path through a feedback mechanism.

2. The method of claim 1, wherein the method further comprises: The preprocessing specifically includes cleaning, standardizing, missing value filling, outlier detection and data format conversion of real-time state data and environmental data of the disinfection equipment.

3. The method of claim 1, wherein the method further comprises: The S4 specifically comprises: S41, collect real-time state data of the disinfection equipment, operation path in the disinfection process, operation parameter of the dynamically adjusted disinfection process, global compliance score and the updated symbolization rule library , and sort into a data set ; S42, ordering the dataset according to the time sequence of the disinfection procedure and the device unique identification performing chunking, each data chunk containing a timestamp, a device number, an operating path, disinfection procedure operating parameters, a global compliance score and an updated library of encoding rules ; S43. Hash operation is performed on each data block to obtain a block hash value, wherein the hash input is the current data block content and the previous block hash value, and the hash algorithm used is consistent with the blockchain ledger; S44, sequentially writing the data block containing the hash value, the timestamp, the operation path, the disinfection process operation parameter, the global compliance score and the updated symbolization rule library into the blockchain ledger in the form of blocks to form a disinfection data chain ; S45. Each data block is authenticated using a digital signature, and the digital signature corresponds one-to-one to the node identity, ensuring non-repudiation of data writing and operation traceability, and the signature content includes the block hash value and the node identifier; S46. The smart contract based on the blockchain ledger automatically detects the global compliance score in each data block comparisons with relevant regulatory standards , if the score is below the set compliance threshold, it is automatically flagged as a non-compliance event and recorded on the chain; S47. For all compliant and non-compliant events, detailed records of disinfection process data and parameter adjustment information are recorded to the blockchain ledger, realizing tamper-proof evidence and traceable audit of the entire process; S48, when the blockchain ledger receives a data query, call or exception alert request, based on the ledger content to retrieve disinfection path parameters, device status, operating parameters, global compliance score , a symbolization rule library and compliance events, supporting data transparency and traceability for the whole process monitoring and compliance verification; S49. The integrity of the blockchain ledger is checked regularly to verify the consistency of the blocks through the hash chain structure.

4. The method of claim 3, wherein the method further comprises: The S5 specifically comprises: S51, compare the disinfection time, equipment temperature and other key operation parameters with the preset standard in real time, and judge according to the compliance rules set by the smart contract; S52, when the smart contract detects that the disinfection time, equipment temperature or other key parameters do not meet the predetermined standard, it immediately determines that the operation is not compliant; S53, for each non-compliant operation, the smart contract automatically triggers the early warning mechanism, generates an alarm information, and sends it to the relevant responsible personnel or management terminal in real time; S54, record all the specific information of non-compliant operation, including operation time, equipment number, abnormal parameter and specific value deviating from the standard, and automatically write into the block chain account book; S55, support to query, search and audit the alarm information and non-compliant operation recorded in the block chain account book, which is convenient for tracing the abnormal situation in the disinfection process and promoting rectification.

5. The platform for monitoring and verifying the compliance of the whole cycle of the digestive endoscope disinfection process according to any one of claims 1-4, characterized in that, It includes: Data acquisition and preprocessing module, used for collecting real-time state data and environmental data and preprocessing to generate standardized data set; Path optimization and decision module, used for simulating multiple operation paths of disinfection process based on standardized data set using Monte Carlo tree search algorithm, and selecting the optimal disinfection path; Compliance reasoning and verification module, used for reasoning the symbolic rules in the disinfection process based on the optimal disinfection path, and outputting the operation parameters and compliance verification results; Block chain storage module, used for writing the real-time state data of disinfection equipment, operation path, operation parameter and compliance verification result into block chain account book in blocks, forming a disinfection data chain that cannot be tampered with; Smart contract processing module, used for real-time judgment of disinfection data chain, automatic triggering of alarm and recording of non-compliant operation; Feedback and query module, used for querying, searching and auditing the whole process information of disinfection process, alarm record and compliance event.

Images