Sanitary disinfection production intelligent collaborative optimization system and control method

Abstract

The invention relates to the technical field of intelligent disinfection, in particular to an intelligent collaborative optimization system for sanitary disinfection production and a control method, which are suitable for high-precision disinfection scenes such as hospital disinfection supply centers, biosafety laboratories, pharmaceutical clean workshops, food processing environments and the like. Intelligent distribution of disinfection tasks, collaborative operation of multiple devices, dynamic optimization of process parameters and real-time verification of effects can be achieved, and the technical problems that in traditional disinfection production, resource allocation is unreasonable, device collaboration is poor, parameter adjustment lags behind, and quality tracing is difficult are solved.

Description

Technical Field

[0001] This invention relates to the field of intelligent disinfection technology, specifically to an intelligent collaborative optimization system and control method for hygiene and disinfection production. Background Technology

[0002] With increasingly stringent public health and safety requirements, disinfection production has become a crucial link in ensuring product quality and personnel safety. Current disinfection production models face four major technical bottlenecks: First, task allocation is inefficient, employing fixed procedures to assign disinfection tasks without considering the dynamic matching of equipment capacity with task demands, resulting in low resource utilization. Second, equipment coordination is inefficient, with multiple disinfection devices operating independently without a unified coordination mechanism, leading to overlapping or missed operations. Third, parameter adjustment relies on experience, depending on manual setting of disinfection parameters without dynamic adjustments based on real-time monitoring data, resulting in unstable disinfection effects. Fourth, quality control is lagging, using post-event sampling and testing methods that cannot detect and correct process deviations in real time, leading to higher quality risks.

[0003] In existing technologies, some solutions improve operational efficiency by introducing automated equipment, but lack intelligent collaboration mechanisms; some solutions attempt automatic parameter adjustment, but have not established a precise model based on disinfection mechanisms; and some solutions strengthen quality inspection, but have not achieved real-time monitoring of the entire process. Therefore, there is an urgent need to develop an intelligent hygiene and disinfection production system that integrates intelligent task allocation, multi-device collaborative control, dynamic parameter optimization, and real-time quality control to achieve intelligent, efficient, and precise operation of the entire disinfection production process. Summary of the Invention

[0004] This invention provides an intelligent collaborative optimization system and control method for hygiene and disinfection production. Through a four-layer collaborative architecture of "task-resource-process-quality," it integrates intelligent task allocation algorithms, multi-device collaborative control technology, dynamic parameter optimization models, and real-time quality assessment methods to achieve intelligent management and control of the entire disinfection production process. The core innovation lies in proposing three key algorithms: an intelligent allocation algorithm based on matching task characteristics with equipment capabilities, a multi-device collaborative operation path planning and conflict resolution algorithm, and a parameter dynamic optimization algorithm based on real-time feedback of disinfection effects. Corresponding mathematical models and efficiency-enhancing mechanisms are established to significantly improve resource utilization efficiency, equipment collaboration level, and disinfection effect stability, while reducing operating costs and quality risks.

[0005] A first aspect of this invention provides an intelligent collaborative optimization method for hygiene and disinfection production, comprising the following steps:

[0006] Disinfection task and resource information collection: Collect task information such as the type, quantity, degree of contamination, material characteristics, and disinfection standards of the items to be disinfected, as well as resource information such as the type, capacity, status, and location of disinfection equipment;

[0007] Intelligent task allocation: Based on the collected data, the system uses an intelligent allocation algorithm that matches task characteristics with equipment capabilities to achieve optimal matching between disinfection tasks and equipment;

[0008] Collaborative path planning: Based on the task allocation results, initiate multi-device collaborative operation path planning and conflict resolution algorithms to generate conflict-free optimal operation paths;

[0009] Dynamic parameter optimization: A parameter dynamic optimization algorithm based on real-time feedback of disinfection effect is adopted to adjust disinfection parameters in real time and control the equipment to perform operations;

[0010] Quality assessment and continuous optimization: Real-time assessment of disinfection quality is achieved through online monitoring and intelligent analysis. The algorithm parameters are continuously iterated and optimized by combining historical data to form a closed-loop control throughout the entire process.

[0011] A second aspect of this invention provides an intelligent collaborative optimization system for hygiene and disinfection production that implements the above-described method, comprising:

[0012] Information collection unit:

[0013] The item identification system (using machine vision and RFID technology, with an identification speed of ≥10 items / second and an identification accuracy of ≥99%) is used to collect information such as the type, quantity, and material of the items to be disinfected.

[0014] Contamination detection module (including ATP biofluorescence detector and microbial sensor, detection range 1-10) 9 CFU (response time ≤ 5 seconds);

[0015] Equipment status monitor (monitoring parameters include operating status, load rate, energy consumption, and fault warning; sampling frequency 1Hz);

[0016] Environmental sensors (temperature 0-100℃, humidity 10-90%RH, air pressure 80-120kPa, measurement accuracy ±0.5℃, ±3%RH, ±0.5kPa respectively);

[0017] Positioning system (using UWB technology, positioning accuracy ±10cm, supports simultaneous positioning of multiple devices).

[0018] Intelligent decision-making unit:

[0019] Industrial server (CPU Intel Xeon E5-2690 v4, 32GB RAM, 1TB SSD storage, processing speed ≥5×10) 11 (times / second);

[0020] Algorithm acceleration module (GPU NVIDIA GTX 1080Ti, supports parallel computing, inference latency ≤10ms);

[0021] Real-time database (supports 100,000 data writes / second, storage capacity of 5TB, data retention period ≥3 years);

[0022] Decision software platform (integrates task allocation, path planning, and parameter optimization algorithms, and supports web access and mobile monitoring).

[0023] Cooperative control unit:

[0024] Main controller (using PLC S7-1200, processing speed 0.1μs / instruction, supporting PROFINET communication);

[0025] Equipment controller (each device is equipped with an independent controller, with a response time ≤10ms);

[0026] Motion control module (controls ≥16 axes, position control accuracy ±0.01mm, speed control accuracy ±0.1%);

[0027] Communication gateway (supports multiple communication methods such as Ethernet, Wi-Fi, and Bluetooth, with a data transmission rate of ≥100Mbps).

[0028] Disinfection execution unit:

[0029] Various types of disinfection equipment (including high-pressure steam sterilizers, hydrogen peroxide low-temperature plasma sterilizers, ultraviolet disinfection cabinets, ultrasonic cleaning and disinfection machines, etc.);

[0030] Intelligent conveying system (consisting of conveyor belt and robotic arm, with adjustable conveying speed of 0-5m / min and positioning accuracy of ±5mm);

[0031] Parameter adjustment device (including power regulator, concentration controller, and temperature regulator, with adjustment accuracies of ±1%, ±0.1%, and ±0.5℃, respectively);

[0032] End effector (a dedicated actuator is configured according to different disinfection objects, with compatibility ≥95%).

[0033] Quality assessment and traceability unit:

[0034] Online quality inspection device (detection methods include optical detection and biosensor detection, with 100% detection coverage);

[0035] Data traceability system (using blockchain technology, ensuring data immutability and supporting full-process traceability);

[0036] Quality analysis software (supports real-time quality monitoring, trend analysis, and anomaly warning);

[0037] Report generation module (automatically generates disinfection reports that comply with GMP, ISO 13485 and other standards).

[0038] Beneficial effects

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] Resource utilization efficiency is significantly improved: The intelligent task allocation algorithm achieves precise matching between tasks and equipment, increases equipment capacity utilization by more than 40%, reduces equipment idle rate by 50%, shortens task completion time by 30%, and greatly improves the overall efficiency of the system.

[0041] The overall level of equipment collaboration has been improved: multi-equipment collaborative path planning shortens the total operation path by more than 35%, reduces operation time by 40%, and increases space utilization by 50%; the real-time conflict resolution mechanism avoids mutual waiting and interference between equipment, increases equipment utilization by 60%, and significantly improves operation smoothness.

[0042] Enhanced stability of disinfection effect: The dynamic parameter optimization algorithm improves the stability of disinfection effect by more than 80%, reduces batch-to-batch differences by 70%, and effectively avoids over-disinfection and under-disinfection; the adaptive learning mechanism enables the system to adapt to different types of disinfection objects, reducing changeover time by 90%.

[0043] Significantly reduced operating costs: Optimized resource allocation and energy consumption control reduced overall operating costs by more than 25%; improved equipment utilization and reduced wear and tear reduced equipment maintenance costs by 30%; and increased automation reduced labor costs by 60%.

[0044] Quality control capabilities have been comprehensively upgraded: real-time quality assessment enables full monitoring of the disinfection process, reducing the response time for quality anomalies from hours to minutes; the blockchain traceability system ensures the authenticity and integrity of the data, meeting stringent industry regulatory requirements.

[0045] The system has wide adaptability: It can be applied to the disinfection needs of multiple fields such as hospitals, pharmaceuticals, and food. It supports the collaborative work of various types of disinfection equipment, is compatible with more than 95% of common disinfection objects, and has broad application prospects.

[0046] This invention, through the synergistic effect of three core algorithms, realizes the transformation of hygiene and disinfection production from "experience-driven" to "data-driven" and from "independent operation" to "collaborative optimization," providing a comprehensive intelligent solution for high-precision disinfection scenarios, with significant economic and social benefits. Attached image description:

[0047] Appendix Figure 1 Process flow diagram. Detailed Implementation

[0048] Example 1

[0049] 1: Intelligent allocation algorithm based on matching task characteristics with equipment capabilities

[0050] Algorithm principle: Construct a quantitative evaluation model of disinfection task characteristics and equipment capabilities, and achieve optimal matching between tasks and equipment through intelligent optimization algorithms to improve resource utilization efficiency and operational efficiency.

[0051] Key innovations:

[0052] The following are proposed: Task feature vector and equipment capability vector: Task feature vector T = [t1, t2, t3, t4, t5], where t1 is the item type, t2 is the degree of contamination, t3 is the material sensitivity, t4 is the disinfection standard level, and t5 is the urgency level; Equipment capability vector E = [e1, e2, e3, e4, e5], where e1 is the equipment type, e2 is the disinfection capability, e3 is the material compatibility, e4 is the accuracy level, and e5 is the current load;

[0053] Establish a task-equipment matching degree calculation model: Where M i,j w represents the matching degree between the i-th task and the j-th device. k is the weight coefficient of the k-th feature, and sim() is the similarity calculation function with a value range of [0,1]. The larger the value, the higher the similarity.

[0054] Design an objective function for task allocation optimization: Where x i,j For the allocation variable (1 indicates allocation, 0 indicates no allocation), C i,j L is the cost of assigning task i to device j. i For the load of task i, Let j be the maximum load capacity of device j.

[0055] Modeling efficiency enhancement principle:

[0056] Precise task-equipment matching avoids the phenomena of "using large equipment for small purposes" and "using small equipment for large purposes", improving equipment capacity utilization by more than 40% and shortening task completion time by 30%.

[0057] Multi-objective optimization achieves a balance between matching degree and cost, reducing operating costs by 25% while ensuring disinfection quality;

[0058] The dynamic allocation mechanism can adjust task allocation in real time according to the device status, reducing device idle rate by more than 50% and significantly improving the overall system efficiency.

[0059] Innovation Point 2: Multi-device collaborative operation path planning and conflict resolution algorithm

[0060] Algorithm principle: Based on the task allocation results, a path planning model for multi-device collaborative operation is constructed. The optimal path without conflict is generated through intelligent algorithms, thereby improving operation efficiency and space utilization.

[0061] Key innovations:

[0062] A path cost function is proposed: taking into account path length, operation efficiency, conflict risk and energy consumption, a path cost model is constructed: Cost(r)=α·L(r)+β·T(r)+γ·R(r)+δ·E(r) is the total cost of path r, L(r) is the path length, T(r) is the operation time, R(r) is the conflict risk coefficient, E(r) is the energy consumption, and α,β,γ,δ are weight coefficients, satisfying α+β+γ+δ=1;

[0063] Designing an improved ant colony algorithm for path planning: Introducing a dynamic pheromone update mechanism. Where τ ij The pheromone evaporation coefficient, The pheromone left by the k-th ant on path ij is added, and a conflict penalty factor is added to increase the cost of the corresponding path when a path conflict is detected.

[0064] Establish a real-time conflict resolution mechanism: Use the time window method to allocate work space and time to each device. When a potential conflict is detected, the device work order is dynamically adjusted by priority adjustment (Priority = ω1·Urgency + ω2τProgress, where Urgency is the urgency of the task and Progress is the work progress) to achieve real-time conflict resolution.

[0065] Modeling efficiency enhancement principle:

[0066] Multi-device collaborative path planning shortens the total operation path by more than 35%, reduces operation time by 40%, and increases space utilization by 50%.

[0067] The real-time conflict resolution mechanism avoids mutual waiting and interference between devices, increases equipment utilization by 60%, and significantly improves operational smoothness;

[0068] Energy-optimized design reduces overall energy consumption by more than 20%, while also reducing equipment wear and tear and extending service life.

[0069] Innovation Point 3: Dynamic parameter optimization algorithm based on real-time feedback of disinfection effect

[0070] Algorithm principle: Construct a dynamic response model between disinfection effect and operating parameters. By monitoring the disinfection effect in real time and feeding back the adjustment parameters, the disinfection process can be precisely controlled, thereby improving the stability and consistency of the disinfection effect.

[0071] Key innovations:

[0072] Establish a dynamic response model for disinfection effectiveness: describe the relationship between disinfection effectiveness and key parameters. Where E(t) represents the disinfection effect (kill rate) at time t, E0 represents the initial effect, and P(t) represents the disinfection effect at time t. α Let C(t) be the disinfection power. β Where is the drug concentration, T(t) is the temperature, α, β, k1, k2 are model parameters, and γ(τ) is the time decay factor;

[0073] Design a parameter tuning strategy based on model predictive control: The objective function within the prediction window is: Where E ref To achieve the desired disinfection effect, The prediction result at time k is... To control the change in the quantity, λ is the regularization coefficient and N is the prediction step size;

[0074] An adaptive learning mechanism is proposed: the model parameters are updated in real time using online least squares method. in Here are the parameter estimates, K(t) is the gain matrix, and E is the gain matrix. meas Φ(t) represents the measured results, and Φ(t) represents the regression vector, enabling the model to adapt to different disinfection objects and environmental conditions.

[0075] Modeling efficiency enhancement principle:

[0076] The dynamic response model enables accurate prediction of disinfection effects, with prediction errors controlled within 5%, providing a scientific basis for parameter adjustment.

[0077] Real-time feedback control improves the stability of disinfection effect by more than 80%, reduces batch-to-batch differences by 70%, and effectively avoids over-disinfection and under-disinfection.

[0078] The adaptive learning mechanism enables the system to adapt to different types of disinfection objects without the need for manual recalibration, reducing changeover time by 90% and significantly lowering operational complexity.

[0079] Example 2

[0080] 1) Dynamic matching algorithm: "Precise pairing" of tasks and devices

[0081] 1. Algorithm Principle: Feature-Driven Intelligent Matching Mechanism

[0082] The core innovation of this algorithm lies in breaking through the traditional "one-size-fits-all" allocation model and constructing an intelligent system of "task feature quantification - equipment capability modeling - dynamic matching optimization". By sensing task requirements and equipment status in real time, it achieves optimal matching between the two, ensuring that each piece of equipment is "fully utilized" and each task is "allocated on demand". Its modeling logic is derived from multi-attribute decision theory, transforming abstract task requirements and equipment capabilities into quantifiable and comparable feature indicators, and improving matching accuracy through dynamic weight adjustment.

[0083] The modeling process is divided into three core stages: task feature quantification, equipment capability modeling, and matching optimization. In the task feature quantification stage, the system collects multi-dimensional information about the items to be disinfected through an item identification system (machine vision + RFID) and a contamination detection module, constructing a task feature matrix. This matrix includes item type (surgical instruments / pharmaceutical packaging / food containers), contamination level (high / medium / low, quantified using an ATP bio-fluorescence detector), material sensitivity (metal / plastic / rubber, corresponding to tolerance to different disinfection methods), disinfection standard (sterile / hygienic / general), and urgency level (levels 1-5, with level 1 being emergency equipment). These qualitative and quantitative indicators are then transformed into unified-dimensional feature values ​​using a fuzzy comprehensive evaluation method, forming a standardized task feature vector.

[0084] The equipment capability modeling phase corresponds to the task characteristics. Operating data from the disinfection equipment is collected through equipment status monitors to construct an equipment capability matrix, including equipment type (high-pressure steam / hydrogen peroxide plasma / ultraviolet), disinfection capability (kill rate range), material compatibility (list of supported materials and tolerance parameters), accuracy level (parameter adjustment accuracy), and current status (load rate, fault warning, energy consumption level). Similarly, through standardization, equipment capabilities are transformed into capability vectors corresponding to task characteristics, ensuring direct comparison between the two. The matching optimization phase is the core of the algorithm, achieving optimal allocation by constructing a dynamic matching degree model and a multi-objective optimization function. The dynamic matching degree model calculates a matching score based on the similarity between task characteristics and equipment capabilities. For example, highly contaminated surgical instruments have a 90% matching degree with high-pressure steam sterilizers, but only a 40% matching degree with ultraviolet disinfection cabinets. Simultaneously, a real-time weight adjustment mechanism is introduced. For emergency instruments with an urgency level of 1, the weight of the "urgency level" feature is increased from the usual 10% to 30%, ensuring priority allocation to idle equipment. Multi-objective optimization functions, on the basis of maximizing the matching degree, take into account both minimizing costs (energy consumption, consumables) and maximizing efficiency (task completion time), avoiding cost spikes in pursuit of high matching degree.

[0085] 2. Solution process: The response time from data acquisition to allocation and execution is in the milliseconds.

[0086] The algorithm's solution process is a real-time closed loop of "data acquisition - feature quantization - matching calculation - optimization decision - allocation execution," with a total time of ≤5 seconds, meeting the rapid response needs of emergency tasks such as emergency medical equipment. Taking a batch of mixed instruments from a hospital's sterilization supply center as an example: the system first quickly identifies the instrument type (including 5 emergency surgical scissors, 10 ordinary syringes, and 8 plastic infusion sets) using RFID, determines the material using machine vision, and uses an ATP detector to determine the degree of contamination (surgical scissors are highly contaminated, and the rest are medium to low contamination). The urgency level is marked as Level 1 for surgical scissors and Level 3 for the rest.

[0087] Meanwhile, the equipment status monitor reports the current equipment status: Autoclave A has a 30% load (idle), hydrogen peroxide plasma sterilizer B has an 80% load (almost complete), and UV sterilizer C is completely idle. After quantifying the task and equipment characteristics, the algorithm calculates the matching degree: surgical scissors match autoclave A 92%, B 88%, and C 45%; syringes match A 75%, B 60%, and C 80%; infusion sets (plastic material) match A 50% (easily deformed at high temperatures), B 90%, and C 85%.

[0088] After comprehensive calculation using a multi-objective optimization function, the allocation results are output as follows: surgical scissors are preferentially allocated to autoclave A (high matching degree + idle time); syringes are allocated to UV sterilizer C (high matching degree + low cost); and infusion sets are allocated after hydrogen peroxide plasma sterilizer B is completed (to avoid high-temperature damage + optimal matching degree). The allocation command is sent to the collaborative control unit in real time, and the equipment immediately starts the preparation process. The entire process, from instrument delivery to allocation completion, takes only 3.2 seconds.

[0089] 3. Efficiency Enhancement Principle: From "Resource Waste" to "High-Efficiency Utilization"

[0090] The core efficiency-enhancing value of this algorithm lies in its complete resolution of the task-equipment mismatch problem, achieving a qualitative leap in resource utilization efficiency. In terms of equipment utilization, under the traditional fixed allocation model, equipment capacity utilization is only 55%, with an idle rate as high as 40%. After adopting this algorithm, equipment capacity utilization increases to over 95%, and the idle rate drops to below 10%. High-end equipment such as autoclaves are no longer used for low-specification tasks, and their core capabilities are fully utilized. In terms of task efficiency, the response time for emergency tasks is shortened from the traditional 30 minutes to less than 5 seconds, and the completion time for ordinary tasks is reduced by an average of 30%. The daily processing capacity of the hospital's sterilization supply center increases from 800 pieces to 1200 pieces.

[0091] From a cost control perspective, precise matching avoids energy waste caused by over-sterilization, reducing the energy consumption of a single operation of the autoclave by 20%, while also reducing instrument damage caused by material incompatibility (the damage rate is 3% in the traditional mode, but reduced to 0.5% by this algorithm). According to calculations by a medium-sized hospital's sterilization supply center, adopting this algorithm can save 150,000 yuan in energy costs annually, reduce instrument replacement costs by 80,000 yuan, and reduce labor costs by 200,000 yuan due to improved task efficiency, resulting in significant overall benefits.

[0092] (II) Path Optimization Algorithm: "Collaborative and Conflict-Free" Cooperation Among Multiple Devices

[0093] 1. Algorithm Principle: Path planning and conflict resolution under time and space constraints

[0094] To address the path conflicts and inefficiencies in multi-device collaborative operations, this algorithm innovatively employs a "spatiotemporal dual-dimensional planning + real-time conflict resolution" strategy. The core of this strategy is to construct a spatiotemporal model of equipment operations, pre-planning conflict-free optimal paths, and simultaneously establishing a dynamic response mechanism to resolve unexpected conflicts in real time, ensuring efficient collaborative operation of multiple devices within a limited space. Its modeling inspiration comes from path optimization technology in logistics and warehousing, combined with the equipment characteristics and operational processes of disinfection production, achieving a transformation from "disorderly collisions" to "orderly collaboration."

[0095] The modeling process is divided into three core stages: workspace modeling, spatiotemporal path planning, and conflict resolution mechanism. In the workspace modeling stage, the system scans the disinfection workshop space using a UWB positioning system (positioning accuracy ±10cm), constructs a 3D digital map, divides functional areas such as equipment operation areas, material conveying areas, and safety buffer zones, and marks the fixed positions and activity ranges of conveyor belts, robotic arms, and disinfection equipment, forming a spatial constraint model. Simultaneously, based on the task allocation results, the system determines the starting point (material grabbing point), ending point (disinfection equipment entrance), and operation time window for each piece of equipment, forming a time constraint model.

[0096] The spatiotemporal path planning stage is the core of the algorithm, generating the optimal path through an improved path optimization algorithm. Unlike traditional path planning that only considers distance, this algorithm constructs a multi-objective cost function, comprehensively considering path length (reducing travel time), energy consumption (reducing equipment operating costs), conflict risk (avoiding intersections with other equipment), and operational efficiency (matching the idle time of disinfection equipment). For example, when the robotic arm transports emergency instruments, the algorithm will prioritize the shortest path but with a slightly higher conflict risk, while coordinating with other equipment to avoid it in advance; while for ordinary material transport, it will choose the path with the lowest energy consumption to ensure optimal cost.

[0097] The conflict resolution mechanism is divided into pre-emptive prevention and in-process resolution. Pre-emptive prevention is achieved through time window allocation, assigning a dedicated time interval to the path of each device. For example, robotic arm A passes through the conveyor belt intersection between 10:00 and 10:02, while robotic arm B is scheduled to pass between 10:03 and 10:05, thus avoiding conflicts at the source. In-process resolution is achieved through real-time monitoring and dynamic adjustment. When the UWB positioning system detects that a device deviates from its planned path or a sudden malfunction causes a potential conflict, the algorithm immediately calculates the conflict level. High-priority tasks (such as emergency equipment) are prioritized to allow other devices to pause and avoid the conflict, while low-priority tasks automatically replan their detour paths, ensuring that the conflict resolution time is ≤0.5 seconds.

[0098] 2. Solution Process: Dynamic Adaptation from Path Planning to Conflict Resolution

[0099] The algorithm's solution process is a dynamic closed loop of "spatial modeling - time allocation - path generation - real-time monitoring - conflict resolution," running continuously throughout the entire equipment operation. Taking a sterile workshop in a pharmaceutical company as an example: the workshop contains two robotic arms (M1, M2), one main conveyor belt (C1), and three sterilization devices (D1, D2, D3). The current task is for M1 to transport material A to D1, and M2 to transport material B to D3. Both require sharing a section of the intersection area on C1.

[0100] After spatial modeling, the algorithm determines the path for M1 as: material area → C1 starting point → intersection area → D1 entrance; and the path for M2 as: material area → C1 branch → intersection area → D3 entrance. During the time allocation phase, considering the idle times of D1 and D3, a time window of 14:00-14:04 is allocated to M1 (D1 idles at 14:05), and 14:06-14:10 is allocated to M2 (D3 idles at 14:10). The passage times to the intersection area are 14:02-14:03 (M1) and 14:08-14:09 (M2), with no time overlap. After path generation, the robotic arm starts operating according to the planned path.

[0101] During operation, D1 experienced a brief malfunction (malfunction at 14:03, restored at 14:04). M1 paused and waited at the end of the junction area, extending its occupation of the junction area to 14:04. The UWB positioning system detected the abnormal position of M1 in real time. The algorithm immediately determined a potential conflict with M2 (M2 arrived as planned at 14:08, without conflict). However, to ensure safety, it sent a command to M1 to accelerate through the remaining path after the malfunction was restored, arriving at D1 at 14:05, without affecting subsequent tasks. If M2 needs to start earlier at 14:05 due to an urgent task, the algorithm will adjust M1's path, causing it to detour through the backup channel of C1 to avoid conflict.

[0102] 3. Principle of Enhanced Efficiency: From "Disorderly Conflicts" to "Smooth Collaboration"

[0103] The core efficiency-enhancing value of this algorithm lies in its complete resolution of path conflict issues in multi-device collaboration, significantly improving operational efficiency and space utilization. In terms of path efficiency, under traditional unplanned operation, the total operational path for multiple devices is lengthy, with the robotic arm's ineffective movement distance accounting for 35% of the total distance. With this algorithm, the total operational path is shortened by more than 35%, ineffective movement distance is reduced to below 5%, and the robotic arm's average operation time is reduced by 40%. Regarding conflict control, the probability of equipment conflict in the traditional mode reaches 15%, with each conflict requiring 2-5 minutes to resolve. This algorithm, through pre-emptive prevention and in-process resolution, reduces the conflict probability to below 1%, the resolution time for sudden conflicts is ≤0.5 seconds, and the mutual waiting time between devices is reduced by 60%.

[0104] From a space utilization perspective, through scientific path planning and functional zoning, the space utilization rate of the disinfection workshop has increased from the traditional 40% to over 60%, increasing processing capacity by 30% without expanding the workshop area. Taking the aseptic disinfection workshop of a pharmaceutical company as an example, after adopting this algorithm, the daily disinfection volume of materials in the workshop increased from 500 batches to 700 batches, equipment utilization increased from 65% to 95%, and the material damage rate due to conflict decreased from 2% to 0.1%, reducing losses by approximately 200,000 yuan annually, while saving several million yuan in workshop expansion costs.

[0105] (III) Parameter Adjustment Algorithm: "Precise and Controllable" Disinfection Effect

[0106] 1. Algorithm Principle: Adaptive Adjustment Driven by Effect Feedback

[0107] To address the issues of disinfection parameters relying on human experience and exhibiting large fluctuations in effectiveness, this algorithm innovatively constructs a closed-loop system of "real-time monitoring - effectiveness evaluation - parameter adjustment." The core of this system is to dynamically adjust parameters such as disinfection power, reagent concentration, and temperature by online monitoring of key indicators during the disinfection process, ensuring stable and compliant disinfection results while avoiding over-disinfection. Its modeling logic is derived from process control theory, using disinfection effectiveness as the core feedback signal to establish a dynamic response relationship between parameters and effects, thus achieving a shift from "experience-based setting" to "data-driven regulation."

[0108] The modeling process is divided into three core stages: disinfection mechanism modeling, real-time effect monitoring, and adaptive parameter adjustment. In the disinfection mechanism modeling stage, a model of the effects of different disinfection methods was established using extensive experimental data. For example, the effectiveness of high-pressure steam sterilization is mainly affected by temperature, pressure, and time; for every 1°C increase in temperature, the kill rate increases by 5%. The effectiveness of hydrogen peroxide plasma sterilization is positively correlated with the agent concentration and plasma power, but the improvement slows down after the concentration exceeds a certain threshold, and it also increases the cost of consumables. Combining the material characteristics of different items (e.g., plastic is not heat-resistant) and the degree of contamination (higher contamination requires higher parameters), a multi-factor correlation model of "parameter-material-contamination degree-effectiveness" was constructed to provide a theoretical basis for parameter adjustment.

[0109] The real-time effect monitoring phase is the foundation of the algorithm. The system monitors the entire disinfection process through online quality detection devices (optical detection + biosensors): optical detection captures images of residual contaminants on the surface of items in real time, biosensors detect changes in microbial concentration in the disinfection environment in real time, and environmental sensors simultaneously collect environmental parameters such as temperature, humidity, and air pressure. This data is transmitted to the intelligent decision-making unit in real time, and the machine learning model evaluates the current disinfection effect to determine whether there is a trend of insufficient disinfection (slow decrease in microbial concentration) or over-disinfection (parameters are too high but the effect is not improved).

[0110] The parameter adaptive adjustment phase is the core of the algorithm, dynamically adjusting parameters based on the effect evaluation results. The adjustment strategy consists of three levels: First, basic adjustment, which retrieves initial parameters from the database based on the initial contamination level and material of the items (e.g., for highly contaminated metal instruments, the initial temperature for high-pressure steam sterilization is 132℃, and the time is 15 minutes); second, real-time adjustment, which increases parameters by a gradient if the biosensor detects that the rate of decrease in microbial concentration is below the threshold (inadequate disinfection trend), and immediately decreases parameters if the concentration has dropped to a safe value but the parameters are still high (over-disinfection trend) (e.g., the temperature drops to 130℃, and the time is shortened to 12 minutes); third, adaptive learning, which stores the final parameters, effect data, and item information in the database after each disinfection, and optimizes the parameter adjustment model through a neural network algorithm to make subsequent adjustments more precise.

[0111] 2. Solution process: Full control from initial parameters to dynamic optimization

[0112] The algorithm's solution process is a complete closed loop of "initial parameter setting - process monitoring - effect evaluation - parameter adjustment - learning optimization," spanning the entire disinfection operation. Taking the disinfection of stainless steel containers in a food processing enterprise as an example (using an ultraviolet disinfection cabinet): the system first identifies the container type (stainless steel) through machine vision, and the contamination detection module determines the degree of contamination (medium contamination, microbial concentration 10). 5CFU), environmental sensors detect the current temperature and humidity (temperature 25℃, humidity 60%).

[0113] The algorithm uses the following basic parameters: UV power 80%, disinfection time 20 minutes. After disinfection begins, the online monitoring device tracks the microbial concentration in real time: within the first 5 minutes, the concentration drops to 10. 4 The CFU (Chemical Fume) was decreasing at a normal rate. At the 6-minute mark, due to a sudden increase in humidity to 75% (due to steam leakage in the workshop), the rate of decrease in microbial concentration slowed, reaching only 5 × 10⁻⁶ at 10 minutes. 3 CFU (lower than expected 10) 3 The effectiveness evaluation module determined that disinfection was insufficient, and the parameter adjustment algorithm responded immediately: increasing the UV power to 90% and extending the disinfection time to 25 minutes. After adjustment, the microbial concentration decreased more rapidly, dropping to 800 CFU (meeting the standard) by the 20th minute. The algorithm determined that there was no need to further increase the parameters, maintaining the current power until the end of the 25-minute period to avoid over-disinfection. After disinfection, the data was stored in the database, and the neural network algorithm analyzed the impact of the humidity fluctuation on the parameters, optimizing the model: if a moderately contaminated stainless steel container with 75% humidity is encountered subsequently, the initial parameters will be automatically set to 85% power and 22 minutes, improving the accuracy of adjustment.

[0114] 3. Principle of Enhanced Efficiency: From "Fluctuating Results" to "Stable Achievement of Targets"

[0115] The core value of this algorithm lies in achieving precise and stable control of disinfection effects while reducing energy consumption and consumable costs. In terms of effect stability, traditional manual parameter settings result in disinfection effect fluctuations of up to 25%, with a batch pass rate of only 80%. Using this algorithm, the effect fluctuation is controlled within 5%, and the batch pass rate increases to over 99.5%, completely resolving the safety risks caused by insufficient disinfection and the resource waste caused by over-disinfection. In terms of parameter accuracy, the algorithm can dynamically adjust parameters according to differences in contamination levels. The disinfection effect compliance rate for highly contaminated items increases from the traditional 70% to 100%, while energy consumption for low-contamination items decreases by an average of 20%.

[0116] In terms of adaptability during product switching, traditional methods require 30 minutes of manual parameter recalibration when changing the type of items to be disinfected, and are prone to errors. This algorithm, through an adaptive learning mechanism, eliminates the need for manual intervention during switching; the system automatically matches the optimal parameters, reducing the switching time to less than one minute and achieving 100% accuracy. Taking a food company's tableware disinfection line as an example, after adopting this algorithm, the disinfection pass rate increased from 82% to 99.8%, the product recall rate due to disinfection issues dropped to zero, and annual energy and material costs were saved by approximately 120,000 yuan, while labor costs were reduced by 60%.

Claims

1. A health disinfection production intelligent collaborative optimization method, characterized in that, The application relates to a disinfection task allocation method and system. The application comprises the following steps: Collecting task information such as the type, quantity, contamination degree, material characteristics and disinfection standard of the to-be-disinfected articles and resource information such as the type, capacity, state and position of the disinfection equipment; Optimal matching of disinfection tasks and equipment is realized through an intelligent allocation algorithm based on task characteristics and equipment capacity matching; An optimal work path without conflict is generated based on a multi-equipment collaborative work path planning and conflict resolution algorithm; A parameter dynamic optimization algorithm based on real-time feedback of disinfection effect is adopted to adjust disinfection parameters in real time and control equipment to perform work; 2. The method of claim 1, wherein, The task-feature-based intelligent distribution algorithm establishes a task-device matching degree calculation model And through multi-objective optimization, the intelligent distribution of tasks is realized.

3. The method of claim 1, wherein, Real-time evaluation of disinfection quality is realized through online monitoring and intelligent analysis, and algorithm parameters are continuously iteratively optimized in combination with historical data, so that full-process closed-loop control is formed.

4. The method of claim 1, wherein, The parameter dynamic optimization algorithm based on the real-time feedback of the disinfection effect establishes a dynamic response model of disinfection effect And the dynamic optimization of parameters is realized through model predictive control.

5. The method of claim 1, wherein, The multi-equipment collaborative work path planning and conflict resolution algorithm constructs a path cost function Cost(r)=alpha.L(r)+beta.T(r)+gamma.R(r)+delta.E(r), and an improved ant colony algorithm is adopted to generate an optimal path. The multi-equipment collaborative work path planning and conflict resolution algorithm establishes a real-time conflict resolution mechanism, and realizes conflict-free work among equipment through a time window method and priority adjustment.

Images