Medical disinfection area environment safety comprehensive monitoring platform

Abstract

The application discloses a medical disinfection area environment safety comprehensive management platform, relates to the medical care information technology field, and comprises the following steps: an adaptive disinfection regulation algorithm carried by an algorithm module is used to analyze data and determine parameters of disinfection operation; disinfection work is performed according to the determined parameters; the environment data and disinfection operation records are encapsulated into data blocks by using encryption hashing and a distributed consensus mechanism, and are stored in a block chain; regional environment risks are dynamically classified according to preset standards, real-time monitoring data after disinfection is compared with preset target values, and parameters of the adaptive disinfection regulation algorithm are corrected according to comparison results. The application provides a comprehensive management platform integrating multi-source data fusion, dynamic weight optimization, nonlinear risk assessment, closed-loop feedback control and safe data storage.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of healthcare information technology, and particularly relates to a comprehensive supervision platform based on medical disinfection area environment safety. BACKGROUND

[0002] The current medical disinfection field generally relies on manual monitoring and traditional timed disinfection methods. The environmental data collection is scattered and single, there is a lack of effective fusion between various sensor data, the environmental risk assessment method is mostly based on static threshold setting, and there is a lack of real-time closed-loop control between disinfection operation and environmental monitoring, which leads to problems such as information asymmetry, response lag and data untraceability in the safety supervision and operation control of the disinfection area. SUMMARY

[0003] In view of the above existing problems, the present application is proposed.

[0004] Therefore, the present application aims to solve the problems of non-integrated multi-source environmental data collection, static risk assessment, lack of closed-loop feedback in operation control and insufficient data record transparency in the prior art. Through multi-sensor data fusion, self-adaptive disinfection control algorithm based on the combination of deep learning and reinforcement learning, dynamic risk assessment grading, closed-loop feedback control and blockchain technology to realize data tamper-proof storage, a comprehensive supervision system covering environmental monitoring, risk assessment, operation scheduling and data management is established, so as to realize dynamic adjustment and efficient scheduling of disinfection operation parameters.

[0005] To solve the above technical problems, the present application provides the following technical solutions. The comprehensive supervision platform for medical disinfection area environment safety comprises:

[0006] Multi-dimensional environmental data is collected in the medical disinfection area through a sensor module, and the multi-dimensional environmental data is preprocessed. Data fusion technology is used to uniformly eliminate noise and redundancy.

[0007] An adaptive disinfection control algorithm carried by an algorithm module is used to analyze the preprocessed multi-dimensional environmental data, determine the parameters of the disinfection operation, including disinfectant concentration, operation path and operation duration.

[0008] The disinfection operation is performed according to the determined parameters, and the multi-dimensional environmental data after disinfection is collected in real time. The multi-dimensional environmental data and disinfection operation records are encapsulated into data blocks using encryption hashing and distributed consensus mechanism, and stored in the blockchain.

[0009] According to the multi-dimensional environmental data, a fusion algorithm calculation module is used to calculate a comprehensive risk index, and according to a preset standard, the regional environmental risk is dynamically classified, and the real-time monitoring data after disinfection is compared with the preset target value, and the parameters of the self-adaptive disinfection control algorithm are corrected according to the comparison result, when the monitoring data exceeds the preset safety threshold, the current disinfection operation is automatically interrupted, and after the multi-dimensional environmental data recovers to the safety range, the automatic restart operation is implemented through step-by-step detection.

[0010] The analysis of the pre-processed multi-dimensional environmental data includes that, in order to realize the weighted fusion of multi-modal data, first, the attention weight is defined as:

[0011]

[0012] Wherein, α i represents the attention weight of the i-th sensor data, w i ∈R d represents the attention weight of the i-th sensor data x i ∈R d The corresponding weight vector is, represents the transpose of w i , exp(·) represents the exponential function, N represents the total number of sensors, d represents the dimension of each sensor data, x j represents the j-th sensor data; represents the transpose of w j ;

[0013] The above attention weight is used to weight and sum each sensor data to obtain the fusion data:

[0014]

[0015] Wherein, represents the data vector after fusion;

[0016] The fusion data is input into a deep neural network controlled by the parameter ξ to obtain the hidden layer feature representation:

[0017] z=f ξ (x)

[0018] Wherein, z∈R m represents the hidden layer feature vector, f ξ represents the deep neural network, ξ represents the parameter of the deep neural network, and m represents the dimension of the hidden layer vector.

[0019] Based on the hidden layer representation, an initial control parameter is generated by using a policy network:

[0020] a0=π φ (z)

[0021] where a0∈R k denotes the initial control parameter, π φ denotes the policy network, φ denotes the parameters of the policy network, and k denotes the dimension of the control parameter.

[0022] At the same time, the auxiliary control parameter is generated for the hidden layer representation z by using the supervision module:

[0023] a h = h ψ (z)

[0024] where a h ∈R k denotes the auxiliary control parameter, h ψ denotes the supervision module, and ψ denotes the parameters of the supervision module.

[0025] The analysis of the preprocessed multi-dimensional environment data further includes introducing an adaptive weight factor to achieve dynamic fusion of the initial and auxiliary parameters, which is defined as:

[0026] β = σ(w T z + b)

[0027] where β ∈ (0, 1) denotes the adaptive fusion weight, σ(·) denotes the Sigmoid function, w ∈ R m denotes the vector for calculating the weight, and b ∈ R denotes the bias term.

[0028] The initial control parameter and the auxiliary control parameter are fused according to the adaptive weight factor to obtain the preliminary control parameter:

[0029] a p = β·a0 + (1-β)·a h

[0030] where a p ∈R k denotes the preliminary control parameter.

[0031] In order to correct the uncertainty of the preliminary control parameter, a Bayesian layer is introduced to model the posterior distribution of the hidden layer features, and an uncertainty correction formula is defined:

[0032] a = a p + δ·σ a

[0033] where a ∈ R k denotes the final control parameter, δ denotes the uncertainty adjustment constant, and σ a ∈R k denotes the standard deviation vector extracted from the Bayesian posterior distribution p(a∣z), i.e., the uncertainty measure.

[0034] The analysis of the pre-processed multi-dimensional environment data further comprises introducing a reinforcement learning module for implementing time difference-based policy evaluation, with a time difference target defined as:

[0035]

[0036] wherein y R represents the time difference target, R(s, a) represents a reward function value corresponding to a state and an action, s represents a current state, g (0, 1) represents a discount factor, a ' represents a candidate action at the next moment, z ' represents an implicit layer feature vector generated by a deep network at the next moment, Q φ represents a Q value function determined by a parameter f; and

[0037] Finally, a multi-objective loss function is constructed to jointly train the entire model, and is defined as:

[0038]

[0039] wherein L R represents a total loss function, represents an error measurement function between a generated final control parameter a and a reference parameter a * , a * R k represents the reference control parameter, g R represents a weight factor of the reinforcement learning error term, h R represents a weight factor of the uncertainty regularization term, and || s a || 2 represents a square of an L2 norm of the uncertainty vector;

[0040] The parameter of the adaptive disinfection regulation algorithm is corrected according to the comparison result, comprising: after the disinfection operation is completed, the system collects real-time disinfection environment data through an environment sensor, denoted as y measured , and compares y measured with preset target environment data y target , to calculate an error signal e, which is defined as:

[0041] e = y target - y measured

[0042] wherein e R represents the error signal, i.e., the difference between the target value and the actual measured value;

[0043] The error signal e is input to a feedback control module, which adopts a proportional-integral-derivative (PID) control algorithm, with a calculation formula as:

[0044]

[0045] wherein denotes a parameter correction amount, is a proportional gain vector, is an integral gain vector, is a differential gain vector, t ∈ R is a current time, τ ∈ [0, t] is an integral variable, and denotes the dimension of the adaptive disinfection control algorithm parameter;

[0046] The parameter correction amount is added to the current parameter to update the new parameter θ new , that is:

[0047] θ new = θ old + Δθ

[0048] where θ old is the parameter set of the current disinfection control algorithm, and the new parameter θ new is used to update the control settings in real time in the disinfection operation, thereby forming a closed-loop feedback system to realize continuous automatic adjustment of the disinfection operation parameters.

[0049] As a preferred scheme of the medical disinfection area environment safety comprehensive supervision platform, the data fusion technology includes applying a filtering algorithm to remove random noise in the original data of temperature and humidity data, air quality data, microbial concentration data, and chemical residue data, respectively, normalizing the data to a unified numerical interval according to the different dimensions and numerical ranges of each sensor data, extracting key feature information from the normalized data through principal component analysis and discrete wavelet transform technology, and fusing the multi-dimensional environmental data according to a predetermined data format to form a unified standard data stream.

[0050] The key feature information includes temperature and humidity data, time sequence characteristics, frequency domain characteristics, correlation characteristics, and discrete characteristics.

[0051] As a preferred scheme of the medical disinfection area environment safety comprehensive supervision platform, the data block encapsulated by the encryption hash and distributed consensus mechanism includes defining a single data record:

[0052] R i = (d i , ω i , t i )

[0053] where R i denotes the i-th data record, and the data record contains multi-dimensional environmental data d i collected by the environmental sensor d i ∈ R pWhere p represents the environmental data dimension, and ω represents the operation records collected by the disinfection equipment. i ω i ∈R q Where q is the dimension of the operation record and t is the timestamp of the record. i ,t i ∈R;

[0054] Generate a data digest, i.e., a hash value:

[0055] h i =H(R) i ||r i ||h i-1 )

[0056] Among them, h i Let H(·) represent the hash value of the i-th data record, H(·) represent the cryptographic hash function, and the symbol || represents the concatenation operation. i Let h represent the random salt generated for the i-th data record. i-1 This represents the hash value of the previous block, with an initial value of h0 defined for the first block;

[0057] Encapsulate data records into data blocks:

[0058] B i =(h i-1 ,h i R i , τ i V i )

[0059] Among them, B i τ represents the i-th data block. i V represents the timestamp of the data block generation. i This is the set of valid digital signatures obtained for the i-th data block.

[0060] As a preferred embodiment of the comprehensive monitoring platform for environmental safety in medical disinfection areas described in this invention, the method of encapsulating data blocks using encrypted hashing and distributed consensus mechanisms further includes defining a set of valid digital signatures generated by each monitoring node in the distributed consensus mechanism.

[0061] V i ={σ i，n |n∈N and Π(σ) i,n )=1}

[0062] Where, σ i,n Indicates that the monitoring node n is responsible for data block B. i The generated digital signature, For the set of regulatory nodes participating in the consensus, is a indicator function, which takes value 1 when the digital signature generated by node n is verified, otherwise takes value 0;

[0063] A dynamic consensus threshold is introduced to determine whether the block consensus is reached:

[0064]

[0065] wherein, τ c represents the minimum number of valid signatures required for consensus, λ1 represents a dynamic adjustment parameter, which takes value in the range of (0, 1] and is used to adjust the consensus strictness, represents the total number of regulatory nodes participating in consensus, symbol represents the ceiling function.

[0066] As a preferred scheme of the medical disinfection area environment safety comprehensive supervision platform, the data block encapsulated by using the encryption hash and the distributed consensus mechanism further comprises determining a block consensus reaching condition:

[0067] |V i |≥τ c

[0068] wherein, |V i | represents the number of valid digital signatures in the set V i , when the number of valid digital signatures is greater than or equal to the consensus threshold τ c , the block B i is considered to pass the consensus and can be stored synchronously among the regulatory nodes;

[0069] A complete block chain structure is constructed:

[0070]

[0071] wherein, represents a block chain structure composed of M data blocks, and M is the total number of data blocks.

[0072] As a preferred scheme of the medical disinfection area environment safety comprehensive supervision platform, the fusion algorithm for calculating the comprehensive risk index comprises: the temperature and humidity data x d , the air quality data x a , the microorganism concentration data x m and the chemical residue data x c are respectively processed by using a minimum-maximum normalization method to obtain normalized data x′ d , x′ a , x′ m and x′ c ; and a dynamic weight optimization module based on a genetic algorithm is used to obtain an initial weight and As input, the weight adjustment function W(x′) is used. d ,x′ a ,x′ m ,x′ c Output the updated weights w d ,w a ,w m and w c Using the weighted fusion formula:

[0073] ε=w d ·x′ d +w a ·x′ a +w m ·x′ m +w c ·x′ c

[0074] The comprehensive risk index ε is calculated; a risk assessment function F is constructed using fuzzy logic control, and a preset risk threshold sequence T1, T2, ..., T is set. o-1 Each value T in the sequence i The risk boundary point is represented by a Bayesian update mechanism, using the prior distribution p(ε) and the likelihood function p(ε|T) corresponding to the observed data. i Using the update formula:

[0075] p(T i |ε)∝p(ε|T i )p(T i )

[0076] The threshold sequence is dynamically adjusted to divide ε into discrete risk levels L1, L2, ..., L o L o This represents the 0th level of risk category; this discrete risk level serves as input data for the dynamic adjustment module of disinfection operation parameters in subsequent steps; p(T i ∣ε) represents the posterior distribution.

[0077] The beneficial effects of this invention are as follows: This invention provides a comprehensive monitoring platform that integrates multi-source data fusion, dynamic weight optimization, nonlinear risk assessment, closed-loop feedback control, and secure data storage. This platform enables disinfection operations to automatically adjust operating parameters based on real-time environmental monitoring data, ensuring real-time correspondence between monitoring information and disinfection execution. This improves the scientific nature and transparency of medical disinfection area management, while also enabling full-process data recording and traceability. This provides an efficient and intelligent comprehensive monitoring technology solution for medical environmental safety. Attached Figure Description

[0078] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description only represent some of the embodiments of the present application, and all other drawings obtained by those skilled in the art without creative effort based on these drawings should also fall within the protection scope of the present application.

[0079] Figure 1 A flowchart of a medical disinfection area environment safety comprehensive monitoring platform provided by an embodiment of the present application is shown. DETAILED DESCRIPTION

[0080] In order to make the above objectives, features and advantages of the present application more apparent, the specific embodiments of the present application will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative effort should also fall within the protection scope of the present application.

[0081] In the following description, many specific details are set forth in order to provide a thorough understanding of the present application. However, the present application can be practiced in other manners different from those described herein, and those skilled in the art can make similar generalizations without departing from the scope of the present application. Therefore, the present application is not limited to the specific embodiments disclosed below.

[0082] Secondly, the term "one embodiment" or "embodiment" as used herein means that the specific features, structures or characteristics can be included in at least one implementation of the present application. The term "in one embodiment" appearing in different places in the specification does not mean the same embodiment, nor is it an embodiment that is independent of or selected from other embodiments.

[0083] The present application is described in detail in combination with the schematic diagrams. In the detailed description of the embodiments of the present application, the cross-sectional view of the device structure is locally enlarged without the general proportion for the convenience of description, and the schematic diagram is only an example, which should not limit the protection scope of the present application herein. In addition, the three-dimensional spatial dimensions of length, width and depth should be included in the actual manufacture.

[0084] Meanwhile, in the description of the present application, it should be noted that the terms "up, down, in and out" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the indicated device or element must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application. In addition, the terms "first, second or third" are only for the purpose of description, and cannot be understood as indicating or implying relative importance.

[0085] Unless otherwise defined, the terms "mounting, connecting, associating" in the present application should be interpreted broadly, for example: it can be fixed connection, detachable connection or integral connection; it can also be mechanical connection, electrical connection or direct connection, it can also be indirectly connected through intermediate medium, or it can be the internal communication of two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0086] Embodiment 1, reference Figure 1 As the first embodiment of the present application, the embodiment provides a medical disinfection area environment safety comprehensive management platform, comprising:

[0087] S1: Collecting multi-dimensional environmental data in the medical disinfection area through a sensor module, and preprocessing the multi-dimensional environmental data, and using data fusion technology to uniformly eliminate noise and redundancy.

[0088] The data fusion technology comprises: applying a filtering algorithm to the raw data of temperature and humidity data, air quality data, microbial concentration data and chemical residue data respectively to remove random noise in the data; for different dimensions and numerical ranges of each sensor data, a normalization method is used to convert the data to a unified numerical interval; and through principal component analysis and discrete wavelet transform technology, key feature information is extracted from the normalized data, and the multi-dimensional environmental data is fused according to a predetermined data format to form a unified standard data stream.

[0089] The key feature information includes temperature and humidity data, time sequence characteristics, frequency domain characteristics, correlation characteristics and discrete characteristics.

[0090] S2: Using an adaptive disinfection control algorithm carried by an algorithm module to analyze the preprocessed multi-dimensional environmental data to determine the parameters of the disinfection operation, including disinfectant concentration, operation path and operation time.

[0091] The analysis of the preprocessed multi-dimensional environmental data comprises: in order to realize the weighted fusion of multi-modal data, first define the attention weight as:

[0092]

[0093] Wherein, alpha i represents the attention weight of the i-th sensor data, w i ∈R d represents the correlation between the i-th sensor data x i ∈R d and the corresponding weight vector, represents the transpose of w i , exp(·) represents the exponential function, N represents the total number of sensors, d represents the dimension of each sensor data, x jrepresents the jth sensor data; represents w j transpose of w

[0094] The fusion data is obtained by weighting and summing each sensor data using the above attention weight:

[0095]

[0096] wherein, represents the data vector after fusion;

[0097] The fusion data is input into a deep neural network controlled by parameter ξ to obtain a hidden layer feature representation:

[0098] z=f ξ (x)

[0099] wherein, z∈R m represents the hidden layer feature vector, f ξ represents a deep neural network, ξ represents the parameters of the deep neural network, and m represents the dimension of the hidden layer vector;

[0100] Based on the hidden layer representation, an initial control parameter is generated using a policy network:

[0101] a0=π φ (z)

[0102] wherein, a0∈R k represents the initial control parameter, π φ represents a policy network, φ represents the parameters of the policy network, and k represents the dimension of the control parameter;

[0103] At the same time, an auxiliary control parameter is generated for the hidden layer representation z using a supervision module:

[0104] a h =h ψ (z)

[0105] wherein, a h ∈R k represents the auxiliary control parameter, h ψ represents a supervision module, and ψ represents the parameters of the supervision module.

[0106] The analysis of the preprocessed multi-dimensional environment data further includes introducing an adaptive weight factor to realize dynamic fusion of the initial and auxiliary parameters, which is defined as:

[0107] β=σ(w T z+b)

[0108] wherein, β∈(0,1) represents the adaptive fusion weight, σ(·) represents a Sigmoid function, w∈R mdenotes a vector for calculating weights, b e R denotes a bias term;

[0109] The initial control parameter and the auxiliary control parameter are fused according to the adaptive weight factor to obtain a preliminary control parameter:

[0110] a p = β · a 0 + (1 - β) · a h

[0111] Wherein, a p e R k denotes a preliminary control parameter;

[0112] In order to correct the uncertainty of the preliminary control parameter, the Bayesian layer is introduced to model the posterior distribution of the hidden layer feature, and the uncertainty correction formula is defined:

[0113] a = a p + δ · σ a

[0114] Wherein, a e R k denotes a final control parameter, δ denotes an uncertainty adjustment constant, σ a e R k denotes a standard deviation vector extracted from the Bayesian posterior distribution p (a|z), that is, an uncertainty measure.

[0115] The analysis of the preprocessed multi-dimensional environment data also includes introducing a reinforcement learning module to realize policy evaluation based on time difference, and the time difference target is defined as:

[0116]

[0117] Wherein, y e R denotes a time difference target, R (s, a) denotes a reward function value corresponding to a state and action pair, s denotes a current state, γ e (0, 1) denotes a discount factor, a ' denotes a candidate action at the next moment, z ' denotes a hidden layer feature vector generated by the deep network at the next moment, Q φ denotes a Q value function determined by the parameter φ;

[0118] Finally, a multi-objective loss function is constructed to jointly train the entire model, which is defined as:

[0119]

[0120] Wherein, L e R denotes a total loss function, denotes an error measure function between the generated final control parameter a and the reference parameter a * , a * e R kdenotes a reference control parameter, λ ∈ R denotes a weight factor of the reinforcement learning error term, η ∈ R denotes a weight factor of the uncertainty regularization term, ||σ a || 2 denotes the square of the L2-norm of the uncertainty vector.

[0121] S3: Perform disinfection operation according to the determined parameters, and collect multi-dimensional environmental data after disinfection in real time, and encapsulate the multi-dimensional environmental data and disinfection operation record into data blocks by using encryption hash and distributed consensus mechanism, and store in the block chain.

[0122] The encapsulation into data blocks by using encryption hash and distributed consensus mechanism includes defining a single data record:

[0123] R i =(d i ,ω i ,t i )

[0124] Wherein, R i denotes the i-th data record, which contains multi-dimensional environmental data d i (d i ∈R p ), wherein p is the dimension of environmental data, operation record ω i (ω i ∈R q ), wherein q is the dimension of operation record, and the timestamp t i (t i ∈R) collected by the disinfection equipment;

[0125] Generate a data digest, that is, a hash value:

[0126] h i =H(R i ||r i ||h i-1 )

[0127] Wherein, h i denotes the hash value of the i-th data record, H(·) denotes an encryption hash function, symbol || denotes a concatenation operation, r i denotes a random salt generated for the i-th data record, h i-1 denotes the hash value of the previous block, and h0 is defined as an initial value for the first block.

[0128] Encapsulate the data record into a data block:

[0129] B i =(h i-1 ,h i ,R i ,τi , V i )

[0130] where B i represents the i-th data block, τ i represents the timestamp of the data block generation, V i is the set of valid digital signatures obtained for the i-th data block.

[0131] The encapsulation of the data block using the encryption hash and the distributed consensus mechanism further comprises defining the set of valid digital signatures generated by each supervision node in the distributed consensus:

[0132] V i = {σ i，n | n∈N and Π(σ i,n ) = 1}

[0133] where σ i,n represents the digital signature generated by the supervision node n for the data block B i , N c is the set of supervision nodes participating in the consensus, and is an indicator function that takes the value 1 when the digital signature generated by the node n passes the verification, otherwise it takes the value 0.

[0134] A dynamic consensus threshold is introduced to determine whether the block consensus is reached:

[0135]

[0136] where τ c represents the minimum number of valid signatures required for consensus, λ1 represents a dynamic adjustment parameter whose value range is (0, 1] and is used to adjust the strictness of the consensus, N i represents the total number of supervision nodes participating in the consensus, and represents the ceiling function.

[0137] The encapsulation of the data block using the encryption hash and the distributed consensus mechanism further comprises determining the block consensus reaching condition:

[0138] |V i | ≥ τ c

[0139] where |V i | represents the number of valid digital signatures in the set V i , when the number of valid digital signatures is greater than or equal to the consensus threshold τ c , the block B i is considered to pass the consensus and can be stored synchronously among the supervision nodes;

[0140] The complete block chain structure is constructed:

[0141]

[0142] wherein, represents a block chain structure composed of M data blocks, M is the total number of data blocks.

[0143] The above formulas describe in turn the whole process from the composition of a single data record, generation of a data digest using an encryption hash algorithm combined with a random salt and the hash value of the previous block, encapsulation into a block, generation of a digital signature among supervisory nodes and determination of block consensus using a dynamic consensus threshold, until the formation of a complete block chain, each formula being the result of the previous formula and each symbol being described in detail.

[0144] S4: Calculate the comprehensive risk index by using the fusion algorithm of the algorithm module according to the multi-dimensional environmental data, dynamically classify the regional environmental risk according to the preset standard, compare the real-time monitoring data after disinfection with the preset target value, correct the parameters of the self-adaptive disinfection control algorithm according to the comparison result, automatically interrupt the current disinfection operation when the monitoring data exceeds the preset safety threshold, and automatically restart the operation through step-by-step detection after the multi-dimensional environmental data returns to the safe range.

[0145] The calculation of the comprehensive risk index by using the fusion algorithm includes that the temperature and humidity data x d (collected by the temperature and humidity sensor), the air quality data x a (collected by the air quality sensor, including PM2.5 concentration, PM10 concentration, carbon dioxide concentration and volatile organic compound concentration), the microbial concentration data x m (collected by the microbial detector, including bacterial concentration, viral concentration and fungal concentration) and the chemical residue data x c (collected by the chemical residue detection sensor, including disinfectant residue concentration and other chemical residue concentration) are respectively processed by the minimum-maximum normalization method to obtain normalized data x′ d , x′ a , x′ m and x′ c ; the dynamic weight optimization module based on genetic algorithm is used to take the initial weights and as input, and output the updated weights w d , w a , w m and w c through the weight adjustment function W(x′ d , x′ a , x′ m , x′ c (where W: R 4 → R 4for the trained weight optimization function); using a weighted fusion formula:

[0146] ε = w d · x' d + w a · x' a + w m · x' m + w c · x' c

[0147] The comprehensive risk index ε is calculated (wherein ε ∈ R represents the fused environmental risk value); the fuzzy logic control method is used to construct the risk assessment function F (wherein F: ε → {L1, L2, …, L n} maps the continuous value ε to the discrete risk level set, and the mapping process uses the local weighted regression method to ensure the smoothness of the mapping function), and a preset risk threshold sequence T1, T2, …, T o-1 (wherein each value T i in the sequence represents a risk boundary point) is set and a Bayesian updating mechanism is used to update the prior distribution p(ε) and the likelihood function p(ε|T i ) using the updating formula

[0148] p(T i |ε) ∝ p(ε|T i ) p(T i )

[0149] The threshold sequence is dynamically adjusted, so that ε is divided into discrete risk levels L1, L2, …, L o (wherein L o represents the oth risk category); the discrete risk levels are used as input data for the disinfection operation parameter dynamic adjustment module in the subsequent steps; the above process from data normalization, dynamic weight optimization, weighted fusion, nonlinear risk mapping to dynamic threshold updating constitutes an environmental risk assessment technology method based on multi-dimensional environmental data fusion; p(T i |ε) represents the posterior distribution.

[0150] The parameters of the adaptive disinfection control algorithm according to the comparison result include that after the disinfection operation is completed, the system collects the post-disinfection environmental data in real time through the environmental sensor, denoted as y measured , and compares y measured with the preset target environmental data y target , and calculates the error signal e, which is defined as:

[0151] e = y target - y measured

[0152] wherein e∈R represents an error signal, i.e. the difference between the target value and the actual measured value;

[0153] The error signal e is input to a feedback control module, which adopts a proportional-integral-derivative (PID) control algorithm, the calculation formula of which is:

[0154]

[0155] wherein, represents a parameter correction amount, is a proportional gain vector, is an integral gain vector, is a differential gain vector, t∈R is the current time, τ∈[0,t] is an integral variable, and represents the dimension of the adaptive disinfection control algorithm parameter;

[0156] The parameter correction amount is added to the current parameter to update a new parameter θ new , i.e.

[0157] θ new = θ old + Δθ

[0158] wherein θ old is the parameter set of the current disinfection control algorithm, and the new parameter θ new is used to update the control setting in real time in the disinfection operation, thereby constituting a closed-loop feedback system and realizing continuous automatic adjustment of the disinfection operation parameters.

[0159] Embodiment 2

[0160] The second embodiment of the present application is different from the previous embodiment in that:

[0161] If the function is realized in the form of a software function unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product, which is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various program code storage media.

[0162] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 means for functionally implementing the steps listed in the flowchart block or blocks.

[0163] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the function specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 means for functionally implementing the steps listed in the flowchart block or blocks.

[0164] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 means for functionally implementing the steps listed in the flowchart block or blocks.

[0165] While the preferred embodiments of the application have been described, additional variations and modifications can be employed by those skilled in the art. Therefore, the appended claims intend to cover all such modifications and variations as fall within the true spirit and scope of the application.

[0166] It is apparent that a person skilled in the art can make various changes and modifications to the application without departing from the spirit and scope thereof. Therefore, if these modifications and changes fall within the scope of the claims and their equivalents, it is intended to include them in the application.

Claims

1. A medical disinfection area environment safety comprehensive monitoring platform based on, characterized in that: The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. wherein α i represents the attention weight of the i-th sensor data, w i ∈ R d represents the i-th sensor data x i ∈ R d corresponding weight vector, represents the transpose of w i , exp(·) represents the exponential function, N represents the total number of sensors, d represents the dimension of each sensor data, X j represents the j-th sensor data; represents the transpose of W j ; The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. wherein represents the fused data vector; The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. z = f ξ (x) wherein, wherein, z ∈ R m represents a hidden layer feature vector, f ξ represents a deep neural network, ξ represents a parameter of the deep neural network, and m represents a dimension of the hidden layer vector The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. a0= π φ (z) where a0∈R k denotes the initial control parameter, π φ denotes the policy network, φ denotes the parameters of the policy network, and k denotes the dimension of the control parameter; The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. a h = h ψ (z) where a h ∈ R k denotes an auxiliary control parameter, h ψ denotes a supervision module, ψ denotes parameters of the supervision module; The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. β = σ(w T z + b) where β ∈ (0, 1) represents an adaptive fusion weight, σ (·) represents a Sigmoid function, w ∈ R m represents a vector for calculating weights, and b ∈ R represents a bias term; The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. a p = β - a0+ (1 - β) - a h wherein a p ∈R k denotes a preliminary control parameter; The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. a = a p + δ · σ a where a ∈ R k denotes the final control parameter, δ denotes an uncertainty regulation constant, σ a ∈ R k denotes the standard deviation vector extracted from the Bayesian posterior distribution p(a|z), i.e. the uncertainty measure; The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. wherein y∈R represents a time difference target, R(s,a) represents a reward function value corresponding to a state and an action, s represents a current state, γ∈(0,1) represents a discount factor, a ' represents a candidate action at a next moment, z ' represents a hidden layer feature vector generated by a deep network at the next moment, Q φ represents a Q value function determined by a parameter φ; The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. L = l(a, a * )+ λ(Q φ (z, a) - y) 2 + η · ||ω a || 2 where L e R denotes the total loss function, l(a, a * ) denotes an error metric function that generates a measure of error between the final control parameter a and the reference parameter a * * a e R k denotes the reference control parameter, l e R denotes a weight factor for the reinforcement learning error term, h e R denotes a weight factor for the uncertainty regularization term, and ||s a || 2 denotes the square of the L2 norm of the uncertainty vector;​ The parameters of the adaptive disinfection control algorithm according to the comparison result include that after the disinfection operation is completed, the system collects the post-disinfection environment data in real time through the environment sensor, denoted as y measured , and compares y measured with the preset target environment data The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. e = y target - y measured The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. wherein, denotes a parameter correction amount, is a proportional gain vector, is an integral gain vector, is a differential gain vector, t e R is a current time, τ e [0, t] is an integral variable, and denotes the dimension of the adaptive disinfection control algorithm parameters; The parameter correction amount is added to the current parameter to update a new parameter θ new That is:​ θ new = θ old + Δθ where θ old is the parameter set of the current disinfection control algorithm, and the new parameter θ new is used to update the control settings in real time in the disinfection operation, thereby forming a closed-loop feedback system to achieve continuous automatic adjustment of the disinfection operation parameters.

2. The medical treatment based disinfection area environment safety integrated management platform according to claim 1, characterized in that: The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition and processing system. The application relates to a medical disinfection area multi-dimensional environment data acquisition The key feature information includes timing features, frequency domain features, correlation features and discrete features.

3. The medical treatment disinfection area environment safety integrated management platform based on claim 2, characterized in that: The encapsulation into data blocks by using encryption hash and distributed consensus mechanism includes defining a single data record: R i = (d i , ω i , t i ) Among them, R i This represents the i-th data record, which contains multidimensional environmental data d collected by environmental sensors. i d i ∈R p Where p represents the environmental data dimension, and ω represents the operation records collected by the disinfection equipment. i ω i ∈R q Where q is the dimension of the operation record and t is the timestamp of the record. i ,t i ∈R; Generating a data digest, i.e. a hash value: h i = H(R i ||r i ||h i-1 ) where h i denotes the hash value of the i-th data record, H(·) denotes the cryptographic hash function, the symbol || denotes the concatenation operation, r i denotes the random salt generated for the i-th data record, h i-1 denotes the hash value of the previous block, with the initial value h0defined for the first block; Encapsulating the data record into a data block: B i = (h i-1 ,h i , R i , τ i , V i ) wherein B i represents the i-th data block, τ i represents the timestamp of the data block generation, V i is the set of valid digital signatures obtained for the i-th data block.

4. The medical treatment disinfection area environment safety integrated management platform based on claim 3, characterized in that: The encapsulation into data blocks by using encryption hash and distributed consensus mechanism further includes defining a valid digital signature set of the i th data block in the distributed consensus: V i = {σ i，n | n ∈ N and Π(σ i，n ) = 1} wherein σ i,n represents a supervisory node n, a data block B i a digital signature generated by the supervisory node n, a set of supervisory nodes participating in consensus, an indicator function, taking a value of 1 when the digital signature generated by the node n passes verification, and otherwise taking a value of 0; A dynamic consensus threshold is introduced to determine whether the block consensus is reached: wherein τ c represents the minimum number of valid signatures required for consensus, λ1represents a dynamic adjustment parameter, which is in the range of (0, 1] and is used to adjust the strictness of consensus, represents the total number of regulatory nodes participating in consensus, and the symbol represents a ceiling function.

5. The medical treatment disinfection area environment safety integrated management platform based on claim 4, characterized in that: The encapsulating as data blocks by using the encryption hash and the distributed consensus mechanism further includes determining a block consensus reaching condition: |V i |≥τ c wherein |V i represents the number of valid digital signatures in the set V i When the number of valid digital signatures is greater than or equal to the consensus threshold τ c , the block B i is considered to pass the consensus and can be stored synchronously among the supervision nodes. Building a complete block chain structure: wherein, represents a blockchain structure composed of M data blocks, M being the total number of data blocks.

6. The medical treatment disinfection area environment safety integrated management platform based on claim 5, characterized in that: The comprehensive risk index calculated by the fusion algorithm module includes, for temperature and humidity data x d , air quality data x a , microorganism concentration data x m , and chemical residue data x c , normalized data x' is obtained by using a minimum-maximum normalization method respectively d , x' a , x' m , and x' c ; an initial weight and is input into a dynamic weight optimization module based on a genetic algorithm, and updated weights w d , w a , w m , and w c are output by a weight adjustment function W(x' d , x' a , x' m , x' c ); and a weighted fusion formula is used ε = w d • x' d + w a • x' a + w m • x' m + w c • x' c The comprehensive risk index ε is calculated: A fuzzy logic control method is used to construct a risk assessment function F, and a preset risk threshold sequence T1, T2,..., T o-1 is set, where each value T i in the sequence represents a risk demarcation point, and a Bayesian updating mechanism is used with a prior distribution p(ε) and a likelihood function p(ε|T i ) corresponding to the observation data. An updating formula is used: p(T i |ε) ∝ p(ε | T i )p(T i ) The threshold sequence is dynamically adjusted, thereby dividing ε into discrete risk levels L1, L2,..., L o where L o represents the oth risk class; p(T i | ε) represents the posterior distribution.

Images