Active and passive integrated clean room environment intelligent monitoring method

Abstract

The invention belongs to the field of intelligent monitoring, and discloses an active and passive fusion clean room environment intelligent monitoring method, which constructs an active and passive fusion monitoring system based on covariable statistics. A two-parameter variation function is established by introducing real-time wind field difference as a key covariable, and the defect that a traditional model ignores flow field dynamics is overcome; a comprehensive blind area guiding index is constructed by using a deviation between a basic information field generated by a fixed sensor and a correction information field formed by robot fusion data and combining an inherent coverage risk and a joint space disparity degree, and an absolute monitoring blind area is accurately identified; therefore, the control mode of fixed route inspection to blind area driving type active detection is converted. According to the method, on the premise that the hardware cost is not increased, the full-domain monitoring coverage rate, precision and sudden pollution response speed of the clean room are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent monitoring, specifically a method for intelligent monitoring of cleanroom environments that integrates active and passive monitoring. Background Technology

[0002] Cleanrooms are widely used in high-end manufacturing fields such as semiconductors and biomedicine. Their cleanliness is affected by the nonlinear coupling of multiple factors, including airflow organization, dust source distribution, personnel activities, and equipment operation, which directly affects product yield and production safety. Achieving high-precision, spatiotemporally continuous, and comprehensive environmental monitoring is crucial for revealing the pollution coupling mechanism and constructing dynamic optimization models.

[0003] Current mainstream methods rely on fixed sensor networks, such as the patent application number CN202310254146.8. However, due to high costs and sparse deployment, these methods struggle to cover the entire space and cannot effectively detect areas without sensors, resulting in significant blind spots. To expand coverage, some solutions employ fixed-path mobile robots for inspection (such as the patent application number 202410519256.7), but these suffer from rigid paths, slow response, and a tendency to miss dynamic pollution sources. While autonomous mobile robots (such as the patent application number CN202511217004.X) can freely plan their paths, they remain merely "mobile single-point sensors," lacking global real-time perception capabilities and struggling to overcome the "spatiotemporal coupling" problem in large-scale, highly time-varying flow fields.

[0004] In summary, existing technologies are either limited by static deployment or constrained by localized motion sensing, and neither can simultaneously achieve broad coverage, fast response, and global accuracy. There is an urgent need for an intelligent monitoring method that integrates active and passive monitoring to achieve efficient, continuous, and comprehensive sensing of the cleanroom environment. Summary of the Invention

[0005] Therefore, to address the aforementioned shortcomings, this invention provides an intelligent monitoring method for cleanroom environments that integrates active and passive monitoring. This method constructs an active-passive integrated monitoring system based on covariate geostatistics: by introducing real-time wind field differences as a key covariate to construct a two-parameter variogram function, it solves the problem of traditional models neglecting flow field dynamics characteristics; furthermore, by utilizing the monitoring deviation between the "basic information field" generated by fixed sensors and the "corrected information field" of fused robot data, combined with inherent coverage risks and joint spatial alienation, a comprehensive blind zone guidance index is constructed to accurately locate absolute blind zones far from all monitoring nodes; finally, based on this index, a shift from fixed-route inspection to "blind zone-driven active detection" control mode is achieved, significantly improving the overall coverage, accuracy, and response speed to sudden pollution in cleanroom environmental monitoring without increasing hardware costs.

[0006] Specifically, a method for intelligent monitoring of cleanroom environments that integrates active and passive monitoring includes the following steps: S1. Construct a monitoring system, which includes a fixed sensor array and a robot. The fixed sensor array has multiple fixed sensors that transmit data to the central processing unit in real time in one direction. The robot carries the sensors and the central processing unit, transmits data to the central processing unit in real time via a wireless network, and receives the optimal inspection path instructions issued by the central processing unit. S2, Input fixed sensor array position information S =[ S 1, S 2,…, S N Information on obstacles in cleanrooms; Initialize the robot's contaminant dataset C robot Flow field dataset V robot Pollutant datasets from fixed sensor arrays C fixed Flow field dataset V fixed Blind spot dataset I Inspection location dataset I D Basic Information Field C base With modified information field C cor ; Initialize the robot's step size L and the robot's inspection starting point. R S Inspection endpoint R E Blind zone threshold I min ; S3, the robot starts from the inspection point. R S The robot starts moving along the currently set route; simultaneously, a fixed sensor array collects and records data in real time. t Position coordinates at time R ( t Pollutant information C robot ( t ) and flow field information V robot ( t ), and update the robot's contaminant dataset accordingly. C robot Flow field dataset V robot Pollutant information collected by a fixed sensor array in the middle. C fixed ( S i, t ) and flow field information V fixed ( S i , t Update the pollutant dataset to the fixed sensors respectively. Flow field dataset ; S4. Calculate the basic information field of the time based on the data collected in step S3. C base ( t ) and modified information field C cor ( t ), and update to the basic information field. C base With modified information field C cor In this process, a dual information field is constructed; S5. By comparing the dual information fields calculated in step S4, calculate the cleanroom. t Comprehensive blind spot guidance indicators at various locations at any given time I ( t And update to the blind spot dataset. middle; S6. The comprehensive blind spot guidance index I ( t ) and the preset blind zone threshold I min Compare the results to determine whether the robot needs to execute step S3 or step S7.

[0007] S7, Traversing the Blind Spot Dataset I The sequence of numbers filters out spatial location coordinates to form a candidate target set; S8. Calculate the Euclidean distance between each coordinate point in the candidate target set and the robot's current position, select the closest coordinate point as the robot's next moving target position; drive the robot to move to that position, and store the coordinates of that position in the inspection position dataset. I D middle; S9. Determine whether the robot's current position requires continued inspection. The step that needs to be performed in step S6 is either step S3 or step S7.

[0008] S10. Based on the inspection location dataset I DThe historical coordinates are dynamically updated to the robot's preset fixed inspection route, and the location coordinates in the dataset are embedded into the fixed route as new routine inspection nodes; then the state is reset, and the process jumps back to step S2 to start the monitoring of the next cycle.

[0009] The present invention has the following advantages: This invention is an intelligent cleanroom environmental monitoring method that integrates active and passive monitoring. It overcomes the limitations of existing cleanroom environmental monitoring technologies, such as the limited coverage of fixed sensor networks due to sparse locations, high costs of high-density deployment, and the difficulty in accurately capturing sudden local contamination in complex flow fields using traditional interpolation. Existing mobile inspection technologies, lacking guidance from overall environmental situational information, suffer from blind path planning, inability to respond to environmental changes in real time, and difficulty in effectively filling monitoring blind spots. This invention introduces wind field differences as a covariate to construct a two-parameter variogram model. It utilizes multi-source data fusion from fixed sensors and mobile robots to establish a basic information field and a corrected information field. Based on monitoring deviation, inherent uncertainty, and spatial alienation, it calculates a comprehensive blind spot guidance index, thereby driving the mobile robot to actively detect high-risk blind spots and dynamically optimize paths. This achieves low-cost, high-precision, and adaptive full-area cleanroom environmental monitoring. Attached Figure Description

[0010] Figure 1 This is a flowchart illustrating the active-passive integrated intelligent monitoring method for cleanroom environments described in this invention. Detailed Implementation

[0011] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0012] In this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, without necessarily requiring or implying any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus.

[0013] As described in the background section, in existing cleanroom environmental monitoring technologies, fixed sensor networks have limited coverage due to sparse locations, high costs for high-density deployment, and difficulty in accurately capturing localized sudden pollution in complex flow fields by relying on traditional interpolation. Meanwhile, existing mobile inspection technologies lack guidance from overall environmental situation information, resulting in blind monitoring path planning, inability to respond to environmental changes in real time, and difficulty in effectively filling monitoring blind spots.

[0014] To address this problem, this embodiment provides a method for intelligent monitoring of cleanroom environments that integrates active and passive monitoring, such as... Figure 1 As shown, the method includes the following steps: S1. Construct a monitoring system, which includes a fixed sensor array and a robot. The fixed sensor array has multiple fixed sensors that transmit data to the central processing unit in real time in one direction. The robot carries the sensors and the central processing unit, transmits data to the central processing unit in real time via a wireless network, and receives the optimal inspection path instructions issued by the central processing unit. The central processing unit is responsible for receiving all data and executing the core algorithm calculations described in steps S3 to S10.

[0015] S2, Input fixed sensor array position information S =[ S 1, S 2,…, S N Information on obstacles in cleanrooms; Initialize the robot's contaminant dataset C robot Flow field dataset V robot Pollutant datasets from fixed sensor arrays C fixed Flow field dataset V fixed Blind spot dataset I Inspection location dataset I D Basic Information Field C base With modified information field C cor ; Initialize the robot's step size L and the robot's inspection starting point. R S Inspection endpoint R E Blind zone threshold I min .

[0016] Parameter configuration can be achieved through step S2.

[0017] S3, the robot starts from the inspection point. R S The robot starts moving along the currently set route; simultaneously, a fixed sensor array collects and records data in real time. t Position coordinates at time R ( t Pollutant information C robot ( t ) and flow field information V robot ( t ), and update the robot's contaminant dataset accordingly. C robot Flow field dataset V robot Pollutant information collected by a fixed sensor array in the middle. C fixed ( S i , t ) and flow field information V fixed ( S i , t Update the pollutant dataset to the fixed sensors respectively. Flow field dataset ; Data collection and updating can be achieved through step S3.

[0018] S4. Calculate based on the data collected in step S3. t The basic information field of time C base ( t ) and modified information field C cor ( t ), and update to the basic information field. C base With modified information field C cor In this process, a dual information field is constructed. This dual information field construction can be achieved through step S4.

[0019] Calculate in step S4 t The basic information field of time C base ( t ) and modified information field C cor ( t The methods include: Step S4.1: Calculate using the following formula t Effective distance between any two fixed sensors at any given timeh eff ( S i , S j ,t ), ; In the formula S i ,S j ∈ S , S This indicates the position information of a fixed sensor array. S i It is the first i Fixed sensor position vector, S j yes j A fixed sensor position vector, and i Not equal to j , V fixed ( S i ,t ), V fixed ( S j ,t )∈ V fixed , V fixed This represents a flow field dataset from a fixed sensor. V fixed ( S i ,t ) is the first i A fixed sensor in t The velocity vector of the flow field at time t. V fixed ( S j , t ) is the first j A fixed sensor in t The velocity vector of the flow field at time t; C fixed ( S i ,t ), C fixed ( S j ,t )∈ C fixed , C fixedThis represents a pollutant dataset from a fixed sensor. C fixed ( S i ,t ) is the first i A fixed sensor in t Pollutant data at any given time, C fixed ( S j ,t ) is the first j A fixed sensor in t Pollutant data at any given time; S i S j Point S i With point S j The vector, Indicates the distance between two points. α This is a correction factor; Step S4.2: Set the effective distance h eff ( S i ,S j ,t The range of values ​​for ) is divided into K The intervals are as follows: h 1 , h 2 , ..., h k , h k It is the first K The effective distance of each interval is determined by the following formula: ; ; in , N for S Number of elements in the dataset; Step S4.3: Calculate the empirical variation value within different distance ranges using the following formula: ; in: N ( h k ) is the first K Logarithm of data points within a range ξ ( h k , t) is the empirical variation value. Step S4.4: Fit the empirical variance value calculated in step S4.3 to... t Empirical variation function of cleanroom at all times ξ ( Si , Sj , t ); Step S4.5: Construct the mutation value matrix and solve for the optimal weight vector. W With Lagrange multiplier Its defining formula is: ; in ξ i,j The matrix of variability values ​​among all sensors, ξ i,p for P Points and all sensors S i Variations between them; Step S4.6: Calculation t time, P Basic Information C base ( P, t ) and uncertainty indicators U base ( P, t ): C base ( P, t )= W · C fixed ( t ); ; in, ; Step S4.7: Repeat step S4.6 to calculate. t Time-based basic information field C base ( t ) and uncertainty field U base ( t ).

[0020] Step S4.8: Integrate the robot's position information, flow field information, and pollutant concentration information into the original fixed sensor dataset to construct an expanded sensor set. N represents the number of fixed sensors. R ( t )express t The location of the robot at all times is to... tThe position information of the robot and the fixed sensors is stored in a single data sequence for easy data processing. ; ; in, C robot ( t Robots are t Pollutant information at any given time V robot ( t Robots are t Flow field information at any given time C fixed,robot It is the merged pollutant dataset. V fixed,robot This is the fused flow field dataset; and steps S4.1 to S4.7 are repeated to calculate the corrected information field. C cor ( t ) and uncertainty field U cor ( t ).

[0021] S5. By comparing the dual information fields calculated in step S4, calculate the cleanroom. t Comprehensive blind spot guidance indicators at various locations at any given time I ( t And update to the blind spot dataset. In the middle. Step S5 can complete the comprehensive blind spot determination.

[0022] The calculation method for the comprehensive blind spot guidance index is as follows: ; In the formula Δ C ( t The deviation between the base field and the correction field is represented by . The larger the deviation, the less accurate the fixed sensor's monitoring at that point is, and the greater the possibility of a blind zone. The formula for calculating this deviation is: ; ω 1 represents the difference weight, indicating the system's sensitivity to correcting real-time monitoring deviations. ω 2 represents the inherent weight, indicating the system's focus on exploring inherent coverage blind spots.

[0023] S6. The comprehensive blind spot guidance index I ( t ) and the preset blind zone threshold I min Comparison: like I ( t )< Imin If the current area monitoring is deemed valid, the robot continues to move along the currently set route and executes step S3. like I ( t )≥ I min If a monitoring blind spot is detected, the active detection mechanism is triggered, and step S7 is executed.

[0024] S7. Candidate blind zone filtering: Traverse the blind zone dataset and filter the spatial location coordinates that simultaneously meet the following two conditions to form a candidate target set: A. The blind spot guidance indicator value at this location is greater than the preset threshold. I min ; B. This position is in the area in front of the robot's current direction of movement.

[0025] S8. Dynamic Target Execution: Calculate the Euclidean distance between each coordinate point in the candidate target set and the robot's current position, select the closest coordinate point as the robot's next moving target position, drive the robot to move to that position, and store the coordinates of that position in the inspection position dataset. I D middle; S9. Determine whether the robot still needs to continue the inspection at its current position.

[0026] For example, the specific method of step S9 is: determine the robot's current position relative to the inspection endpoint. R E Or the starting point of the inspection R S Is the distance D less than or equal to 5L, where L is the robot's step size? If D≤5L, the robot proceeds to the inspection endpoint. R E Or the starting point of the inspection R S And execute step S10; If D > 5L, then proceed to step S3.

[0027] S10. Based on the inspection location dataset I D The system dynamically updates the historical coordinates and the robot's preset fixed inspection route, embedding the location coordinates in the dataset as new routine inspection nodes into the fixed route; then, it resets the state, jumps back to step S2, and starts the monitoring for the next cycle. This step achieves path correction and iteration.

[0028] The above steps S1-S10 overcome the limitations of existing cleanroom environmental monitoring technologies. Fixed sensor networks suffer from limited coverage due to sparse locations, high costs for high-density deployment, and difficulty in accurately capturing sudden localized contamination in complex flow fields using traditional interpolation. Existing mobile inspection technologies, lacking guidance from overall environmental situational information, suffer from blind path planning, inability to respond to environmental changes in real time, and difficulty in effectively filling monitoring blind spots. This invention introduces wind field differences as a covariate to construct a two-parameter variogram model. It utilizes multi-source data fusion from fixed sensors and mobile robots to establish a basic information field and a corrected information field. Based on monitoring deviation, inherent uncertainty, and spatial alienation, it calculates a comprehensive blind spot guidance index, thereby driving the mobile robot to actively detect high-risk blind spots and dynamically optimize paths. This achieves low-cost, high-precision, and adaptive cleanroom environmental monitoring across the entire area.

[0029] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for intelligent monitoring of a clean room environment with active and passive fusion, characterized by, The method comprises the following steps: S1, constructing a monitoring system comprising a fixed sensor array and a robot, wherein the fixed sensor array comprises a plurality of fixed sensors for transmitting data to a central processing unit in real time in one direction; the robot carries a sensor and a central processing unit, transmits data to the central processing unit in real time through a wireless network, and receives an optimal inspection path instruction issued by the central processing unit; S2, input fixed sensor array position information S [ S 1, S 2,…, S N ], clean room obstacle information; Initialization of the robot's pollutant dataset C robot , flow field dataset V robot , pollutant dataset of a fixed sensor array C fixed , flow field dataset V fixed , blind zone dataset I , inspection location dataset I D , base information field C base and correction information field C cor ; Initialize the moving step length L of the robot, the starting point of the inspection of the robot R S , the ending point of the inspection R E , the blind area threshold I min ; S3, the robot starts from the inspection starting point R S The robot moves along the current set route; at the same time, the fixed sensor array collects real-time records of the robot t The position coordinates of the engraving R ( t ), pollutant information C robot ( t ) and flow field information V robot ( t ), and respectively update the pollutant dataset C robot of the robot, the flow field dataset V robot ; the pollutant information collected by the fixed sensor array C fixed ( S i , t ) and flow field information V fixed ( S i , t ) are respectively updated to the pollutant dataset , the flow field dataset ； S4. Calculate the data collected according to step S3 t time-based information field C base ( t ) and the correction information field C cor ( t ), and update to the basic information field C base and the correction information field C cor , and build a double information field; S5. Calculate the clean room by comparing the two information fields calculated in step S4 t the integrated blind area guidance index for each position at the moment I ( t ) and update to the blind area dataset in the blind area dataset S6、compare the comprehensive blind area guidance index with a preset blind area threshold value I ( t ) with a preset blind area threshold value I min to determine the steps that the robot needs to perform.

2. The method of claim 1, wherein the method is a method of monitoring a cleanroom environment with active and passive fusion, and wherein the method further comprises: The method further comprises: S7, traversing the blind area dataset I The spatial position coordinates are screened out from the number series to form a candidate target set. S8, calculate the Euclidean distance between each coordinate point in the candidate target set and the current position of the robot, select the coordinate point with the shortest distance as the next moving target position of the robot; drive the robot to move to the position, and store the position coordinates in the inspection position data set I D In some embodiments, the method further comprises: S9, judging whether the robot needs to continue to perform inspection at the current position; The step that needs to be performed in the step S6 is the step S3 or the step S7.

3. The method of claim 2, wherein the method is a method of monitoring a cleanroom environment with active and passive fusion, and wherein the method further comprises: The method further comprises: S10、according to the inspection position dataset I D According to the historical coordinates, the fixed inspection route preset for the robot is dynamically updated, and the position coordinates in the dataset are embedded into the fixed route as new normalized inspection nodes; then the state is reset, and the monitoring of the next cycle is started.

4. The method of claim 2, wherein the method is a method of monitoring a cleanroom environment with active and passive fusion, and wherein the method further comprises: The method of calculating the basis information field of the time in step S4 C base ( t ) and the correction information field C cor ( t ) includes: Step S4.1: Calculate by the following equation t Effective distance between any two fixed sensors at time t h eff ( S i ,S j , t ), ； wherein S i ,S j ∈ S , S represents fixed sensor array position information, S i is the i fixed sensor position vector, S j is the j fixed sensor position vector, and i is not equal to j , V fixed S i ,t , V fixed S j ,t ∈ V fixed , V fixed represents a fixed sensor flow field data set, V fixed S i ,t is the flow field velocity vector of the i fixed sensor at the t time instant, V fixed S j ,t is the flow field velocity vector of the j fixed sensor at the t time instant; C fixed S i ,t , C fixed S j ,t ∈ C fixed , C fixed represents a fixed sensor pollutant data set, C fixed S i ,t is the pollutant data of the i fixed sensor at the t time instant, C fixed ​​​​​​​​ S j ,t ) is the first j fixed sensor at t time t; S i S j denotes the vector S i between the points S j , and denotes the distance between the two points, and a is a correction factor. Step S4.2: determining the effective distance h eff ( S i ,S j ,t ) is divided into K intervals, respectively h 1 , h 2 , …, h k , h k is the effective distance of the K interval, K the determination formula is: ； ； wherein: , N is S number of elements in the data set; Step S4.3: calculating the experience variation value in different distance ranges through the following formula: ； where: N h k is the number of data points in the K ξ h k t is the empirical variation value,​​​​ Step S4.4: fitting the empirical variance value calculated according to step S4.3 to t The empirical variance function of a clean room at a time ξ Si Sj,t ;​​ Step S4.5: Constructing the variation value matrix, solving the optimal weight vector W with the Lagrange multiplier whose determination formula is: ； wherein ξ i,j is the matrix of variation values between all sensors, ξ i,p is the matrix of variation values between all sensors, P point and all sensors S i between the point and all sensors. Step S4.6: Calculate t Time, P Point-based information C base ( P,t ) with uncertainty index U base ( P,t ): C base ( P,t )= W · C fixed （ t ）； ； wherein: , Step S4.7: Repeat step S4.6 to calculate t time-based information field C base ( t ) and uncertainty field U base ( t ) containing the base information and uncertainty indicators for all locations indoors; Step S4.8: fuse the position information of the robot, the flow field information and the pollutant concentration information to the original fixed sensor dataset, and construct an extended sensor set ; ； ； wherein, C robot t ) the robot is t contaminant information at time t, V robot t ) the robot is t flow field information at time t, N denotes the number of fixed sensors, R t ) denotes the position of the robot at time t, C fixed,robot is the fused contaminant dataset, V fixed,robot is the fused flow field dataset; and repeating steps S4.1 ~ S4.7 to calculate the correction information field C cor t ) and the uncertainty field U cor t ).​​​​​ 5. The method of claim 4, wherein the method further comprises: The calculation method of the comprehensive blind area guiding index in the step S5 is as follows: ； where Δ C ( t ) represents the deviation of the modified field from the base field, and is calculated as: ; ω 1 is a difference weight, indicating the system's sensitivity to real-time monitoring deviation correction, ω 2 is an inherent weight, indicating the system's exploration attention to inherent coverage blind area.

6. The method of claim 2, wherein the method is a method of monitoring a cleanroom environment with active and passive fusion, and wherein the method further comprises: In the step S6, the comprehensive blind area guiding index is compared with a preset blind area threshold value, and the specific method for judging whether the robot performs the step S3 or the step S7 is as follows: comparing the integrated blind area guidance indicator I ( t ) to a preset blind area threshold I min ​ If I ( t )< I min then it is determined that the current zone monitoring is valid, and the robot continues to move according to the current set route and performs step S3; If I ( t )≥ I min then it is determined that there is a monitoring blind spot, a proactive probing mechanism is triggered, and step S7 is performed.

7. The method of claim 2, wherein the method is a method of monitoring a cleanroom environment with active and passive fusion, and wherein the method further comprises: determining a location of the mobile device; and determining a location of the mobile device. In the step S7, the spatial position coordinates screened out by the number sequence need to satisfy the following two conditions at the same time: Condition 1, the blind area guidance index value of the position is greater than a preset threshold I min ; Condition 2: the position is in the front area of the current motion direction of the robot.

8. The method of claim 2, wherein the method is a method of monitoring a cleanroom environment with active and passive fusion, and wherein the method further comprises: In the step S9, the specific method for judging whether the robot needs to continue to perform inspection at the current position is as follows: whether the distance D between the current position of the robot and the inspection end point R E or the inspection start point R S is less than or equal to 5L, L being the robot movement step length: If D < 5L, the robot goes to the end of the inspection R E or the start of the inspection R S and performs step S10; If D>5L, the step S3 is performed.

Images