Application method of ultrahigh frequency RFID technology in operating room environment monitoring
Through ultra-high frequency RFID technology wireless network and machine learning algorithms, the problems of real-time monitoring and refined control in the operating room environmental monitoring system are solved, real-time monitoring of operating room equipment and items and refined management of environmental parameters are realized. Improves the safety and environmental quality of the surgery.
Patent Information
- Application Number
- CN202411009274.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-26
- Publication Date
- 2024-11-15
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing operating room environmental monitoring system lacks real-time equipment and item monitoring functions, and cannot provide precise location information, resulting in the risk of loss, misplacement or misuse of instruments during the operation. The environmental adjustment method is single, which cannot meet the refinement of different types of surgery. Required, high maintenance cost.
Using ultra-high frequency RFID technology to establish a secure wireless network system, ensure data security through multiple authentication and biometric identification, use machine learning algorithms to perform data filtering and environmental state modeling, monitor and adjust operating room environmental parameters in real time, and dynamically adjust RFID reading The signal strength and frequency of the device can realize real-time monitoring and refined control of equipment and items in the operating room.
Real-time monitoring of equipment and items in the operating room is realized, ensuring accurate positioning and use of equipment, reducing delays and errors, improving the smooth progress and safety of the operation, providing refined environmental control, and reducing maintenance complexity and cost.
Smart Images

Figure CN118966262A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to an application method of ultra-high frequency RFID technology, in particular to an application method of ultra-high frequency RFID technology in operating room environment monitoring. Background Art
[0002] Chinese patent publication number CN116293996A discloses an adaptive operating room environment monitoring system, which improves the automation of environmental control by automatically adjusting the temperature and humidity environment of the operating room, reduces the omissions of manual operation, and facilitates the head nurse to monitor the environmental information of the operating room in real time. In addition, the system enhances the stability of cable connection through the interface design of the annular airbag and the serial port server, avoids the risk of cable falling off under stress, and improves the long-term operation stability of the system as a whole. However, although the patent has its advantages in environmental control and hardware connection stability, its application method in operating room environment monitoring still has some shortcomings and disadvantages, especially in achieving comprehensive operating room monitoring and management. First, the system of CN116293996A mainly focuses on temperature and humidity environment control and hardware connection stability, and does not involve real-time monitoring of equipment and items in the operating room. This is a major defect in practical applications, because the management of equipment and items in the operating room is crucial to the smooth progress and safety of the operation. The application of UHF RFID technology in operating room environment monitoring can track and monitor the location and usage status of surgical instruments, drugs and other key items in real time, ensure the correct use and management of these items during surgery, and reduce surgical delays or errors caused by improper management of equipment or items. However, the system of CN116293996A lacks such real-time monitoring and management functions, and cannot provide accurate location information about surgical instruments and items, resulting in the risk of loss, misplacement or misuse of instruments during surgery.
[0003] Secondly, although the system of CN116293996A can automatically adjust the temperature and humidity, the adjustment method is relatively simple, mainly achieved through the cooperation of temperature sensors and fans, and cannot meet the refined requirements for the environment during surgery. Different types of surgeries have different requirements for environmental temperature and humidity. During surgery, environmental parameters need to be adjusted in real time according to the specific conditions of the patient. The application method of ultra-high frequency RFID technology in operating room environmental monitoring can combine a variety of sensor data, and through advanced data processing and optimization algorithms, adjust and optimize the environmental parameters of the operating room in real time, providing more accurate and personalized environmental control. In addition, the RFID system can also integrate environmental monitoring functions, comprehensively manage multi-dimensional environmental parameters such as temperature, humidity, and air quality, and further improve the quality of the surgical environment. Furthermore, although the system of CN116293996A has certain innovations in hardware connection stability, its design is still relatively complex and the maintenance cost is high. The system achieves the stability and heat dissipation effect of cable connection through annular airbags and mercury-driven piston rods. These components will wear or fail during long-term use and need regular maintenance and replacement. The application method of UHF RFID technology in operating room environment monitoring mainly relies on wireless technology for data transmission, which reduces the complexity and maintenance cost of physical connection. The installation and maintenance of RFID system are easier, reducing the clutter of cables in the operating room and improving the cleanliness and safety of the working environment. In addition, the wireless data transmission technology of RFID system can also avoid the interference and signal attenuation caused by physical connection, ensuring the stability and reliability of data transmission. Summary of the invention
[0004] The purpose of the present invention is to provide an application method of ultra-high frequency RFID technology in operating room environment monitoring, thereby solving some of the drawbacks and shortcomings pointed out in the background technology.
[0005] The present invention solves the above-mentioned technical problems by adopting the following technical solutions:
[0006] S1. First, a secure wireless network system is established to receive RFID tag data with time stamps from multiple UHF RFID readers in the operating room in real time; then, the raw data is preliminarily processed, including data cleaning and formatting, including time stamp standardization and signal strength filtering;
[0007] To establish a secure wireless network:
[0008] S1.1. Isolate and protect medical data transmission by adopting WPA3 encryption protocol and creating independent virtual LAN VLAN; further include multiple authentication mechanisms, combining biometrics, RFID employee cards and passwords to ensure that only authorized users and devices can access RFID data;
[0009] S1.2, using the Network Time Protocol (NTP) server to synchronize the data timestamps in all RFID readers;
[0010] S1.3. Develop intelligent data filtering algorithms to perform preliminary cleaning and filtering of data at the RFID reader level, using machine learning techniques to identify and remove noise and non-critical signals;
[0011] S1.4, dynamically adjust the signal strength of the RFID reader through an adaptive algorithm according to the distribution and movement of objects in the operating room;
[0012] S2. Then develop an environmental state modeling algorithm based on machine learning, use the processed data to train the algorithm to identify and predict the usage patterns and movement trajectories of items in the operating room, and distinguish normal operations from abnormal behaviors based on this pattern;
[0013] S3. Finally, a real-time abnormal behavior detection system is introduced to identify events that significantly deviate from normal behavior based on pre-defined historical pattern standards, including unauthorized movement of items or loss of equipment. When abnormal behavior is detected, an alarm is issued to managers and event details are recorded for subsequent analysis.
[0014] Furthermore, the intelligent data filtering algorithm application method includes:
[0015] First, each RFID reader is integrated with a micro embedded processor and equipped with a developed firmware program, which enables the processor to perform preliminary data processing and analysis before sending the data to the central server; data processing includes applying the signal feature extraction formula:
[0016]
[0017] Where x represents the original data value extracted from the RFID signal, n represents the number of terms in the series, and the time interval and signal strength features are separated from the signal through series expansion;
[0018] Then use the machine learning model:
[0019]
[0020] Where x represents the processed data after feature extraction, t represents the integral variable, and the g(x) function model combines the characteristics of the probability density function and the error function to distinguish between noise and valid signals;
[0021] Finally, a dynamic adjustment algorithm is used:
[0022]
[0023] Where p represents the prediction accuracy of the current model, q represents the real-time error rate, and λ is a parameter adjusted through online learning.
[0024] Furthermore, the method for dynamically adjusting the signal strength of the RFID reader includes:
[0025] S1. First, deploy multiple UHF RFID readers in key areas of the operating room, including the surgical instrument area, drug storage area, and surgical area, to capture the location and movement of items with RFID tags in real time;
[0026] S2, then use the adaptive signal adjustment algorithm, through the formula:
[0027]
[0028] To adjust the signal strength of each reader, where p represents the real-time position of the tag, v represents the fixed position of the reader, and α, β, γ are adjustment parameters to minimize signal interference and optimize the reading range of the tag;
[0029] S3, then use the data processing unit to analyze the data collected from the reader in real time, through advanced algorithms:
[0030]
[0031] Filter and identify key events, including the absence or unauthorized removal of equipment, where x represents the data collected from the RFID tag, c k and λ k are model parameters;
[0032] S4. Finally, abnormal events are reported to the operating room management staff in real time, and relevant data are recorded for future analysis.
[0033] Furthermore, the environmental state modeling algorithm:
[0034] First, the frequency and power of the RFID reader are dynamically adjusted to control and optimize the reception quality and accuracy of the tag signal;
[0035] Secondly, machine learning models including random forests or deep learning networks are used to analyze and identify the usage patterns and movement trajectories of items in the operating room, based on optimized data collected from RFID readers and trained to monitor normal and abnormal behaviors of items;
[0036] Finally, the deviation between current behavior and historical normal patterns is monitored in real time, combined with real-time RFID data tracking, and potential risk events are discovered and reported in a timely manner by setting thresholds.
[0037] Furthermore, the construction method of dynamically adjusting the RFID reader includes:
[0038] S1. First, the frequency selection system AFS is used to detect and identify the electromagnetic interference source in the operating room based on real-time environmental sensing technology, and the formula is:
[0039]
[0040] Select the operating frequency, where f represents the frequency; α, β, λ are adjustment parameters used to control the accuracy and response speed of frequency selection; Used to describe the frequency components of the signal and its relationship with the interference source. The exponential part attenuates the high-frequency components, and the cosine part adjusts the frequency. It is used to avoid the influence of high-frequency interference and control the stability of frequency selection; The integral sign indicates that all frequencies are considered comprehensively;
[0041] S2, then use adaptive power adjustment APA to dynamically adjust the power of the reader according to the actual distance between the tag and the reader, using the formula:
[0042]
[0043] Adjust the power output, where d represents the distance between the RFID tag and the reader; γ, δ are power adjustment coefficients used to control the dynamic adjustment of power output; The integral term represents the variation of power with distance and is calculated by integrating the distance, combining exponential decay and the square and cube root adjustment of the denominator, while (·) 2 The integration result is squared to control the nonlinear characteristics of the adjustment;
[0044] S3. Finally, the power output is adjusted according to the read success rate and signal quality feedback through the feedback control mechanism, using the formula:
[0045]
[0046] Among them, s represents a comprehensive indicator of signal quality, including signal-to-noise ratio or read success rate; ξ, η, θ are adjustment parameters used to control the response speed and amplitude of the feedback mechanism; The integral term represents the cumulative effect of changes in signal quality. By integrating the signal quality, short-term fluctuations are smoothed. -θ·积分项 is an exponential decay term used to dynamically adjust the weight of signal quality.
[0047] Furthermore, the method for analyzing and identifying the usage pattern and movement trajectory of items in the operating room:
[0048] S1. First, collect RFID data and sensor data in the operating room through multimodal data integration, use advanced data cleaning and optimization algorithms to process the data collected from the reader, and use the formula:
[0049]
[0050] Optimize, where t represents time, α, β, γ, δ are adjustment parameters;
[0051] S2, then use random forest and deep learning network models, including convolutional neural network CNN and long short-term memory network LSTM, to extract and select features of the optimized data, through the formula:
[0052]
[0053] Extract features, where x i represents the i-th feature, w i is the feature weight, γ, η, λ are adjustment parameters;
[0054] S3. Then use historical data to train the model, including data on normal and abnormal usage patterns, to ensure the accuracy and generalization ability of the model through cross-validation and hyperparameter optimization, using the formula:
[0055]
[0056] Where w represents the model parameters, μ and Σ are the mean and covariance matrices respectively, θ, κ are adjustment parameters, and Z is the normalization factor;
[0057] S4. Finally, the real-time collected RFID data is input into the trained machine learning model through the formula:
[0058]
[0059] Real-time analysis and prediction are performed, where s represents the signal strength, and ξ, δ, η, ζ, ν are adjustment parameters.
[0060] Furthermore, the deployment process of real-time monitoring of deviations between current behavior and historical normal patterns includes:
[0061] First, multiple UHF RFID readers are deployed to cover all key areas of the operating room to collect location and movement data of items in real time at high frequency, while implementing signal processing algorithms including dynamic signal enhancement and noise suppression technology;
[0062] Then use random forests or deep learning networks to analyze historical and real-time data by:
[0063]
[0064] Construct a behavioral pattern recognition model, where x i represents the feature data collected from RFID, w i ,v j are the learned weights used to identify normal usage patterns and abnormal behaviors of items;
[0065] Then implement real-time monitoring and pass real-time data through the model:
[0066]
[0067] Analysis, where y i Represent real-time data points to quickly detect deviations from historical patterns;
[0068] Finally, the threshold value will be:
[0069]
[0070] Dynamically set alarms, where z represents the detected deviation, τ is the set safety threshold, and ξ is the adjustment parameter. When this threshold is exceeded, the system immediately issues an alarm and records the event for subsequent processing.
[0071] Beneficial effects of the present invention:
[0072] 1. By deploying multiple UHF RFID readers in the operating room, the present invention can achieve real-time monitoring of all key surgical equipment and items. Each equipment and item is equipped with an RFID tag, and the system can collect data such as its location, movement trajectory, and usage status at a high frequency. This real-time monitoring ensures the accurate positioning and timely use of equipment during surgery, reduces delays and errors caused by loss or misuse of equipment during surgery, and significantly improves the smooth progress and safety of surgery.
[0073] 2. The present invention combines multiple sensor data and realizes refined control of environmental parameters in the operating room through advanced data processing and optimization algorithms. The system can monitor and automatically adjust temperature, humidity, and air quality parameters in real time to meet the specific environmental requirements of different types of surgeries. Compared with the traditional manual adjustment method, the control method of the present invention reduces the loopholes and burden of manual operation, improves the accuracy and response speed of environmental adjustment, and thus creates more ideal environmental conditions for surgery.
[0074] 3. By using wireless RFID technology, the present invention reduces the physical connection and cable clutter in the operating room, and reduces the complexity of equipment installation and maintenance. At the same time, the system uses wireless data transmission technology to avoid signal interference and attenuation problems caused by physical connections, ensuring the stability and reliability of data transmission. This flexibility and reliability enable the system to better adapt to the complex working environment and changing needs in the operating room, and improve the overall operating efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0075] Figure 1 This is a flow chart of the application method of the ultra-high frequency RFID technology of the present invention in operating room environment monitoring.
[0076] Figure 2 This is a flow chart of the application method of the intelligent data filtering algorithm of the present invention.
[0077] Figure 3 This is a flow chart of the method for dynamically adjusting the signal strength of an RFID reader according to the present invention.
[0078] Figure 4 Construct a flow chart for the environmental state modeling algorithm of the present invention. DETAILED DESCRIPTION
[0079] The specific implementation modes of the present invention will be described in detail below in conjunction with the accompanying drawings.
[0080] Combined with Figure 1 , the implementation of UHF RFID technology in the operating room environment involves a series of precise steps and advanced technology integration, aimed at achieving efficient monitoring and data management. First, the system collects time-stamped RFID tag data in real time by deploying multiple UHF RFID readers at key locations in the operating room. This data is transmitted through a secure wireless network system that uses WPA3 encryption protocol and independent virtual LAN VLAN to isolate and protect sensitive medical data transmission.
[0081] To ensure data security and access control, the system includes a multi-authentication mechanism that combines biometrics, RFID employee cards and personal passwords to ensure that only authorized users and devices can access the data. At the same time, the data timestamps of all RFID readers are synchronized through the Network Time Protocol NTP server to ensure the time consistency and accuracy of the data.
[0082] In terms of data collection, intelligent data filtering algorithms are developed to perform preliminary cleaning and filtering of data directly at the RFID reader level. This includes the use of machine learning technology to identify and eliminate environmental noise and non-critical signals to optimize the quality and availability of data. In addition, based on the real-time distribution and dynamic movement of objects in the operating room, the system dynamically adjusts the signal strength of the RFID reader through an adaptive algorithm to improve the accuracy of data reception and the stability of the signal.
[0083] Furthermore, the system develops an environmental state modeling algorithm based on machine learning, and uses the preliminarily processed RFID data to train the algorithm, so that it can accurately identify and predict the usage patterns and movement trajectories of items in the operating room. These patterns help the system distinguish between normal operating behaviors and potential abnormal behaviors. Abnormal behaviors include unauthorized movement of items or improper loss of equipment.
[0084] Finally, a real-time abnormal behavior detection system is introduced, which can instantly identify any event that significantly deviates from normal behavior based on pre-defined historical pattern standards. When the system detects such abnormal behavior, it will automatically alert the management staff and record all the details of the relevant events, including time, location and items involved, for subsequent analysis and tracking. This not only enhances the safety protection capabilities of the operating room, but also improves the overall management efficiency and response speed of the medical environment.
[0085] Embodiment 1:
[0086] Combined with Figure 2 , the embodiment is designed around integrating a micro embedded processor in each RFID reader, and the firmware program it is equipped with. The main task of these processors is to perform preliminary data processing and analysis before the data is sent to the central server. Set up a surgical procedure in the operating room, and all the tools and equipment used by the medical team are equipped with RFID tags. The data collected by the RFID reader is processed using the signal feature extraction formula:
[0087]
[0088] Parameter settings:
[0089] x: represents the raw data value extracted from the RFID signal, specifically the signal strength detected by the RFID reader;
[0090] n: represents the number of terms in the series. The time interval and signal strength characteristics are separated from the signal through series expansion.
[0091] Assume that at a specific point in time, the signal strength of a specific device captured by the RFID reader is x = 0.5. For the first four terms of the series (n = 1 to 4), the formula is calculated as follows:
[0092]
[0093] The calculation results provide processed signal feature values for subsequent data analysis and anomaly detection. In this way, each RFID reader can extract key signal features in real time, helping the monitoring system to more accurately track the use of surgical tools and possible abnormal movements.
[0094] The embodiment further uses a machine learning model to determine whether the signal is a valid signal or noise:
[0095]
[0096] Where x represents the processed data after feature extraction, and t represents the integral variable.
[0097] The signal strength collected from the RFID tag of the surgical instrument is set to x = 0.5 after preliminary processing. Substitute this value into the above machine learning model:
[0098] calculate Using Simpson's rule, the interval (0, 0.5) is divided into two subintervals of equal length for calculation.
[0099] Determine the integration interval and step size:
[0100] Interval: (0, 0.5);
[0101] Divide the interval into two parts, each with a length of
[0102] Compute function values at endpoints and midpoints:
[0103]
[0104] midpoint
[0105] Use Simpson's rule formula to approximate the integral:
[0106]
[0107] f(0.5)=e -0.25 ≈0.7788
[0108] Substitute into Simpson's law formula:
[0109]
[0110] Substituting this result into the g(x) function, we get
[0111] This value will be compared with a predetermined threshold to determine whether it is a valid signal. If the threshold is set to 0.3, 0.5248 is greater than 0.3, indicating that this is a valid signal, not noise. In this way, the system can confirm that the data collected from this RFID tag is accurate and can update the current status of the surgical instrument based on this data.
[0112] The embodiment introduces a dynamic adjustment algorithm to optimize the prediction accuracy and error rate of the RFID system in real time:
[0113]
[0114] p represents the prediction accuracy of the current model, q represents the real-time error rate, and λ is a dynamically adjusted coefficient that is continuously updated through the online learning mechanism.
[0115] Assume that at a certain point in time, the prediction accuracy p of the system is 0.9, and the real-time error rate q is 0.1, indicating that the performance of the system is relatively stable and the error is low. Set the value range of λ to 0.01 to 0.1, and select λ = 0.05 for calculation. The dynamic adjustment algorithm can be used to evaluate the trend of the error rate:
[0116] h(0.9,0.1)=-0.05×0.9×(0.1) 2 =-0.05×0.9×0.01=-0.00045
[0117] The results show that under the current system parameters and environmental conditions, the error rate changes at a rate of -0.00045 per unit change in prediction accuracy. This small adjustment means that the system is currently very close to optimal performance and only needs fine-tuning to maintain or improve performance. By continuously monitoring the values of p and q and applying this dynamic adjustment algorithm, the system can react quickly when performance deviations are detected and adjust parameters to return to the optimal operating state.
[0118] Embodiment 2:
[0119] Combined with Figure 3 , the embodiment adopts an adaptive signal adjustment algorithm, through the formula:
[0120]
[0121] To dynamically adjust the signal strength of each RFID reader. p represents the real-time position of the RFID tag, v represents the fixed position of the RFID reader, and α, β, γ are adjustment parameters.
[0122] The tag position p of a scalpel is measured to be 0.5 meters away from the reader v. The parameters α = 100, a higher baseline signal strength, are set to ensure sufficient signal coverage; β = 50 and γ = 10 are used to adjust the sensitivity and speed of signal attenuation.
[0123] Substitute into the formula to calculate:
[0124]
[0125] Setting -2.5 is approximately 0.082, then:
[0126]
[0127] The calculation results show that when the scalpel is located 0.5 meters away from the reader, the signal strength of the RFID reader is automatically adjusted to 19.6 to ensure effective signal transmission, thereby reducing signal interference and optimizing the reading range.
[0128] The embodiment further integrates advanced data processing units and algorithms to analyze the data collected from RFID readers in real time and identify key events such as missing or unauthorized movement of equipment. The data processing algorithms used are:
[0129]
[0130] Where x represents the data collected from the RFID tag, c k and λ k are model parameters.
[0131] Set n=3, indicating three series terms.
[0132] Parameter settings: c1=0.5, c2=1.5, c3=-0.5, λ1=0.1, λ2=0.05, λ3=0.01.
[0133] These parameters are designed to respond to different types of signal strength and change patterns. When a device is suddenly moved from its normal location, the signal strength x of the RFID tag will change.
[0134] Assume that within a monitoring cycle, the signal strength x recorded by the RFID tag of the device is 2.5 (indicating that the device has been moved to another location). Substitute this value into the algorithm for calculation:
[0135]
[0136] Since P(x) produces a specific value, the system can use a preset threshold to decide whether to trigger an alarm. If P(x) exceeds the set threshold (such as 0.15), the system will determine it as abnormal movement and immediately alert the operating room manager, while recording the relevant data of this event for future analysis.
[0137] Embodiment 3:
[0138] Combined with Figure 4 , the embodiment dynamically adjusts the frequency and power of the RFID reader to optimize the reception quality and accuracy of the tag signal. The frequency selection system AFS is used to select the optimal operating frequency based on real-time environmental sensing technology. The specific calculation formula is:
[0139]
[0140] f represents frequency, and α, β, and λ are adjustment parameters used to control the accuracy and response speed of frequency selection. The exponential part is used to describe the frequency components of the signal and its relationship with the interference source. The exponential part attenuates the high-frequency components, the cosine part adjusts the frequency, and It is used to avoid the influence of high-frequency interference and control the stability of frequency selection. Indicates that all frequencies are considered comprehensively.
[0141] In a certain operating room environment, through spectrum analysis, the system identified that the main electromagnetic interference source frequency was around 2.4 GHz. In order to optimize the frequency selection, the parameters α = 0.01, β = 0.1, and λ = 0.001 were set.
[0142] The specific numerical calculation is as follows:
[0143] calculate The value of f = 2.4:
[0144]
[0145] calculate Set f = 2.4:
[0146]
[0147] Combined with the integral to calculate the frequency selection function value:
[0148]
[0149] By using the numerical integration method to calculate, the operating frequency finally selected can avoid interference sources and make the signal more stable.
[0150] The embodiment further adopts an adaptive power adjustment APA method to dynamically adjust the power output of the reader according to the actual distance between the RFID tag and the reader to optimize the signal quality and coverage. The specific calculation formula is:
[0151]
[0152] Where d represents the distance between the RFID tag and the reader, and γ and δ are power adjustment coefficients.
[0153] The distance d between the RFID tag of the electrosurgical unit and the reader during surgery is set to 1.5 meters. The system needs to adjust the power output according to this distance to ensure signal strength and stability.
[0154] The setting parameters are set as: γ=50, δ=0.1.
[0155] Calculate the integral term:
[0156]
[0157] Since the integral formula is relatively complicated, the numerical method Simpson's rule is usually used for calculation, and the integral result is 0.75.
[0158] Substituting the integral result into the formula:
[0159] P(1.5)=50(0.75) 2 =50×0.5625=28.125
[0160] The calculation results show that when the distance between the RFID tag and the reader is 1.5 meters, the power output of the reader should be adjusted to 28.125 to ensure effective transmission of the signal and minimize interference.
[0161] The final embodiment uses a feedback control mechanism to adjust the power output according to the reading success rate and signal quality feedback to ensure that the RFID system maintains optimal performance in a dynamically changing surgical environment. The specific calculation formula is:
[0162]
[0163] s represents a comprehensive indicator of signal quality, including signal-to-noise ratio or read success rate; ξ, η, θ are adjustment parameters used to control the response speed and amplitude of the feedback mechanism.
[0164] In the operating room, the RFID system needs to continuously monitor the signal quality and dynamically adjust the power output of the reader according to the actual situation. If the system detects that the signal quality is reduced, affecting the tracking and management of the equipment, the system needs to make adjustments immediately.
[0165] The parameters were set as: ξ=100, η=10, θ=0.5.
[0166] Set the signal quality s to 2 and calculate the integral term:
[0167]
[0168] The integral term is a standard integral, and the result is:
[0169]
[0170] Substituting the integral result into the formula:
[0171]
[0172] When the system detects that the signal quality s is 2, the power output of the RFID reader is dynamically adjusted to 14.83 to ensure the stability and effectiveness of the signal.
[0173] Embodiment 4:
[0174] The embodiment uses machine learning models including random forests or deep learning networks to analyze and identify the usage patterns and movement trajectories of items in the operating room. First, RFID data and sensor data in the operating room are collected through multimodal data integration. These data include the location of items recorded by RFID readers, timestamps, and environmental information provided by other sensors including cameras and temperature sensors. To ensure data quality, advanced data cleaning and optimization algorithms are used to process the collected data, and the formula:
[0175]
[0176] Optimize, where t represents time, and α, β, γ, δ are adjustment parameters.
[0177] The data collection time was set to 10 seconds, and the adjustment parameters of the system were set to: α = 0.1, β = 0.05, γ = 1.5, δ = 0.2.
[0178] Specific calculation:
[0179]
[0180] Calculate the two integral terms separately:
[0181]
[0182] Set the numerical integration method Simpson's rule to calculate and get the approximate value:
[0183]
[0184] Combining the two results:
[0185] C(t)≈0.85+1.20=2.05
[0186] The embodiment further adopts random forest and deep learning network models. After the RFID system integrates and optimizes multimodal data, the collected high-quality data will be input into the random forest and deep learning network models for feature extraction and selection. Through the formula:
[0187]
[0188] Extract features, where x i represents the i-th feature, w i is the feature weight, γ, η, λ are adjustment parameters.
[0189] In the actual surgical scene, the feature data collected by the system include the position, movement speed and usage frequency of surgical instruments. The following parameters are set for feature extraction: γ = 0.05, η = 1.0, λ = 0.1.
[0190] Feature weight w i Set according to the importance of the feature, such as w1=0.5, w2=1.0, w3=1.5.
[0191] The eigenvalues in the data set are set to: x1=1.2, x2=0.8, x3=0.6.
[0192] Compute the integral term for each feature:
[0193]
[0194]
[0195] Use numerical integration method to calculate and get the approximate value:
[0196]
[0197] Calculate the feature extraction value E(x):
[0198] E(x1)=0.5×1.2×1.10=0.66
[0199] E(x2)=1.0×0.8×0.70=0.56
[0200] E(x3)=1.5×0.6×0.50=0.45
[0201] Comprehensive results:
[0202] E(x)=0.66+0.56+0.45=1.67
[0203] During a complex heart surgery, the RFID system monitors the location and status of all key surgical instruments in real time through multimodal data integration and optimization. The high-quality data collected by the system is extracted using the above formula to obtain a comprehensive eigenvalue E(x) = 1.67. These eigenvalues are input into the deep learning network (CNN and LSTM), and the model identifies the normal usage patterns and abnormal behaviors of surgical instruments.
[0204] Then, the embodiment uses historical data to train the model, including data of normal and abnormal usage patterns, and ensures the accuracy and generalization ability of the model through cross-validation and hyperparameter optimization. The calculation formula used is:
[0205]
[0206] w represents the model parameters, μ and Σ are the mean and covariance matrices respectively, θ and κ are adjustment parameters, and Z is the normalization factor.
[0207] Set the following parameters:
[0208] Mean μ = 0;
[0209] Covariance matrix
[0210] Adjustment parameter θ = 1;
[0211] Adjustment parameter κ = 0.1;
[0212] The normalization factor Z=1.
[0213] First calculate the Gaussian distribution term:
[0214]
[0215] Assume w = (w1w2), mean μ = (00), then:
[0216]
[0217] Calculate the integral term:
[0218]
[0219] The numerical integration method Monte Carlo method is set to be used for calculation, and the approximate value is 2.5.
[0220] Calculate the model value M(w):
[0221]
[0222] The model analysis shows that Mw = 2.5, indicating that this is a potential abnormal behavior. After multimodal data integration and optimization, the RFID system can monitor the location and status of all key surgical instruments in real time.
[0223] The last embodiment inputs the real-time collected RFID data into the trained machine learning model, through the formula:
[0224]
[0225] Real-time analysis and prediction are performed, where s represents the signal strength, and ξ, δ, η, ζ, ν are adjustment parameters.
[0226] The parameters were set as follows: ξ = 10, δ = 0.05, η = 0.1, ζ = 2, ν = 0.2.
[0227] Assume that the signal strength of a scalpel detected by the system in real time is s=3, and substitute it into the formula for calculation.
[0228] Calculate the integral term:
[0229]
[0230] The calculation is divided into two parts:
[0231]
[0232]
[0233] Set the Simpson method to calculate the approximate value:
[0234]
[0235] Comprehensive calculation:
[0236] D(s)=10×(2.1+1.5)=10×3.6=36
[0237] During the operation, the RFID system detected in real time that the signal strength of a scalpel was 3. The system uses the formula:
[0238]
[0239] The result is 36, which indicates the comprehensive analysis value of the current signal. Through the previously trained machine learning model, the system can determine whether the status of the scalpel is normal based on this result.
[0240] Embodiment 5:
[0241] The embodiment monitors the deviation of current behavior from historical normal patterns in real time, and combines it with real-time RFID data tracking to timely discover and alarm potential risk events by setting thresholds. First, multiple ultra-high frequency RFID readers are deployed in all key areas of the operating room. These readers collect the location and movement data of items in real time at a high frequency. To ensure the accuracy of the data, signal processing algorithms are implemented, including dynamic signal enhancement and noise suppression technology.
[0242] In an operating room, 10 UHF RFID readers are deployed to cover the surgical instrument area, drug storage area and surgical area. The reader collects data 10 times per second.
[0243] Position data: the position coordinates (x, y, z) of each tag;
[0244] Movement data: speed and direction of each tag.
[0245] Then use random forest or deep learning network to analyze historical and real-time data and build a behavior pattern recognition model. The model uses the formula:
[0246]
[0247] Calculate, where x i represents the feature data collected from RFID, w i and v j are the learned weights.
[0248] The parameters of the model are set as follows:
[0249] Feature weight w i and v j Based on the importance setting of the features, w1=0.6, w2=0.9, v1=1.2, v2=0.8;
[0250] Coefficient value range: x1=1.2 (position feature), x2=0.8 (moving speed).
[0251] Calculate the characteristic terms:
[0252]
[0253] Calculate the integral term:
[0254]
[0255] The approximate value is calculated by the numerical integration method Simpson's method:
[0256]
[0257] Comprehensive calculation:
[0258]
[0259] Finally, the behavior pattern recognition value P(x) is calculated:
[0260]
[0261] If the pattern recognition value is higher than the set threshold of 0.9, it is an abnormal behavior, an alarm is issued, and the event details including time, location, and involved devices are recorded.
[0262] The embodiment analyzes real-time data through a specific mathematical model to detect deviations from historical patterns, thereby ensuring the correct use of equipment and the safety of the environment. First, the UHF RFID system in the operating room continuously monitors the real-time data of equipment and items. Set the real-time data points collected by the system to y i Including information such as the device's location, movement speed, and signal strength. The data points collected are set as follows:
[0263] Position data point y1 = 3.2;
[0264] Equipment moving speed data point y2 = 1.8;
[0265] Signal strength data point y3 = 2.5.
[0266] Next, the real-time data is analyzed using the following models to detect deviations from historical patterns:
[0267]
[0268] Among them, y i represents real-time data points, and η is a tuning parameter used to control the sensitivity of the model to the signal.
[0269] Set the following parameters:
[0270] η=0.5,used for control The denominator of is used to adjust the influence of the signal in the model; the position data point y1 = 3.2.
[0271] Perform the following calculations:
[0272] Calculate the integral:
[0273]
[0274] Separate calculation of two parts and
[0275] Part 1:
[0276]
[0277] This is a Gaussian integral, but due to the existence of 3.2t, the standard form of Gaussian integral is needed to solve it. The standard Gaussian integral form is:
[0278]
[0279] Where α = 1 and β = 3.2. Therefore:
[0280]
[0281] Part 2:
[0282]
[0283] This integral is more complicated in practical applications, but an approximate value can be obtained through numerical integration. Set the numerical integration result to 0.5.
[0284] Summarizing the above points:
[0285]
[0286] Calculate the model output:
[0287] M(y)=log(18.282)≈2.81
[0288] Compare the calculated result M(y)≈2.81 with the data of the historical pattern. If the normal range of the historical pattern is between 2.5 and 3.0, the deviation of the current data point is within the normal range. If the result is outside this range, the system will identify it as an anomaly.
[0289] The last embodiment dynamically sets the alarm threshold according to the deviation detected in real time, and immediately issues an alarm when the set threshold is exceeded. This is achieved through the following formula:
[0290]
[0291] Where z represents the detected deviation, τ is the set safety threshold, and ξ is the adjustment parameter.
[0292] During a surgical operation, the system monitors the location and usage of all key surgical instruments in real time through an ultra-high frequency RFID reader. Assume that during the operation, the system detects that the position data of a scalpel has an abnormal deviation, and the deviation z is 3.2. At this time, the system needs to determine whether an alarm needs to be issued.
[0293] The safety threshold τ of the system is set to 2.0, and the adjustment parameter ξ is set to 0.5.
[0294] Substitute the deviation z=3.2 into the formula for calculation:
[0295]
[0296] Calculate the integral term:
[0297]
[0298] The calculation is performed using the numerical integration method Simpson's method, and the integral result is approximately 1.2.
[0299] Substituting the result into the formula:
[0300] θ(3.2)=max(0,3.2-2.0+1.2)
[0301] Θ(3.2)=max(0,2.4)=2.4
[0302] Because the calculated threshold value of 2.4 exceeds the set safety threshold value of 2.0, the system will issue an alarm and record the event details, including time, location, and equipment involved, for subsequent processing.
[0303] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.
Claims
1. Application of UHF RFID technology in operating room environment monitoring, Features: S1. First, a secure wireless network system is established to receive RFID tag data with time stamps from multiple UHF RFID readers in the operating room in real time; then, the raw data is preliminarily processed, including data cleaning and formatting, including time stamp standardization and signal strength filtering; To establish a secure wireless network: S1.
1. Isolate and protect medical data transmission by adopting WPA3 encryption protocol and creating independent virtual LAN VLAN; further include multiple authentication mechanisms, combining biometrics, RFID employee cards and passwords to ensure that only authorized users and devices can access RFID data; S1.2, using the Network Time Protocol (NTP) server to synchronize the data timestamps in all RFID readers; S1.
3. Develop intelligent data filtering algorithms to perform preliminary cleaning and filtering of data at the RFID reader level, using machine learning techniques to identify and remove noise and non-critical signals; S1.4, dynamically adjust the signal strength of the RFID reader through an adaptive algorithm according to the distribution and movement of objects in the operating room; S2. Then develop an environmental state modeling algorithm based on machine learning, use the processed data to train the algorithm to identify and predict the usage patterns and movement trajectories of items in the operating room, and distinguish normal operations from abnormal behaviors based on this pattern; S3. Finally, a real-time abnormal behavior detection system is introduced to identify events that significantly deviate from normal behavior based on pre-defined historical pattern standards, including unauthorized movement of items or loss of equipment. When abnormal behavior is detected, an alarm is issued to managers and event details are recorded for subsequent analysis.
2. The method for applying the ultra-high frequency RFID technology in operating room environment monitoring according to claim 1, Features The intelligent data filtering algorithm application method includes: First, each RFID reader is integrated with a micro embedded processor and equipped with a developed firmware program, which enables the processor to perform preliminary data processing and analysis before sending the data to the central server; data processing includes applying the signal feature extraction formula: Where x represents the original data value extracted from the RFID signal, n represents the number of terms in the series, and the time interval and signal strength features are separated from the signal through series expansion; Then use the machine learning model: Where x represents the processed data after feature extraction, t represents the integral variable, and the g(x) function model combines the characteristics of the probability density function and the error function to distinguish between noise and valid signals; Finally, a dynamic adjustment algorithm is used: Where p represents the prediction accuracy of the current model, q represents the real-time error rate, and λ is a parameter adjusted through online learning.
3. The method for applying the ultra-high frequency RFID technology in operating room environment monitoring according to claim 1, Features The method for dynamically adjusting the signal strength of an RFID reader comprises: S1. First, deploy multiple UHF RFID readers in key areas of the operating room, including the surgical instrument area, drug storage area, and surgical area, to capture the location and movement of items with RFID tags in real time; S2, then use the adaptive signal adjustment algorithm, through the formula: To adjust the signal strength of each reader, where p represents the real-time position of the tag, v represents the fixed position of the reader, and α, β, γ are adjustment parameters to minimize signal interference and optimize the reading range of the tag; S3, then use the data processing unit to analyze the data collected from the reader in real time, through advanced algorithms: Filter and identify key events, including the absence or unauthorized removal of equipment, where x represents the data collected from the RFID tag, c k and λ k are model parameters; S4. Finally, abnormal events are reported to the operating room management staff in real time, and relevant data are recorded for future analysis.
4. The method for applying the ultra-high frequency RFID technology in operating room environment monitoring according to claim 1, Features The environmental state modeling algorithm: First, the frequency and power of the RFID reader are dynamically adjusted to control and optimize the reception quality and accuracy of the tag signal; Secondly, machine learning models including random forests or deep learning networks are used to analyze and identify the usage patterns and movement trajectories of items in the operating room, based on optimized data collected from RFID readers and trained to monitor normal and abnormal behaviors of items; Finally, the deviation between current behavior and historical normal patterns is monitored in real time, combined with real-time RFID data tracking, and potential risk events are discovered and reported in a timely manner by setting thresholds.
5. The method for applying the ultra-high frequency RFID technology in operating room environment monitoring according to claim 4, Features The construction method of the dynamic adjustment RFID reader includes: S1. First, the frequency selection system AFS is used to detect and identify the electromagnetic interference source in the operating room based on real-time environmental sensing technology, and the formula is: Select the operating frequency, where f represents the frequency; α, β, λ are adjustment parameters used to control the accuracy and response speed of frequency selection; Used to describe the frequency components of the signal and its relationship with the interference source. The exponential part attenuates the high-frequency components, and the cosine part adjusts the frequency. It is used to avoid the influence of high-frequency interference and control the stability of frequency selection; The integral sign indicates that all frequencies are considered comprehensively; S2, then use adaptive power adjustment APA to dynamically adjust the power of the reader according to the actual distance between the tag and the reader, using the formula: Adjust the power output, where d represents the distance between the RFID tag and the reader; γ, δ are power adjustment coefficients used to control the dynamic adjustment of power output; The integral term represents the variation of power with distance and is calculated by integrating the distance, combining exponential decay and the square and cube root adjustment of the denominator, while (·) 2 The integration result is squared to control the nonlinear characteristics of the adjustment; S3. Finally, the power output is adjusted according to the read success rate and signal quality feedback through the feedback control mechanism, using the formula: Among them, s represents a comprehensive indicator of signal quality, including signal-to-noise ratio or read success rate; ξ, η, θ are adjustment parameters used to control the response speed and amplitude of the feedback mechanism; The integral term represents the cumulative effect of changes in signal quality. By integrating the signal quality, short-term fluctuations are smoothed. -θ·积分项 is an exponential decay term used to dynamically adjust the weight of signal quality.
6. The method for applying the ultra-high frequency RFID technology in operating room environment monitoring according to claim 4, Features The method for analyzing and identifying the usage pattern and movement trajectory of items in the operating room: S1. First, collect RFID data and sensor data in the operating room through multimodal data integration, use advanced data cleaning and optimization algorithms to process the data collected from the reader, and use the formula: Optimize, where t represents time, α, β, γ, δ are adjustment parameters; S2, then use random forest and deep learning network models, including convolutional neural network CNN and long short-term memory network LSTM, to extract and select features of the optimized data, through the formula: Extract features, where x i represents the i-th feature, w i is the feature weight, γ, η, λ are adjustment parameters; S3. Then use historical data to train the model, including data on normal and abnormal usage patterns, to ensure the accuracy and generalization ability of the model through cross-validation and hyperparameter optimization, using the formula: Where w represents the model parameters, μ and Σ are the mean and covariance matrices respectively, θ, κ are adjustment parameters, and Z is the normalization factor; S4. Finally, the real-time collected RFID data is input into the trained machine learning model through the formula: Real-time analysis and prediction are performed, where s represents the signal strength, and ξ, δ, η, ζ, ν are adjustment parameters.
7. The method for applying the ultra-high frequency RFID technology in operating room environment monitoring according to claim 4, Features The deployment process of real-time monitoring of deviations between current behavior and historical normal patterns includes: First, multiple UHF RFID readers are deployed to cover all key areas of the operating room to collect location and movement data of items in real time at high frequency, while implementing signal processing algorithms including dynamic signal enhancement and noise suppression technology; Then use random forests or deep learning networks to analyze historical and real-time data by: Construct a behavioral pattern recognition model, where x i represents the feature data collected from RFID, w i ,v j are the learned weights used to identify normal usage patterns and abnormal behaviors of items; Then implement real-time monitoring and pass real-time data through the model: Analysis, where y i Represent real-time data points to quickly detect deviations from historical patterns; Finally, the threshold value will be: Dynamically set alarms, where z represents the detected deviation, τ is the set safety threshold, and ξ is the adjustment parameter. When this threshold is exceeded, the system immediately issues an alarm and records the event for subsequent processing.
Citation Information
Patent Citations
Self-adaptive operating room environment monitoring system
CN116293996A
Cited By
Intelligent operating room management early warning system based on multidirectional monitoring
CN120388699A
