Stem cell culture process monitoring method based on Internet of Things

Abstract

The invention relates to the field of remote monitoring, in particular to a stem cell culture process monitoring method based on the Internet of Things. Comprising the steps that environmental parameters are collected, sensor data are obtained, the sensor data are verified through a median filtering algorithm, and whether the sensor data are abnormal or not is judged; based on the sensor data marked as normal, calculating by using a nonlinear dynamic coupling analysis algorithm to obtain an environment state index; based on the environment state index, using a non-linear state monitoring prediction algorithm to predict and obtain an environment state index at a future moment; based on the predicted environmental state index of the future moment, using an intelligent early warning trigger algorithm based on environmental state prediction to obtain an early warning state, and generating an early warning signal according to the early warning state; and based on the early warning signal, a dynamic feedback adjustment algorithm is used to automatically generate a control signal. The technical problem that normal operation is interfered due to the fact that a wrong early warning signal is easily triggered by misjudgment when the environment state changes critically is solved.

Description

Technical Field

[0001] The present invention relates to the field of remote monitoring, and in particular to a method for monitoring a stem cell culture process based on the Internet of Things. Background Art

[0002] Stem cell technology has important research value in the fields of regenerative medicine, biomedicine, and drug development. It can repair damaged tissues, treat degenerative diseases, and construct disease models. However, the culture and expansion process of stem cells is complex and has extremely strict environmental requirements. Factors such as temperature, humidity, gas concentration, nutrient supply, and metabolite excretion will directly affect the success or failure of the culture. Therefore, the monitoring method of the stem cell culture process has become a key link in stem cell research and clinical applications. The traditional monitoring method of the stem cell culture process relies on the culture personnel to regularly detect the composition of the culture medium, check whether the temperature control system is stable, and require the culture personnel to use a microscope to observe the cell morphology. This is not only time-consuming and labor-intensive, but also easy to introduce contamination risks. In addition, the monitoring data is discrete and cannot form a complete dynamic analysis process. It is difficult to respond and correct in real time when an abnormality occurs, which can easily lead to culture failure. In order to solve the above problems, in recent years, the Internet of Things technology has gradually been introduced into the monitoring of the stem cell culture process, aiming to achieve real-time monitoring, analysis and intelligent control of key indicators.

[0003] With the continuous development of Internet of Things technology, the Internet of Things-based stem cell culture process monitoring method will be further intelligent and automated. It will not only be limited to laboratories and R&D institutions, but can also be applied to the production of cell therapy preparations and personalized medical services, providing strong support for promoting the development of regenerative medicine and precision medicine.

[0004] However, the existing monitoring methods for stem cell culture processes have the following technical problems: in a complex biological culture environment, the failure of a single sensor will affect the accuracy of the entire monitoring system, which may cause erroneous status assessment and decision-making. In the process of multi-node data transmission, continuous monitoring may not be ensured due to network jitter, delay or data loss. Traditional monitoring systems find it difficult to capture nonlinear interactions between environmental parameters, resulting in insufficient accuracy in environmental status monitoring and an inability to fully reflect the real-time status of the environment. In remote monitoring systems, changes in environmental status may be sudden or periodic, and traditional monitoring systems are often unable to accurately predict future trends of change, thereby increasing the risk of delayed response. In addition, when the environmental status changes critically, it is easy to trigger an erroneous warning signal due to misjudgment, resulting in interference with normal operation. Summary of the invention

[0005] The present invention provides a stem cell culture process monitoring method based on the Internet of Things to solve the problems that in a complex biological culture environment, a single sensor failure will affect the accuracy of the entire monitoring system, which may cause erroneous state evaluation and decision-making. In the process of multi-node data transmission, continuous monitoring may not be ensured due to network jitter, delay or data loss. Traditional monitoring systems are difficult to capture nonlinear interactions between environmental parameters, resulting in insufficient accuracy in environmental state monitoring and inability to fully reflect the real-time state of the environment. In a remote monitoring system, changes in environmental states may be sudden or periodic, and traditional monitoring systems are often unable to accurately predict future change trends, thereby increasing the risk of response lag. In addition, when the environmental state changes critically, it is easy to trigger an erroneous warning signal due to misjudgment, resulting in technical problems that interfere with normal operation.

[0006] The present invention provides a stem cell culture process monitoring method based on the Internet of Things, which specifically includes the following technical solutions:

[0007] A method for monitoring a stem cell culture process based on the Internet of Things comprises the following steps:

[0008] S 1: Collect environmental parameters, obtain sensor data, verify the sensor data through the median filter algorithm, and determine whether there is any abnormality in the sensor data;

[0009] S2: Based on the sensor data marked as normal, the environmental status index is calculated using the nonlinear dynamic coupling analysis algorithm;

[0010] S3: Based on the environmental state indicators, a nonlinear state monitoring prediction algorithm is used to predict the environmental state indicators at future moments; based on the predicted environmental state indicators at future moments, an intelligent early warning trigger algorithm based on environmental state prediction is used to obtain the early warning state, and an early warning signal is generated according to the early warning state;

[0011] S4: Based on the early warning signal, a dynamic feedback regulation algorithm is used to automatically generate a control signal.

[0012] Preferably, the S1 specifically includes:

[0013] In the implementation process of the median filtering algorithm, different sensor data of the same environmental parameter at the same time are sorted, and the median of the environmental parameter is calculated. The sensor data of each environmental parameter is compared with the median at the same time, and the difference between the two is calculated. When the difference between the sensor data and the median is less than or equal to the preset allowable error threshold of the environmental parameter, the sensor data is marked as normal, otherwise, the sensor data is marked as abnormal. The formula is as follows:

[0014]

[0015] Among them, V i，j (t) represents the verification status of the ith sensor of the ith environmental parameter at time t; S i，j (t) represents the sensor data measured by the jth sensor of the i-th environmental parameter at time t; S′ i (t) represents the median of all sensor data of the i-th environmental parameter at time t; δ i Represents the allowable error threshold of the i-th environmental parameter.

[0016] Preferably, the S1 specifically includes:

[0017] When a sensor is marked as abnormal, the abnormal sensor data will be eliminated and the abnormal sensor will be reported through the IoT management platform.

[0018] Preferably, the S2 specifically includes:

[0019] In the process of implementing the nonlinear dynamic coupling analysis algorithm, the environmental state index is generated by performing nonlinear analysis on the changes in sensor data of environmental parameters, adjusting the time-related weights and the coupling relationship between environmental parameters; and designing interactive coupling factors to calculate the coupling strength between different environmental parameters.

[0020] Preferably, the S2 specifically includes:

[0021] The calculation formula of environmental status index is:

[0022]

[0023] Among them, P(t) represents the environmental status index at time t; Indicates the accumulation of all environmental parameters: N represents the number of environmental parameters; represents the accumulation of all sensors of the i-th environmental parameter; Mi represents the number of sensors of the i-th environmental parameter; α i，j represents the weight coefficient of the jth sensor of the i-th environmental parameter; S i，j (t) represents the sensor data measured by the jth sensor of the i-th environmental parameter at time t; represents the time-dependent nonlinear adjustment factor; β i represents the time adjustment coefficient of the i-th environmental parameter; tan -1 represents the inverse tangent function; cos(γ i ·S i，j (t)) represents the cosine function; γ i represents the nonlinear adjustment coefficient related to the i-th environmental parameter; ∑ k≠i It means to sum all different environmental parameters except the i-th environmental parameter; Indicates that for the kth environmental parameter, calculate M k The coupling strength between the environmental parameters of the sensors; M k represents the number of sensors of the kth environmental parameter; i，k (t) represents the interaction coupling factor between the i-th environmental parameter and the k-th environmental parameter at time t.

[0024] Preferably, the S3 specifically includes:

[0025] In the implementation process of the nonlinear state monitoring prediction algorithm, the historical state changes of the monitoring system are captured through the data of the time window, and the environmental state indicators at future moments are predicted based on the weights of different time periods. The calculation formula is:

[0026]

[0027] Wherein, P(T) represents the predicted environmental status index at the future time T; P(t) represents the environmental status index at the time t; represents the weighted sum of the changes in each time window; L represents the number of time windows; wn represents the weight of the nth time window; Δt n represents the length of the nth time window; P(tn·Δt n ) represents the historical environmental status indicator in the nth time window; Indicates the change rate of the current environmental status index and the historical environmental status index in the nth time window; represents the sine function; Represents the phase factor of the sine function.

[0028] Preferably, the S3 specifically includes:

[0029] The warning status formula is:

[0030]

[0031] Among them, Q(t) represents the warning state; P(t) min Represents the lower threshold of the environmental status indicator; P(t) max Indicates the upper threshold of the environmental status indicator;

[0032] When the predicted environmental status indicator at a future moment is within the safety range of the lower and upper thresholds of the environmental status indicator, no intervention is performed; when the predicted environmental status indicator at a future moment is not within the safety range of the environmental status indicator, an early warning signal is triggered.

[0033] Preferably, the S4 specifically includes:

[0034] In the process of implementing the dynamic feedback regulation algorithm, the stem cell culture environment is dynamically feedback controlled in combination with error integration. An adaptive neural network is introduced to output a control signal based on the input environmental state indicator at the future moment. The deviation between the current environmental state indicator and the target environmental state indicator is calculated through error integration to adjust the control signal.

[0035] Preferably, the S4 specifically includes:

[0036] In the process of implementing the dynamic feedback regulation algorithm, an adaptive neural network is used to transmit the environmental state indicators at future moments layer by layer, and each hidden layer uses a hyperbolic tangent activation function to perform nonlinear transformation on the input environmental state indicators at future moments to obtain a control signal. The calculation formula of the control signal is:

[0037]

[0038] Wherein, C(t) represents the control signal generated at time t; represents the weighted sum of the output results of all neuron layers l = 1 to H; ω l represents the weight of the lth hidden layer; tanh represents the hyperbolic tangent activation function; W l represents the weight matrix of the lth hidden layer; b l represents the bias value of the lth hidden layer; η represents the integral control coefficient; represents the error integral term; τ represents the time integral variable; P o Indicates the target environment status indicator.

[0039] The beneficial effects of the technical solution of the present invention are:

[0040] 1. Use the sensor network to collect environmental parameters such as temperature, humidity, CO2 concentration, pH value, etc. in real time to obtain sensor data, and verify the sensor data through the median filter algorithm to eliminate abnormal sensor data. The data is transmitted to the IoT platform through the wireless communication module and the IoT platform for analysis and monitoring, ensuring that even when a single or multiple sensors fail, the monitoring system can still maintain the integrity and accuracy of the data, thereby avoiding monitoring errors caused by individual sensor failures and improving the robustness and stability of the monitoring system.

[0041] 2. Based on the sensor data marked as normal, the coupling relationship and change trend between different environmental parameters are analyzed through a nonlinear dynamic coupling analysis algorithm, and a comprehensive environmental status index is generated. The dynamic relationship between environmental parameters such as temperature and humidity is captured, providing a comprehensive understanding of environmental status indicators, avoiding the limitations that may be brought about by a single environmental parameter analysis, and enabling the monitoring system to cope with complex and changeable environmental conditions.

[0042] 3. Use nonlinear state monitoring and prediction algorithms to analyze long-term and short-term fluctuations in historical environmental status indicators, predict environmental status indicators at future times, and compare environmental status indicators at future times with safety thresholds to automatically trigger early warning signals to avoid training failures caused by delayed environmental changes. This early prediction capability improves the foresight and resilience of the monitoring system.

[0043] 4. Based on the early warning signal, a dynamic feedback adjustment algorithm is used to generate a control signal, and the control signal instruction is sent to the execution device (such as a temperature control system or a gas supply device) through the wireless communication module to achieve adaptive adjustment of environmental parameters. It can dynamically adjust the operating status of the equipment according to the changes in the actual environmental status indicators to ensure that the culture environment is always maintained in an ideal state. This adaptive adjustment reduces human intervention, improves the automation level of the monitoring system, and ensures the repeatability of stem cell culture.

[0044] 5. The sensor data collection, environmental status analysis, future environmental status indicator prediction, warning signal triggering, and feedback control links are organically combined to form a closed-loop control system, which improves the accuracy and continuity of stem cell culture process monitoring and can still operate efficiently under complex conditions. In addition, through the remote monitoring and control function, users can understand the environmental status at any time through mobile phones or computer terminals, realizing remote and intelligent management.

[0045] 6. The IoT platform realizes data transmission between sensor nodes through wireless communication modules and sends monitoring instructions to on-site equipment. Abnormal conditions and warning status will also be reported to the user terminal in real time through the IoT platform, realizing remote monitoring and control. Users can obtain environmental status information at any time and perform remote intervention when necessary, which improves the flexibility and response speed of the monitoring system, and is particularly suitable for intelligent management in distributed environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 This is a flow chart of a method for monitoring a stem cell culture process based on the Internet of Things described in the present invention. DETAILED DESCRIPTION

[0047] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0049] The specific scheme of the method for monitoring the stem cell culture process based on the Internet of Things provided by the present invention is described in detail below with reference to the accompanying drawings.

[0050] See attached Figure 1 , which shows a flow chart of a stem cell culture process monitoring method based on the Internet of Things provided by an embodiment of the present invention, the method comprising the following steps:

[0051] S1. Collect environmental parameters to obtain sensor data, verify the sensor data through the median filter algorithm, and determine whether there is any abnormality in the sensor data;

[0052] Arrange a sensor network, use sensors to collect environmental parameters in real time, such as temperature, humidity, CO2 concentration, and pH value, obtain sensor data, and realize data transmission and monitoring command interaction between sensor nodes through wireless communication modules and the Internet of Things platform to ensure that sensors collect sensor data in real time and upload it to the cloud or local server;

[0053] In order to ensure the accuracy and robustness of sensor data in the monitoring process, even when a single or multiple sensors fail, they can still operate normally and output reliable sensor data. The sensor data is verified by the median filtering algorithm to determine whether there are abnormalities in the sensor data and eliminate abnormal sensor data to ensure the continuity and reliability of data collection. Specifically, different sensor data of the same time and the same environmental parameters are sorted, and the median of the environmental parameters is calculated. Then, the sensor data of each environmental parameter is compared with the median at the same time, and the difference between the two is calculated. If the difference between a certain sensor data and the median is less than or equal to the preset allowable error threshold of the environmental parameter, the sensor data is marked as normal, otherwise, the sensor data is marked as abnormal. The formula is as follows:

[0054]

[0055] Among them, V i，j (t) represents the verification status of the jth sensor of the i-th environmental parameter at time t; S i，j (t) represents the sensor data measured by the jth sensor of the i-th environmental parameter at time t; S′ i (t) represents the median of all sensor data of the i-th environmental parameter at time t, which is used to reduce the impact of extreme values ​​(such as abnormal data) on the overall data; δ irepresents the allowable error threshold of the i-th environmental parameter, which is used to determine whether the sensor data is within a reasonable range. It can be set according to the specific implementation scenario and is not limited here;

[0056] The IoT platform will monitor the status of sensors in real time. When a sensor is marked as abnormal, the abnormal sensor data will be removed and the abnormal sensor will be reported through the IoT management platform. At the same time, even if some sensors fail, redundant monitoring will be completed based on the remaining sensor data to ensure the integrity of the data chain and improve the reliability of monitoring. It can still provide accurate monitoring information in the face of uncertain factors, so that occasional errors will not affect the normal operation of monitoring.

[0057] S2, based on the sensor data marked as normal, the environmental status index is calculated using a nonlinear dynamic coupling analysis algorithm;

[0058] After completing the sensor data verification, in order to reflect the dynamic characteristics of the environmental state, based on the sensor data marked as normal, the environmental state index is calculated using a nonlinear dynamic coupling analysis algorithm. The nonlinear dynamic coupling analysis algorithm is scheduled by the IoT platform for analysis to capture the coupling relationship and dynamic characteristics between different environmental parameters and provide a comprehensive understanding of the environmental state.

[0059] The nonlinear dynamic coupling analysis algorithm generates an environmental status index by performing nonlinear analysis on the sensor data changes of each environmental parameter, adjusting the time-related weights and the coupling relationship between environmental parameters; specifically, the sensor data change rate of each environmental parameter is calculated to identify the change trend of the environmental parameter at different time points, and the time-related nonlinear adjustment factor is used to make adjustments to reflect the complex changes of different environmental parameters over time. At the same time, the influence of periodic fluctuations is analyzed by embedding the cosine function, so as to capture the common regular changes in the stem cell culture environment;

[0060] The interactive coupling factor is designed to calculate the coupling strength between different environmental parameters. By comparing the differences in sensor data measured by different sensors for any two different environmental parameters, the similarity between the two is determined. The higher the similarity, the greater the coupling strength of the two environmental parameters. At the same time, in order to avoid misjudgment caused by excessive differences, an attenuation processing mechanism is introduced. When the difference in sensor data between two environmental parameters increases, the influence of the interactive coupling factor will be rapidly weakened.

[0061] The calculation formula of environmental status index is:

[0062]

[0063] Among them, P(t) represents the environmental status index at time t; Represents the outer summation symbol, which accumulates all environmental parameters: N represents the number of environmental parameters; represents the inner summation symbol, which accumulates all sensors of the i-th environmental parameter; M i represents the number of sensors of the i-th environmental parameter; α i，j represents the weight coefficient of the jth sensor of the i-th environmental parameter, which is used to adjust the contribution of each sensor in the calculation; S i，j (t) represents the sensor data measured by the jth sensor of the i-th environmental parameter at time t; represents the time-dependent nonlinear adjustment factor; β i represents the time adjustment coefficient of the i-th environmental parameter, controls the time correlation of environmental parameters, and ensures that the weights of different environmental parameters are dynamically adjusted over time; tan -1 represents the inverse tangent function, which is used to control time-dependent nonlinear transformations; cos(γ i ·S i，j (t)) represents the cosine function, which is used to perform nonlinear transformation on sensor data and enhance the sensitivity to environmental parameter fluctuations; γ i represents the nonlinear adjustment coefficient related to the i-th environmental parameter, which is used to adjust the sensitivity of environmental parameter fluctuations; ∑ k≠i It means to sum all different environmental parameters except the i-th environmental parameter to capture the relationship between environmental parameters; Indicates that for the kth environmental parameter, calculate M k The coupling strength between the environmental parameters of the sensors; M k represents the number of sensors of the kth environmental parameter; i，k (t) represents the interaction coupling factor between the i-th environmental parameter and the k-th environmental parameter at time t, reflecting the dynamic mutual influence between environmental parameters. The specific calculation formula is:

[0064]

[0065] Among them, κ represents the coupling coefficient, which is used to measure the coupling strength between two environmental parameters; represents the parameter change rate of the jth sensor of the i-th environmental parameter; represents the parameter change rate of the mth sensor of the kth environmental parameter; represents an exponential decay function. When the difference between the sensor data of the two environmental parameters increases, the influence of the interaction coupling factor will weaken rapidly; represents the attenuation coefficient of the sensor data difference between the i-th environmental parameter and the k-th environmental parameter; |S i，j (t)-S k，m(t)| represents the absolute value of the difference between the sensor data of the jth sensor of the i-th environmental parameter and the sensor data of the mth sensor of the k-th environmental parameter at time t, which is used to capture the instantaneous difference between the two environmental parameters;

[0066] S3. Based on the environmental state indicators, a nonlinear state monitoring prediction algorithm is used to predict the environmental state indicators at future moments; based on the predicted environmental state indicators at future moments, an intelligent early warning trigger algorithm based on environmental state prediction is used to obtain an early warning state, and an early warning signal is generated according to the early warning state;

[0067] Based on the environmental status indicators, a nonlinear state monitoring prediction algorithm is used to predict the environmental status indicators at future moments;

[0068] The nonlinear state monitoring prediction algorithm aims to capture the historical state changes of the monitoring system through data from multiple time windows, and predict the environmental state indicators at future moments according to the weights of different time periods; in order to capture the long-term and short-term trends of the environmental state indicators, the historical environmental state indicators are divided into time windows, the number of time windows is determined according to demand, each time window has a different weight, and the weight is dynamically adjusted according to the importance of the time window. When calculating the environmental state indicators at future moments, not only the environmental state indicators in each time window are considered, but also the changing trend of the environmental state indicators over time is analyzed. In order to accurately capture the dynamic changes, a nonlinear mapping method is used to link the amplitude and speed of the changes in the historical environmental state indicators with the current environmental state indicators, simulating the complex interactive relationship of the environmental state indicators; further introduce a dynamic adjustment mechanism to simulate the nonlinear changes in the real environment, refer to the change pattern of the sinusoidal waveform, superimpose the short-term fluctuations on the long-term trend, so that the environmental state indicators at future moments conform to the dynamic changes of the real environment;

[0069] The calculation formula for the predicted environmental status index at the future moment is:

[0070]

[0071] Wherein, P(T) represents the predicted environmental status index at the future time T; P(t) represents the environmental status index at the time t; represents the weighted sum of the changes in each time window; L represents the number of time windows; wn represents the weight of the nth time window, which is used to give different influences to different time windows; Δt n represents the length of the nth time window, which can be set according to the specific implementation scenario and is not limited here; P(tn·Δt n ) represents the historical environmental status indicator in the nth time window; It represents the change rate between the current environmental status index and the historical environmental status index in the nth time window, which is used to predict future changes; Represents a sine function, which is used to simulate the periodic characteristics of changes in environmental state indicators, capture the nonlinear oscillation trend in environmental state indicators, and improve the accuracy of predicting environmental state indicators; Represents the phase factor of the sine function, ensuring that the function value changes within the range of [-1, 1], adjusting the period of the sine function, and adapting to the frequency of changes in environmental status indicators;

[0072] The above steps provide accurate and intelligent environmental status prediction capabilities for the monitoring process of stem cell culture, and can accurately predict environmental status indicators at future moments when environmental status indicators fluctuate or delay;

[0073] An intelligent early warning trigger algorithm based on environmental state prediction is used to compare the predicted environmental state index at a future moment with a preset safety range to obtain an early warning state; if the predicted environmental state index at a future moment is within the safety range of the lower threshold and the upper threshold of the environmental state index, no intervention is made; otherwise, an early warning signal is immediately triggered, and the lower threshold and the upper threshold of the environmental state index can be specifically set according to the specific implementation scenario and are not limited here;

[0074] The warning status formula is as follows:

[0075]

[0076] Among them, Q(t) represents the warning state; P(t) min Represents the lower threshold of the environmental status indicator; P(t) max Indicates the upper threshold of the environmental status indicator;

[0077] The IoT platform will automatically trigger the warning signal according to the warning status and push the warning signal to the user terminal, such as mobile phone APP or computer console;

[0078] S4, based on the early warning signal, automatically generate the control signal using the dynamic feedback regulation algorithm;

[0079] Based on the early warning signal, a dynamic feedback adjustment algorithm is used to automatically generate a control signal, and the control signal instruction is sent to the execution device, such as the temperature control system or gas supply equipment, through the wireless network to achieve dynamic adjustment of the environment;

[0080] The dynamic feedback regulation algorithm uses a combination with error integration to achieve dynamic feedback control of the stem cell culture environment, introduces an adaptive neural network, outputs a preliminary control signal based on the input future environmental state index, and then calculates the deviation between the current environmental state index and the target environmental state index through error integration, and performs secondary adjustment on the control signal;

[0081] Using an adaptive neural network, the environmental state indicators at future moments are transmitted layer by layer. Each hidden layer uses a hyperbolic tangent activation function to perform nonlinear transformation on the input environmental state indicators at future moments, limiting the output value of the adaptive neural network to the range of (-1, 1) to ensure nonlinear control and avoid interference from extreme values. Each neuron of the adaptive neural network multiplies the input environmental state indicators at future moments by the corresponding weight matrix, and then adds the bias value and processes it through a hyperbolic tangent activation function. Further, by calculating the difference between the current environmental state indicator and the target environmental state indicator, the size and direction of the deviation are judged, and a basis is provided for adjusting the next control signal. The difference is further integrated to eliminate extremely small deviations and ensure stability during long-term operation, thereby realizing intelligent monitoring and feedback control of the stem cell culture environment.

[0082] The calculation formula of the control signal is:

[0083]

[0084] Wherein, C(t) represents the control signal generated at time t, which is used to adjust the environmental parameters; represents the weighted sum of the output results of all neuron layers l = 1 to H; ω l represents the weight of the lth hidden layer, which is used to adjust the contribution of the output of the hidden layer in the lth layer of the adaptive neural network to the control signal; tanh represents the hyperbolic tangent activation function, which is used to limit the output value of the adaptive neural network to the range of (-1, 1) to ensure nonlinear control and avoid interference from extreme values; W l represents the weight matrix of the lth hidden layer, which is used to map the input environment state indicators at future moments to the output of the corresponding layer in the adaptive neural network; b l represents the bias value of the lth hidden layer, ensuring flexibility and nonlinear response capability; η represents the integral control coefficient, which is used to adjust the influence of the error integral term on the control signal and balance the relationship between the cumulative effect of historical errors and the current control; represents the error integral term, which is used to correct the long-term deviation from the target state; represents the time integral variable; P o represents the target environment state index; P(τ) represents the environment state index at time τ;

[0085] In the embodiments of the present invention, the Internet of Things and intelligent control technology are combined to achieve real-time, accurate and intelligent monitoring of the stem cell culture environment, which fully reflects the in-depth application of the Internet of Things in the biomedical field. The closed-loop control chain formed by sensor monitoring, environmental state prediction, and feedback control can effectively reduce human intervention and improve the degree of automation of the stem cell culture process. It not only optimizes the culture efficiency, but also ensures the repeatability of the experiment and the reliability of the data, providing strong support for the promotion of stem cell technology in the fields of regenerative medicine and drug research and development.

[0086] In summary, a stem cell culture process monitoring method based on the Internet of Things was completed.

[0087] The order of the embodiments of the invention is for description only and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0088] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0089] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. A method for monitoring stem cell culture process based on the Internet of Things, characterized in that: The following steps are involved: S1: Collect environmental parameters, obtain sensor data, verify the sensor data through the median filter algorithm, and determine whether there is any abnormality in the sensor data; S2: Based on the sensor data marked as normal, the environmental status index is calculated using the nonlinear dynamic coupling analysis algorithm; S3: Based on the environmental state indicators, a nonlinear state monitoring prediction algorithm is used to predict the environmental state indicators at future moments; based on the predicted environmental state indicators at future moments, an intelligent early warning trigger algorithm based on environmental state prediction is used to obtain the early warning state, and an early warning signal is generated according to the early warning state; S4: Based on the early warning signal, a dynamic feedback regulation algorithm is used to automatically generate a control signal.

2. The method for monitoring stem cell culture process based on the Internet of Things according to claim 1, characterized in that: The S1 specifically includes: In the implementation process of the median filtering algorithm, different sensor data of the same environmental parameter at the same time are sorted, and the median of the environmental parameter is calculated. The sensor data of each environmental parameter is compared with the median at the same time, and the difference between the two is calculated. When the difference between the sensor data and the median is less than or equal to the preset allowable error threshold of the environmental parameter, the sensor data is marked as normal, otherwise, the sensor data is marked as abnormal. The formula is as follows: Among them, V i，j (t) represents the verification status of the ith sensor of the ith environmental parameter at time t; S i，j (t) represents the sensor data measured by the jth sensor of the i-th environmental parameter at time t; S′ i (t) represents the median of all sensor data of the i-th environmental parameter at time t; δ i Represents the allowable error threshold of the i-th environmental parameter.

3. The method for monitoring stem cell culture process based on Internet of Things according to claim 2, characterized in that: The S1 specifically includes: When a sensor is marked as abnormal, the abnormal sensor data will be eliminated and the abnormal sensor will be reported through the IoT management platform.

4. The method for monitoring stem cell culture process based on Internet of Things according to claim 1, characterized in that: The S2 specifically includes: In the process of implementing the nonlinear dynamic coupling analysis algorithm, the environmental state index is generated by performing nonlinear analysis on the changes in sensor data of environmental parameters, adjusting the time-related weights and the coupling relationship between environmental parameters; and designing interactive coupling factors to calculate the coupling strength between different environmental parameters.

5. The method for monitoring stem cell culture process based on Internet of Things according to claim 4, characterized in that: The S2 specifically includes: The calculation formula of the environmental status index is: Among them, P(t) represents the environmental status index at time t; Indicates the accumulation of all environmental parameters: N represents the number of environmental parameters; Indicates the accumulation of all sensors of the i-th environmental parameter; M i Represents the number of sensors of the i-th environmental parameter; α i，j represents the weight coefficient of the jth sensor of the i-th environmental parameter; S i，j (t) represents the sensor data measured by the jth sensor of the i-th environmental parameter at time t; represents the time-dependent nonlinear adjustment factor; β i represents the time adjustment coefficient of the i-th environmental parameter; tan -1 represents the inverse tangent function; cos(γ i ·S i，j (t)) represents the cosine function; γ i represents the nonlinear adjustment coefficient related to the i-th environmental parameter; ∑ k≠i It means to sum all different environmental parameters except the i-th environmental parameter; Indicates that for the kth environmental parameter, calculate M k The coupling strength between the environmental parameters of the sensors; M k represents the number of sensors of the kth environmental parameter; i，k (t) represents the interaction coupling factor between the i-th environmental parameter and the k-th environmental parameter at time t.

6. The method for monitoring stem cell culture process based on Internet of Things according to claim 1, characterized in that: The S3 specifically includes: In the implementation process of the nonlinear state monitoring prediction algorithm, the historical state changes of the monitoring system are captured through the data of the time window, and the environmental state indicators at future moments are predicted based on the weights of different time periods. The calculation formula is: Wherein, P(T) represents the predicted environmental status index at the future time T; P(t) represents the environmental status index at the time t; represents the weighted summation of the changes in each time window; L represents the number of time windows; wn represents the weight of the nth time window; Δtn represents the length of the nth time window; P(tn·Δt n ) represents the historical environmental status indicator in the nth time window; Indicates the change rate of the current environmental status index and the historical environmental status index in the nth time window; represents the sine function; Represents the phase factor of the sine function.

7. The method for monitoring stem cell culture process based on Internet of Things according to claim 1, characterized in that: The S3 specifically includes: The warning status formula is: Among them, Q(t) represents the warning state; P(t) min Indicates the lower threshold of the environmental status indicator; P(t) max Indicates the upper threshold of the environmental status indicator; When the predicted environmental status indicator at a future moment is within the safety range of the lower and upper thresholds of the environmental status indicator, no intervention is performed; when the predicted environmental status indicator at a future moment is not within the safety range of the environmental status indicator, an early warning signal is triggered.

8. The method for monitoring stem cell culture process based on Internet of Things according to claim 1, characterized in that: The S4 specifically includes: In the implementation process of the dynamic feedback regulation algorithm, the stem cell culture environment is dynamically feedback controlled in combination with error integration. An adaptive neural network is introduced to output a control signal based on the input environmental state indicator at the future moment. The deviation between the current environmental state indicator and the target environmental state indicator is calculated through error integration to adjust the control signal.

9. The method for monitoring stem cell culture process based on Internet of Things according to claim 8, characterized in that: The S4 specifically includes: In the process of implementing the dynamic feedback regulation algorithm, an adaptive neural network is used to transmit the environmental state indicators at future moments layer by layer, and each hidden layer uses a hyperbolic tangent activation function to perform nonlinear transformation on the input environmental state indicators at future moments to obtain a control signal. The calculation formula of the control signal is: Wherein, C(t) represents the control signal generated at time t; represents the weighted sum of the output results of all neuron layers l = 1 to H; ω l represents the weight of the lth hidden layer; tanh represents the hyperbolic tangent activation function; W l represents the weight matrix of the lth hidden layer; b l represents the bias value of the lth hidden layer; η represents the integral control coefficient; represents the error integral term; τ represents the time integral variable; P o Indicates the target environment status indicator.

Images