Intelligent pediatric ward environment adaptive control system

Abstract

The application relates to the technical field of medical environment control, and particularly discloses an intelligent children's ward environment adaptive control system, which comprises a multi-modal perception module, a coupling evaluation module, a fusion modeling module, a simulation deduction module and an adaptive collaborative regulation module. The scheme constructs a coupling evaluation model to establish a quantitative mapping relationship between physical environment parameters and children's emotional states, generates a coupling evaluation index of the environment and the emotions, and provides scientific basis and quantitative support for environment regulation decision-making; a unified multi-sensory environment fusion model covering light environment, sound environment, thermal environment and olfactory environment is constructed, collaborative modeling and joint optimization of multi-dimensional environment elements are realized, and the comprehensive effect of environment regulation is effectively improved; an adaptive learning controller based on a context slot machine is introduced, a regulation strategy is dynamically optimized according to the age characteristics, historical preferences and real-time feedback of sick children, and personalized environment adaptive control is realized.

Description

Technical Field

[0001] This invention relates to the field of medical environment control technology, specifically to an intelligent adaptive control system for children's ward environment. Background Technology

[0002] Children exhibit significant developmental differences across age groups, sensory processing abilities, autonomy, and emotional regulation styles. Their environmental needs are highly heterogeneous across different age groups, posing higher demands for refined management of ward environments. Existing pediatric ward environmental control systems are mostly passively responsive, lacking quantitative correlation models between environmental parameters and children's psychological states. This makes it difficult to establish precise quantitative mappings between changes in environmental parameters and children's emotional states and stress levels, resulting in a lack of scientific basis for environmental control decisions. Children's perception of the environment is a comprehensive experience integrating multiple senses, including vision, hearing, touch, and smell. However, existing systems are often limited to a single sensory dimension, lacking unified modeling and collaborative optimization of multiple dimensions such as light, sound, heat, and olfactory environments. Children of different ages show significant differences in their sensitivity and preferences to sensory stimuli, while existing systems often employ a "one-size-fits-all" uniform control strategy, failing to dynamically adjust environmental parameters based on children's age characteristics, emotional states, and individualized needs. Summary of the Invention

[0003] To address the above issues and overcome the shortcomings of existing technologies, this invention provides an intelligent adaptive control system for pediatric ward environments. Addressing the lack of a quantitative correlation model between environmental parameters and children's psychological states, this solution constructs a coupled evaluation model to establish a quantitative mapping relationship between physical environmental parameters and children's emotional states, generating a coupled evaluation index of environment and emotion, providing a scientific basis and quantitative support for environmental control decisions. Addressing the insufficient collaborative control capability of multi-sensory environmental elements, this solution constructs a unified multi-sensory environment fusion model covering light, sound, heat, and olfactory environments, achieving collaborative modeling and joint optimization of multi-dimensional environmental elements, effectively improving the overall effect of environmental control. Addressing the lack of adaptive mechanisms for children's age differences and individual preferences, this solution introduces a context-based adaptive learning controller, dynamically optimizing control strategies based on the child's age characteristics, historical preferences, and real-time feedback, achieving personalized adaptive environmental control.

[0004] This invention provides an intelligent adaptive control system for a pediatric ward environment, comprising a multimodal sensing module, a coupled evaluation module, a fusion modeling module, a simulation and deduction module, and an adaptive collaborative control module;

[0005] The multimodal perception module collects environmental parameter vectors, children's physiological parameter vectors, and behavioral feature vectors inside the pediatric ward, constructs emotion category labels and corresponding emotion intensity confidence scores, and sends all collected data to the other modules.

[0006] The coupling assessment module establishes a quantitative mapping relationship between environmental parameter vectors and children's emotional states, calculates the coupling assessment index, constructs an emotional state prediction model based on deep learning algorithms, predicts the real-time emotional category label of children and their corresponding emotional intensity confidence based on the current ward's environmental parameter vectors, and sends the emotional state prediction model to the simulation and deduction module.

[0007] The fusion modeling module constructs a multi-sensory environment fusion model based on environmental parameter vectors to describe the relationship between the ward environment and the child's emotional state, calculates the comprehensive environmental comfort index, and sends the multi-sensory environment fusion model to the simulation and deduction module.

[0008] The simulation and deduction module constructs a set of candidate control strategies based on environmental parameter vectors for simulation and deduction. It calculates the change in the comprehensive environmental comfort index of each candidate control strategy according to the multi-sensory environment fusion model, predicts the change in the child's emotional category label through the emotional state prediction model, calculates the child's emotional improvement degree, generates a simulation evaluation report and sends it to the adaptive collaborative control module.

[0009] The adaptive collaborative control module, based on the simulation evaluation report, uses optimization algorithms and adaptive learning controllers to solve for the optimal control strategy, which is then decomposed into a set of control instructions for joint control of the pediatric ward environment.

[0010] Furthermore, the coupling evaluation module includes a feature extraction unit, a temporal prediction unit, a coupling calculation unit, and an evaluation output unit;

[0011] The feature extraction unit performs standardization and feature fusion on environmental parameter vectors, physiological parameter vectors, behavioral feature vectors, and emotion category labels to construct coupled feature vectors.

[0012] The temporal prediction unit divides the coupled feature vector into a training set and a validation set, and constructs an emotion state prediction model based on a long short-term memory network to predict the emotion category label of children in future time periods. The formula used is as follows: ;

[0013] In the formula, Indicates time, Indicates the length of the future time window. express Mood category labels for real-time prediction Indicates the length of the historical time window. Indicates from Time's up The sequence of historical coupled feature vectors at each moment. This represents a model for predicting emotional states.

[0014] The coupling calculation unit calculates the coupling evaluation index of environment and emotion based on the current environmental parameter vector and the predicted emotion category label. The larger the value, the higher the matching degree between the current environment and the child's emotional state.

[0015] The evaluation output unit outputs a coupled evaluation index, a predicted emotion category label, and its corresponding emotion intensity confidence level.

[0016] The evaluation output unit outputs a coupled evaluation index, a predicted emotional state, and its confidence vector.

[0017] Furthermore, the fusion modeling module constructs a multi-sensory environment fusion model based on environmental parameter vectors, specifically by calculating the light environment comfort index, sound environment comfort index, thermal environment comfort index, and olfactory environment comfort index, and then weighting and fusing the above comfort indices to obtain a comprehensive environmental comfort index.

[0018] Furthermore, the simulation and deduction module includes a scenario definition unit, a state deduction unit, an effect evaluation unit, and a report generation unit;

[0019] The scenario definition unit automatically enumerates the adjustment range and adjustment step size of each parameter in the environmental parameter vector to generate a set of candidate control strategies, where each candidate control strategy is a structured set of control instructions.

[0020] The state simulation unit simulates each candidate regulation strategy based on the multi-sensory environment fusion model, sets the simulation time step and simulation time window, calculates the change of the comprehensive environmental comfort index of the current candidate regulation strategy in each simulation time step, and calls the emotion state prediction model to predict the change of the child's emotion category label.

[0021] The effect evaluation unit calculates the degree of improvement in children's emotions under each candidate regulation strategy based on the results of state deduction, using the following formula: ;

[0022] In the formula, Indicates candidate regulatory strategies, Indicates candidate regulatory strategies The improvement in children's mood This represents the total number of simulation time steps. An index representing the simulation time step. Indicates the first Candidate control strategies are adopted within each simulation time step. Coupling evaluation index at time, Indicates the first The coupling evaluation index when no control strategy is adopted within each simulation time step;

[0023] The report generation unit generates a complete simulation evaluation report for each candidate control strategy.

[0024] Furthermore, the adaptive collaborative control module includes an objective function construction unit, a constraint modeling unit, an optimization solution unit, a learning controller, and a policy execution unit;

[0025] The objective function construction unit constructs a multi-objective optimization function based on the simulation evaluation report. The optimization objective is to maximize the improvement of children's emotions and the comprehensive environmental comfort index, while minimizing the control energy consumption cost. The form of the multi-objective optimization function is as follows: ;

[0026] In the formula, Represents a multi-objective optimization function. This indicates the operation of finding the maximum value of a multi-objective optimization function. Indicates candidate regulatory strategies The corresponding comprehensive environmental comfort index is the average value over the entire simulation time window. Indicates candidate regulatory strategies The corresponding energy consumption control costs, , and These represent the weighting coefficients corresponding to the degree of mood improvement, the overall environmental comfort index, and the energy consumption control cost, respectively.

[0027] The constraint modeling unit constructs the physical and safety constraints of candidate control strategies within the pediatric ward;

[0028] The optimization solution unit uses a genetic algorithm to iteratively optimize all candidate regulation strategies to obtain a set of optimal environmental regulation strategies.

[0029] The learning controller employs a context-based gambling machine algorithm as its adaptive learning mechanism. It selects the optimal control strategy from the set of optimal environmental control strategies and calculates the reward signal after execution based on the degree of emotion improvement and the comprehensive environmental comfort index. The strategy selection is optimized through online learning, using the following formula: ;

[0030] In the formula, Indicates in The optimal control strategy selected at any given time. This represents the index of the optimal environmental control strategy set. express The system state vector at time t. Indicates the optimal control strategy The estimated parameter vector, Indicates the optimal control strategy Online updated design matrix The exploration coefficient represents the strategy selection factor;

[0031] The strategy execution unit decomposes the optimal control strategy into a set of executable control instructions.

[0032] The beneficial effects achieved by adopting the above solution are as follows:

[0033] (1) In view of the lack of a quantitative correlation model between environmental parameters and children’s psychological state, this scheme constructs a coupled assessment model to establish a quantitative mapping relationship between physical environmental parameters and children’s emotional state, and generates a coupled assessment index of environment and emotion, so as to provide scientific basis and quantitative support for environmental regulation decision-making.

[0034] (2) In response to the problem of insufficient collaborative regulation of multi-sensory environmental elements, this solution constructs a unified multi-sensory environmental fusion model covering light environment, sound environment, thermal environment and olfactory environment, realizes collaborative modeling and joint optimization of multi-dimensional environmental elements, and effectively improves the comprehensive effect of environmental regulation.

[0035] (3) In response to the lack of adaptive mechanisms for children’s age differences and individual preferences, this solution introduces an adaptive learning controller based on contextual gambling machines. The control strategy is dynamically optimized according to the child’s age characteristics, historical preferences and real-time feedback to achieve personalized environmental adaptive control. Attached Figure Description

[0036] Figure 1 This is a schematic diagram of an intelligent adaptive control system for a pediatric ward environment proposed in this invention.

[0037] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. Detailed Implementation

[0038] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0039] Example 1, see Figure 1The present invention provides an intelligent adaptive control system for a pediatric ward environment, comprising a multimodal perception module, a coupled evaluation module, a fusion modeling module, a simulation and deduction module, and an adaptive collaborative control module;

[0040] The multimodal perception module collects environmental parameter vectors, children's physiological parameter vectors, and behavioral feature vectors inside the pediatric ward, constructs emotion category labels and corresponding emotion intensity confidence scores, and sends all collected data to the other modules.

[0041] The coupling assessment module establishes a quantitative mapping relationship between environmental parameter vectors and children's emotional states, calculates the coupling assessment index, constructs an emotional state prediction model based on deep learning algorithms, predicts the real-time emotional category label of children and their corresponding emotional intensity confidence based on the current ward's environmental parameter vectors, and sends the emotional state prediction model to the simulation and deduction module.

[0042] The fusion modeling module constructs a multi-sensory environment fusion model based on environmental parameter vectors to describe the relationship between the ward environment and the child's emotional state, calculates the comprehensive environmental comfort index, and sends the multi-sensory environment fusion model to the simulation and deduction module.

[0043] The simulation and deduction module constructs a set of candidate control strategies based on environmental parameter vectors for simulation and deduction. It calculates the change in the comprehensive environmental comfort index of each candidate control strategy according to the multi-sensory environment fusion model, predicts the change in the child's emotional category label through the emotional state prediction model, calculates the child's emotional improvement degree, generates a simulation evaluation report and sends it to the adaptive collaborative control module.

[0044] The adaptive collaborative control module, based on the simulation evaluation report, uses optimization algorithms and adaptive learning controllers to solve for the optimal control strategy, which is then decomposed into a set of control instructions for joint control of the pediatric ward environment.

[0045] Example 2, see Figure 1 This embodiment is based on the above embodiment, and the multimodal perception module includes an environmental perception unit, a physiological perception unit, a behavioral perception unit, and an emotion perception unit;

[0046] The environmental sensing unit is equipped with a light sensor, an illuminance sensor, a noise sensor, a temperature sensor, a humidity sensor, and a volatile organic compound sensor. It collects environmental data in the ward at a sampling frequency of once per minute, reads the indoor color temperature from the lighting system, performs analog-to-digital conversion on the above data, and generates an environmental parameter vector in the following form: ;

[0047] In the formula, Represents an environment parameter vector. Indicates light intensity. Indicates indoor color temperature. This represents the uniformity of indoor illuminance. Indicates the equivalent continuous A-weighted sound level. Indicates indoor temperature. Indicates indoor relative humidity. Indicates the concentration of volatile organic compounds;

[0048] The physiological sensing unit collects children's physiological data, including heart rate, heart rate variability, respiratory rate, skin conductance, and body motion frequency, through a wearable device at a sampling frequency of once per second, generating a physiological parameter vector in the following form: ;

[0049] In the formula, Represents a vector of physiological parameters. Indicates heart rate, Indicates heart rate variability. Indicates respiratory rate, Indicates skin conductivity, Indicates the frequency of body movement;

[0050] The behavior perception unit collects children's behavioral characteristics, including facial expression features, body posture features, and activity trajectory features, through high-definition cameras and infrared sensors deployed in the ward at a sampling frequency of 30 frames per second. After image preprocessing and feature extraction, a behavior feature vector is generated. The unit also collects sound signals in the ward through a microphone, classifies and outputs sound event types, counts the occurrence of each type of sound event within the collection time window, and calculates the sound event density.

[0051] The emotion perception unit, based on behavioral feature vectors and physiological parameter vectors, uses a multimodal fusion algorithm to initially identify the child's real-time emotional state. The emotional state is categorized into six basic emotions: pleasure, calmness, anxiety, sadness, fear, and anger. The unit outputs emotion category labels and corresponding emotion intensity confidence scores, in the following format: ; ; ;

[0052] In the formula, Indicates emotion category labels, This represents the six basic emotional states. This represents the confidence vector of emotional intensity. Indicates an index of emotional states. This indicates the confidence level of the emotional intensity.

[0053] Example 3, see Figure 1This embodiment is based on the above embodiment, and the coupled evaluation module includes a feature extraction unit, a temporal prediction unit, a feature matching unit, and an evaluation output unit;

[0054] The feature extraction unit performs Z-score normalization and feature fusion on environmental parameter vectors, physiological parameter vectors, behavioral feature vectors, and emotion category labels to construct coupled feature vectors;

[0055] The time-series prediction unit divides the coupled feature vectors from the past 3 to 6 months into a training set and a validation set in an 8:2 ratio. An emotion state prediction model is constructed based on a long short-term memory network, using cross-entropy loss as the optimization objective. The Adam optimization algorithm is used to iteratively train the emotion state prediction model. Training terminates when the validation set loss no longer decreases after 10 consecutive rounds. After training, the predicted emotion category label for the child in the future time period is output, using the following formula: ;

[0056] In the formula, Indicates time, Indicates the length of the future time window. express Mood category labels for real-time prediction Indicates the length of the historical time window. Indicates from Time's up The sequence of historical coupled feature vectors at each moment. This represents a model for predicting emotional states.

[0057] The coupling calculation unit calculates the coupling evaluation index between environment and emotion based on the current environmental parameter vector and the predicted emotion category label. The larger the value, the higher the matching degree between the current environment and the child's emotional state. The formula used is as follows: ; ;

[0058] In the formula, An assessment index representing the coupling between environment and emotion. , , , , and These represent the weighting coefficients for temperature, humidity, light intensity, noise, PM2.5 concentration, and volatile organic compound concentration, respectively. , , , , and These represent the suitability functions corresponding to the child's current emotional state for each environmental data point. In this embodiment, a temperature suitability function is constructed, in the following form: ;

[0059] In the formula, This indicates the most suitable temperature for the child in their current emotional state. This parameter represents the sensitivity of the control fitness function to temperature deviations. and The parameters are read from a pre-set parameter library based on the child's age and current emotional state.

[0060] The evaluation output unit outputs a coupled evaluation index, a predicted emotion category label, and its corresponding emotion intensity confidence level.

[0061] By performing the aforementioned operations, this solution addresses the lack of a quantitative correlation model between environmental parameters and children's psychological states. It constructs a coupled assessment model to establish a quantitative mapping relationship between physical environmental parameters and children's emotional states, generating a coupled assessment index of environment and emotion, thus providing a scientific basis and quantitative support for environmental regulation decisions.

[0062] Example 4, see Figure 1 This embodiment is based on the above embodiment. The fusion modeling module constructs a multi-sensory environment fusion model based on environmental parameter vectors, including a light environment modeling unit, an acoustic environment modeling unit, a thermal environment modeling unit, an olfactory environment modeling unit, and a model fusion unit.

[0063] The light environment modeling unit calculates the light environment comfort index based on light intensity, color temperature, and illuminance uniformity. The impact of the light environment on children's emotions is mainly achieved through two dimensions: light intensity and color temperature. High color temperature (above 5000K) cool light has a refreshing effect, while low color temperature (below 3000K) warm light has a relaxing and soothing effect. The formula used is as follows: ;

[0064] In the formula, This indicates the comfort index of the lighting environment. , and These represent comfort sub-functions for light intensity, color temperature, and illuminance uniformity, respectively. , and This represents the corresponding weighting coefficient;

[0065] The acoustic environment modeling unit calculates the acoustic environment comfort index based on the equivalent continuous A-weighted sound level, sound event type, and sound event density, using the following formula: ;

[0066] In the formula, This indicates the acoustic environment comfort index. This indicates the most suitable equivalent sound level for the child in their current emotional state. This represents the sound level sensitivity parameter. Indicates the type of sound event. Indicates the density of sound events. This represents the adjustment coefficients for the effect of sound event type and sound event density on the comfort of the acoustic environment, which are retrieved from a preset parameter table based on the sound event type.

[0067] The thermal environment modeling unit calculates the thermal comfort index based on indoor temperature and relative humidity using a simplified form of the predicted average voting index. The formula used is as follows: ;

[0068] In the formula, This represents the thermal comfort index. This indicates the most suitable temperature for the child in their current emotional state. This represents the temperature sensitivity parameter. This indicates the most suitable relative humidity for a child's current emotional state. This indicates a humidity sensitivity parameter;

[0069] The olfactory environment modeling unit calculates the olfactory environment comfort index based on the type and concentration of volatile organic compounds, using the following formula: ;

[0070] In the formula, This indicates the olfactory environment comfort index. This indicates the comfort threshold concentration of volatile organic compounds. Indicates the type of volatile organic compounds. This represents the adjustment coefficient of volatile organic compound type on olfactory comfort, which is obtained from a preset parameter table based on the volatile organic compound type.

[0071] The model fusion unit weights and fuses the light environment comfort index, sound environment comfort index, thermal environment comfort index, and olfactory environment comfort index. The fusion weights are dynamically adjusted according to the child's age and current emotional state: for infants and toddlers, the weights of the thermal and sound environments are relatively high; for school-aged children, the weights of the light and olfactory environments are appropriately increased. The formula used to calculate the comprehensive environmental comfort index is as follows: ;

[0072] In the formula, This represents the overall environmental comfort index. , , and These represent the fusion weights corresponding to the light environment, sound environment, thermal environment, and olfactory environment, respectively.

[0073] By performing the aforementioned operations, this solution addresses the problem of insufficient collaborative regulation capabilities among multi-sensory environmental elements. It constructs a unified multi-sensory environment fusion model encompassing light, sound, heat, and olfactory environments, enabling collaborative modeling and joint optimization of multi-dimensional environmental elements and effectively improving the overall effect of environmental regulation.

[0074] Example 5, see Figure 1 This embodiment is based on the above embodiment, and the simulation deduction module includes a scenario definition unit, a state deduction unit, an effect evaluation unit, and a report generation unit;

[0075] The scenario definition unit automatically enumerates the adjustment range and adjustment step size of each parameter in the environmental parameter vector to generate a set of candidate control strategies, where each candidate control strategy is a structured set of control instructions.

[0076] The state simulation unit simulates each candidate regulation strategy based on the multi-sensory environment fusion model, sets the simulation time step to 1 minute, the simulation time window to the next 1 to 24 hours, calculates the change of the comprehensive environmental comfort index of the current candidate regulation strategy in each simulation time step, and calls the emotion state prediction model to predict the change of the child's emotion category label.

[0077] The effect evaluation unit calculates the degree of improvement in children's emotions under each candidate regulation strategy based on the results of state deduction, using the following formula: ;

[0078] In the formula, Indicates candidate regulatory strategies, Indicates candidate regulatory strategies The improvement in children's mood This represents the total number of simulation time steps. An index representing the simulation time step. Indicates the first Candidate control strategies are adopted within each simulation time step. Coupling evaluation index at time, Indicates the first The coupling evaluation index when no control strategy is adopted within each simulation time step;

[0079] The report generation unit generates a complete simulation evaluation report for each candidate regulation strategy. The report includes environmental parameter vectors for each simulation time step, changes in emotion category labels, comprehensive environmental comfort index, and emotion improvement.

[0080] Example 6, see Figure 1 This embodiment is based on the above embodiment, and the adaptive collaborative control module includes an objective function construction unit, a constraint modeling unit, an optimization solution unit, a learning controller, and a policy execution unit;

[0081] The objective function construction unit constructs a multi-objective optimization function based on the simulation evaluation report. The optimization objective is to maximize the improvement in children's emotions and the overall environmental comfort index, while minimizing the energy consumption costs, including lighting energy consumption, temperature control energy consumption, and audio equipment energy consumption. The form of the multi-objective optimization function is as follows: ;

[0082] In the formula, Represents a multi-objective optimization function. This indicates the operation of finding the maximum value of a multi-objective optimization function. Indicates candidate regulatory strategies The corresponding comprehensive environmental comfort index is the average value over the entire simulation time window. Indicates candidate regulatory strategies The corresponding energy consumption control costs, , and These represent the weighting coefficients corresponding to the degree of mood improvement, the overall environmental comfort index, and the energy consumption control cost, respectively.

[0083] The constraint modeling unit constructs the physical and safety constraints of candidate control strategies in the pediatric ward. The physical constraints include the temperature adjustment range, the light intensity adjustment range, and the upper limit of the equivalent sound level. The safety constraints include environmental parameter change rate constraints and equipment operation safety constraints.

[0084] The optimization unit uses a genetic algorithm to iteratively optimize all candidate regulatory strategies. Real-number encoding is used, with a population size of 80 individuals, each representing a complete candidate regulatory strategy. Tournament selection with a size of 3 is used for selection, simulated binary crossover with a crossover probability of 0.85, and polynomial mutation with a mutation probability of 0.1. The optimal solution is considered to have an improvement of less than 10% over 50 consecutive generations. -6 The iteration stops when the condition is met, and a set of optimal environmental control strategies is obtained upon convergence.

[0085] The learning controller employs a context-based gambling machine algorithm as its adaptive learning mechanism. It selects the optimal control strategy from the set of optimal environmental control strategies and calculates the reward signal after execution based on the degree of emotion improvement and the comprehensive environmental comfort index. The strategy selection is optimized through online learning, using the following formula: ; ;

[0086] In the formula, express Momentary reward signals express The change in the coupling evaluation index relative to the previous time step. express The change in the overall environmental comfort index at any given time relative to the previous time. express The change in regulation energy consumption cost at any given time relative to the previous time. , , These represent the reward weight coefficients corresponding to each change; Indicates in The optimal control strategy selected at any given time. This represents the index of the optimal environmental control strategy set. express The system state vector at time t. Indicates the optimal control strategy The estimated parameter vector, Indicates the optimal control strategy Online updated design matrix The exploration coefficient represents the strategy selection factor;

[0087] The strategy execution unit decomposes the optimal control strategy into an executable set of light environment control instructions, a set of sound environment control instructions, a set of thermal environment control instructions, and a set of olfactory environment control instructions, and performs joint control of the light environment, sound environment, thermal environment, and olfactory environment in the pediatric ward.

[0088] By performing the aforementioned operations, this solution addresses the lack of adaptive mechanisms for children's age differences and individual preferences. It introduces a context-based adaptive learning controller that dynamically optimizes control strategies based on the child's age characteristics, historical preferences, and real-time feedback, thereby achieving personalized environmental adaptive control.

[0089] Example 7, see Figure 1This embodiment is based on the above embodiment. The initial values ​​of each parameter in the preset parameter table are set comprehensively based on hospital building design specifications, existing research results in environmental physiology, and clinical expert experience. During system operation, the data recording unit continuously records the control effect data and performs statistical calibration and incremental updates to the parameter table according to the operating cycle. For long-term hospitalized children, personalized offsets generated by the learning controller are superimposed on the preset parameters. The parameter table is set with a version number, and the reason for the update and the information of the reviewers are recorded synchronously after each update for periodic review by the hospital's expert committee.

[0090] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0091] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

[0092] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. An intelligent adaptive control system for a pediatric ward environment, characterized in that: It includes a multimodal perception module, a coupled evaluation module, a fusion modeling module, a simulation and deduction module, and an adaptive collaborative control module; The multimodal perception module collects environmental parameter vectors, children's physiological parameter vectors, and behavioral feature vectors inside the pediatric ward, constructs emotion category labels and corresponding emotion intensity confidence scores, and sends all collected data to the other modules. The coupling assessment module establishes a quantitative mapping relationship between environmental parameter vectors and children's emotional states, calculates the coupling assessment index, constructs an emotional state prediction model based on deep learning algorithms, predicts the real-time emotional category label of children and their corresponding emotional intensity confidence based on the current ward's environmental parameter vectors, and sends the emotional state prediction model to the simulation and deduction module. The fusion modeling module constructs a multi-sensory environment fusion model based on environmental parameter vectors to describe the relationship between the ward environment and the child's emotional state, calculates the comprehensive environmental comfort index, and sends the multi-sensory environment fusion model to the simulation and deduction module. The simulation and deduction module constructs a set of candidate control strategies based on environmental parameter vectors for simulation and deduction. It calculates the change in the comprehensive environmental comfort index of each candidate control strategy according to the multi-sensory environment fusion model, predicts the change in the child's emotional category label through the emotional state prediction model, calculates the child's emotional improvement degree, generates a simulation evaluation report and sends it to the adaptive collaborative control module. The adaptive collaborative control module, based on the simulation evaluation report, uses optimization algorithms and adaptive learning controllers to solve for the optimal control strategy, which is then decomposed into a set of control instructions for joint control of the pediatric ward environment.

2. The intelligent adaptive control system for a pediatric ward environment according to claim 1, characterized in that: The coupling evaluation module includes a feature extraction unit, a temporal prediction unit, a coupling calculation unit, and an evaluation output unit; The feature extraction unit performs standardization and feature fusion on environmental parameter vectors, physiological parameter vectors, behavioral feature vectors, and emotion category labels to construct coupled feature vectors. The temporal prediction unit divides the coupled feature vector into a training set and a validation set, and constructs an emotion state prediction model based on a long short-term memory network to predict the emotion category label of children in the future time period. The coupling calculation unit calculates the coupling evaluation index of environment and emotion based on the current environmental parameter vector and the predicted emotion category label. The larger the value, the higher the matching degree between the current environment and the child's emotional state. The evaluation output unit outputs a coupled evaluation index, a predicted emotion category label, and its corresponding emotion intensity confidence level.

3. The intelligent adaptive control system for a pediatric ward environment according to claim 2, characterized in that: The time-series prediction unit outputs the predicted emotion category label for the child within a future time period, using the following formula: ； In the formula, Indicates time, Indicates the length of the future time window. express Mood category labels for real-time prediction Indicates the length of the historical time window. Indicates from Time's up The sequence of historical coupled feature vectors at each moment. This represents a model for predicting emotional states.

4. The intelligent adaptive control system for a pediatric ward environment according to claim 1, characterized in that: The simulation and deduction module includes a scenario definition unit, a state deduction unit, an effect evaluation unit, and a report generation unit; The scenario definition unit automatically enumerates the adjustment range and adjustment step size of each parameter in the environmental parameter vector to generate a set of candidate control strategies, where each candidate control strategy is a structured set of control instructions. The state simulation unit simulates each candidate regulation strategy based on the multi-sensory environment fusion model, sets the simulation time step and simulation time window, calculates the change of the comprehensive environmental comfort index of the current candidate regulation strategy in each simulation time step, and calls the emotion state prediction model to predict the change of the child's emotion category label. The effect evaluation unit calculates the degree of improvement in children's emotions under each candidate regulation strategy based on the results of state deduction. The report generation unit generates a complete simulation evaluation report for each candidate control strategy.

5. The intelligent adaptive control system for a pediatric ward environment according to claim 4, characterized in that: In the effect evaluation unit, the formula for calculating the degree of mood improvement is as follows: ； In the formula, Indicates candidate regulatory strategies, Indicates candidate regulatory strategies The improvement in children's mood This represents the total number of simulation time steps. An index representing the simulation time step. Indicates the first Candidate control strategies are adopted within each simulation time step. Coupling evaluation index at time, Indicates the first The coupling evaluation index when no control strategy is adopted within a simulation time step.

6. The intelligent adaptive control system for a pediatric ward environment according to claim 1, characterized in that: The adaptive collaborative control module includes an objective function construction unit, a constraint modeling unit, an optimization solution unit, a learning controller, and a policy execution unit; The objective function construction unit constructs a multi-objective optimization function based on the simulation evaluation report. The optimization objective is to maximize the improvement of children's emotions and the comprehensive environmental comfort index, while minimizing the control energy consumption cost. The constraint modeling unit constructs the physical and safety constraints of candidate control strategies within the pediatric ward; The optimization solution unit uses a genetic algorithm to iteratively optimize all candidate regulation strategies to obtain a set of optimal environmental regulation strategies. The learning controller uses a context-based gambling machine algorithm as an adaptive learning mechanism. It selects the optimal control strategy from the set of optimal environmental control strategies and executes it. It also calculates the reward signal after execution based on the degree of emotion improvement and the comprehensive environmental comfort index, and optimizes the strategy selection through online learning. The strategy execution unit decomposes the optimal control strategy into a set of executable control instructions.

7. The intelligent adaptive control system for a pediatric ward environment according to claim 6, characterized in that: The formula used for strategy selection in the learning controller is as follows: ； In the formula, Indicates in The optimal control strategy selected at any given time. This represents the index of the optimal environmental control strategy set. express The system state vector at time t. Indicates the optimal control strategy The estimated parameter vector, Indicates the optimal control strategy Online updated design matrix The exploration coefficient represents the strategy selection.

8. The intelligent adaptive control system for a pediatric ward environment according to claim 6, characterized in that: In the objective function construction unit, the multi-objective optimization function takes the following form: ； In the formula, Represents a multi-objective optimization function. This indicates the operation of finding the maximum value of a multi-objective optimization function. Indicates candidate regulatory strategies The corresponding comprehensive environmental comfort index is the average value over the entire simulation time window. Indicates candidate regulatory strategies The corresponding energy consumption control costs, , and These represent the weighting coefficients corresponding to the degree of mood improvement, the overall environmental comfort index, and the energy consumption cost of regulation, respectively.

Images