Intelligent radiotherapy patient pain behavior monitoring system

Abstract

The invention relates to the technical field of radiotherapy pain monitoring, and discloses an intelligent radiotherapy patient pain behavior monitoring system. The system comprises a behavior monitoring module, a pain recognition module and a behavior regulation and control module. The behavior monitoring module comprises a plurality of behavior capturing units and a behavior consistency evaluation unit, the behavior capturing units are distributed at monitoring point positions, and the evaluation unit receives output signals and calculates behavior consistency indexes. The pain recognition module is connected with the evaluation unit, judges whether pain behavior capture operation is executed or not according to the consistency index, and collects pain behavior feature information during execution. The behavior regulation and control module is connected with the pain recognition module, processes the feature information to derive an optimal pain intervention scheme, and adjusts the monitoring parameter configuration of the radiotherapy environment according to the optimal pain intervention scheme. The system can comprehensively capture behaviors of a patient at different point locations, improves the monitoring reliability through consistency evaluation, specifically collects pain information, and optimizes the pain monitoring and intervention process in a radiotherapy environment.

Description

Technical Field

[0001] This invention relates to the field of radiotherapy pain monitoring technology, specifically to an intelligent radiotherapy patient pain behavior monitoring system. Background Technology

[0002] In the field of cancer treatment, radiotherapy is a commonly used method. However, patients often experience varying degrees of pain during radiotherapy, which not only affects their treatment experience but may also indirectly impact the treatment process. Currently, pain monitoring for radiotherapy patients largely relies on regular rounds by medical staff or proactive feedback from patients. Regular rounds are limited by time intervals, making it difficult to capture subtle changes in patients' pain behavior in real time, and easily missing the optimal time for pain intervention. Proactive patient feedback, on the other hand, is subject to subjective differences; some patients may be unwilling or unable to accurately express their pain status for various reasons, making it difficult for medical staff to obtain accurate and comprehensive pain information. Existing technologies also include pain monitoring methods based on single monitoring devices. These devices typically only capture some of the patient's behaviors at fixed locations. Since patients in a radiotherapy environment may move across multiple areas, the coverage of a single monitoring device is limited, making it difficult to comprehensively reflect the patient's pain-related behaviors at different monitoring points. Furthermore, data collected by different monitoring devices are often independent, lacking effective integration and comparison mechanisms, making it impossible to determine the consistency of monitoring results across different locations. This affects the reliability of the monitoring data and consequently, the accurate assessment of the patient's pain status.

[0003] When potential pain behaviors are detected in patients, existing systems often employ a uniform capture mode, failing to adapt to the specific conditions of different monitoring points. This results in the collection of some invalid data, increasing the burden of subsequent data processing. Regarding pain intervention, existing protocols are largely based on experience, lacking in-depth analysis of individual patient pain behavior characteristics. This makes it difficult to develop personalized intervention plans and to adjust the monitoring parameters of the radiotherapy environment in real time based on the intervention's effectiveness, leading to insufficient coordination between the entire monitoring and intervention process. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent pain behavior monitoring system for radiotherapy patients to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides an intelligent pain behavior monitoring system for radiotherapy patients, the system comprising: The behavior monitoring module includes multiple behavior capture units and a behavior consistency assessment unit. The multiple behavior capture units are respectively arranged at the multiple patient monitoring points. The behavior consistency assessment unit is used to receive the output signal of the behavior capture unit and calculate the behavior consistency index of the behavior capture unit accordingly. The pain recognition module is connected to the behavior consistency assessment unit and is used to determine whether a pain behavior capture operation needs to be performed based on the behavior consistency index of each behavior capture unit, and to collect the pain behavior feature information of the behavior capture unit when it is determined that a pain behavior capture operation should be performed. A behavior control module, connected to the pain recognition module, is used to process the pain behavior characteristic information to deduce the optimal pain intervention plan for the patient, and adjust the monitoring parameter configuration of the radiotherapy environment according to the optimal pain intervention plan.

[0006] Preferably, the behavior capture unit includes a behavior recording device installed around the patient monitoring point. The behavior recording device maintains a fixed relative position with the patient monitoring point. The behavior recording device integrates multiple behavior sensors and is also equipped with a positioning marker.

[0007] Preferably, the behavior consistency evaluation unit calculates the behavior consistency index of the behavior capture unit based on the output signal of the behavior capture unit. The specific process is as follows: for each behavior sensor, the original behavior data acquired by the behavior sensor in a continuous time series is aggregated by time window to generate aggregated behavior data and discrete behavior components. The aggregated behavior data and discrete behavior components are combined using a filtering algorithm to output optimized behavior data in a continuous time series. Based on the optimized behavior data in a continuous time series corresponding to each behavior sensor, the behavior consistency index of the behavior capture unit is calculated.

[0008] Preferably, the behavior consistency evaluation unit calculates the behavior consistency index of the behavior capture unit based on the optimized behavior data in the continuous time series corresponding to each behavior sensor. The specific process is as follows: for each behavior sensor, multiple data points at multiple sampling times are selected from the optimized behavior data in the continuous time series corresponding to the behavior sensor, and the behavior data values ​​of the behavior sensor at the multiple sampling times are obtained; for each sampling time, based on the behavior data values ​​of all the behavior sensors at the sampling time, a single-point behavior consistency metric corresponding to that sampling time is calculated; and the behavior consistency index of the behavior capture unit is determined by combining the single-point behavior consistency metrics corresponding to all the sampling times.

[0009] Preferably, the pain recognition module determines whether to perform a pain behavior capture operation based on the behavior consistency index of each behavior capture unit. The specific process is as follows: select any two groups of behavior capture units and calculate the difference in behavior consistency index between the two groups of behavior capture units; calculate the behavior consistency fluctuation coefficient based on the difference in behavior consistency index between all behavior capture units; when the behavior consistency fluctuation coefficient is less than a preset fluctuation threshold and the behavior consistency index of any behavior capture unit exceeds a preset consistency standard, it is determined that a pain behavior capture operation should be performed.

[0010] Preferably, the behavior recording device is further provided with multiple calibration points; the pain recognition module collects pain behavior feature information of the behavior capture unit, including: collecting behavior feature information of the behavior capture unit under multiple observation positions; the behavior control module derives the optimal pain intervention plan for the patient based on the pain behavior feature information, including: for each observation position, analyzing the geometric correlation features between the key behavior points and the multiple calibration points based on the behavior feature information collected under that observation position; integrating the geometric correlation features between the key behavior points and the multiple calibration points corresponding to all observation positions to construct a global geometric model between the key behavior points and the multiple calibration points; and deriving the optimal pain intervention plan for the patient based on the global geometric model between the key behavior points and the multiple calibration points.

[0011] Preferably, the behavior control module derives the optimal pain intervention plan for the patient based on a global geometric model between the behavioral key points and multiple calibration points. The specific process is as follows: generating multiple potential pain intervention plans based on the global geometric model; acquiring the patient's physiological state data at continuous time points; acquiring environmental condition data at continuous time points; and combining the patient's physiological state data and environmental condition data at continuous time points to optimize and screen the multiple potential pain intervention plans, thereby determining the optimal pain intervention plan.

[0012] Preferably, the system further includes an alarm activation module, which is connected to the behavior control module and is used to trigger an alarm mechanism in the radiotherapy environment by using the patient's physiological state data at continuous time points, the environmental condition data at continuous time points, and the optimal pain intervention plan.

[0013] Preferably, the behavior control module adjusts the monitoring parameter configuration of the radiotherapy environment according to the optimal pain intervention plan. The specific process is as follows: a preset monitoring parameter strategy is generated by referring to historical pain intervention records, wherein the preset monitoring parameter strategy includes preset monitoring parameter values ​​for different time periods; when the deviation between the optimal pain intervention plan and the preset monitoring parameter values ​​for the corresponding time period exceeds a preset deviation limit, the monitoring parameter configuration and the preset monitoring parameter strategy are updated according to the optimal pain intervention plan.

[0014] Preferably, the system further includes an interference suppression module connected to the behavior recording device. This module activates an electromagnetic shielding mechanism when a radiotherapy device start-up signal is detected. Specifically, the process involves: acquiring the original data fluctuation range of the behavior sensor within a preset time period before the radiotherapy device starts; monitoring the instantaneous data offset amplitude of the behavior sensor in real time after the radiotherapy device starts; if the instantaneous data offset amplitude exceeds a preset proportional threshold of the original data fluctuation range, injecting a reverse calibration signal into the behavior sensor; and dynamically adjusting the acquisition frequency of the behavior sensor based on the reverse calibration signal until the instantaneous data offset amplitude returns to the original data fluctuation range.

[0015] Compared with the prior art, the beneficial effects of the present invention are: By setting up multiple behavior capture units and deploying them at various patient monitoring points within the radiotherapy environment, comprehensive capture of patient behavior in different areas can be achieved, overcoming the limitations of a single monitoring point and ensuring that all types of patient behavior during radiotherapy are covered. The behavior consistency assessment unit processes the output signals from each behavior capture unit and calculates a behavior consistency index. This index reveals the correlation and matching degree between monitoring results from different points, helping to more comprehensively determine the authenticity and stability of patient behavior and avoiding misjudgments caused by biases in monitoring at a single point. The pain recognition module determines whether to perform pain behavior capture based on behavioral consistency indicators, making pain behavior collection more targeted, reducing unnecessary and invalid data collection, and allowing the system to focus its resources on pain behaviors that truly require attention. When performing pain behavior capture, this module focuses on collecting pain behavior characteristic information, which provides specific and detailed evidence for subsequent pain analysis, making the judgment of the patient's pain state more accurate and realistic. The behavior control module processes pain behavior characteristics to deduce the optimal pain intervention plan. It can develop an individualized plan based on the patient's specific pain manifestations, making the intervention measures more aligned with the patient's actual needs. Simultaneously, it adjusts the monitoring parameters of the radiotherapy environment according to the optimal pain intervention plan, achieving linkage between the monitoring system and the intervention measures. This allows the monitoring of the entire radiotherapy environment to dynamically adjust as the patient's pain state changes, making the monitoring process more adaptive and synchronized with the patient's real-time condition. The entire system forms a coherent process, from behavior capture and consistency assessment to pain identification, intervention plan derivation, and monitoring parameter adjustment. The modules work together to make the monitoring of patients' pain behavior in the radiotherapy environment more comprehensive and accurate, and the intervention measures more adaptive, so that patients can receive more personalized attention and treatment during radiotherapy. Attached Figure Description

[0016] Figure 1 This is a schematic diagram illustrating the working principle of the intelligent radiotherapy patient pain behavior monitoring system described in this invention. Figure 2 A flowchart for calculating the behavioral consistency index; Figure 3 A flowchart detailing the calculation of the behavioral consistency index; Figure 4 A flowchart generated for the optimal pain intervention plan; Figure 5 A flowchart for adjusting monitoring parameter configurations. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Please see Figure 1 This invention provides an intelligent pain behavior monitoring system for radiotherapy patients, the system comprising: This system is used in radiotherapy environments that include multiple patient monitoring points. It comprises a behavior monitoring module, a pain recognition module, and a behavior regulation module.

[0019] The behavior monitoring module includes multiple behavior capture units and behavior consistency assessment units, with the multiple behavior capture units being deployed at multiple patient monitoring points.

[0020] The behavior consistency assessment unit receives the output signal from the behavior capture unit and calculates the behavior consistency index of the behavior capture unit accordingly.

[0021] The pain recognition module is connected to the behavior consistency assessment unit. It determines whether a pain behavior capture operation needs to be performed based on the behavior consistency index of each behavior capture unit, and collects the pain behavior feature information of the behavior capture unit when it is determined to perform the pain behavior capture operation.

[0022] The behavior control module connects to the pain recognition module, processes pain behavior characteristics information to deduce the best pain intervention plan for the patient, and adjusts the monitoring parameter configuration of the radiotherapy environment according to the best pain intervention plan.

[0023] Example 1: See Figure 2 As a core component of the intelligent radiotherapy patient pain behavior monitoring system, the behavior capture unit must be designed to meet the requirements of stability and accuracy in the radiotherapy environment. This unit includes behavior recording devices installed around the patient monitoring points, maintaining a fixed relative position to ensure spatial consistency of the monitoring data. The behavior recording devices employ a modular design, facilitating flexible deployment at different locations within the radiotherapy room. The device housing is made of lightweight materials with electromagnetic shielding properties to adapt to the electromagnetic environment during radiotherapy equipment operation.

[0024] The behavior recording device integrates multiple behavior sensors employing multimodal sensing technologies, including but not limited to infrared thermal imaging sensors, millimeter-wave radar, and inertial measurement units. The infrared thermal imaging sensor captures changes in the patient's body surface temperature distribution, the millimeter-wave radar detects micro-movements in the patient's limbs, and the inertial measurement unit records changes in the patient's body posture. The sensors are mounted in a distributed layout on the surface of the behavior recording device, forming multi-angle coverage and avoiding blind spots. The sensor sampling frequency is dynamically configured according to the characteristics of the radiotherapy process, automatically increasing the sampling density during critical treatment phases.

[0025] Positioning markers, made of highly reflective material and permanently mounted on the outer casing of the behavior recording device, facilitate rapid localization by the visual recognition system. The geometries of these markers are specially designed and contain machine-readable coded information, enabling the system to accurately distinguish devices at different monitoring points. Within the radiotherapy room, multiple behavior recording devices use these markers to calibrate their spatial coordinates, establishing a unified monitoring reference system.

[0026] The behavioral consistency assessment unit receives raw data streams from each behavioral sensor and first performs time alignment processing to eliminate time deviations caused by transmission delays. For each behavioral sensor, the system employs a sliding time window mechanism to aggregate and analyze the raw data within a continuous time series. The time window length is dynamically adjusted according to the radiotherapy stage; a shorter window is used during treatment to improve response speed, while a longer window is used during intervals to enhance data stability. During the aggregation process, the system extracts statistical features within the window as aggregated behavioral data and simultaneously calculates discrete behavioral components to reflect data fluctuation characteristics.

[0027] The filtering algorithm employs a multi-stage processing architecture. Primary filtering suppresses noise from individual sensor data, while secondary filtering performs cross-validation by combining the correlation of data from multiple sensors. The optimized behavioral data retains valid signal characteristics while significantly reducing the impact of environmental interference. The system calculates the coordination index between the sensors based on the optimized behavioral data, serving as the basis for behavioral consistency assessment.

[0028] The calculation of the behavioral consistency index employs a hierarchical evaluation strategy. The bottom layer evaluates the stability of a single sensor over different time periods; the middle layer compares the data correlation between multiple sensors at the same location; and the top layer comprehensively evaluates the degree of coordination between different monitoring points. The system generates the final behavioral consistency index by weighted fusion of the evaluation results from each layer. This index reflects the stability of patient behavior patterns in real time, providing a quantitative basis for pain recognition.

[0029] At the hardware level, the behavior recording device has a built-in preprocessing unit with edge computing capabilities, enabling real-time data acquisition and preliminary processing. The device connects to the central processing system via wired or wireless means, with transmission efficiency meeting the real-time requirements of radiotherapy monitoring. The power management system supports continuous operation and is equipped with a backup power supply to handle sudden power outages.

[0030] The system software architecture adopts a microservice design, with the behavior consistency assessment function running as an independent service, facilitating functional expansion and performance optimization. The data processing algorithm employs a fault-tolerant mechanism, maintaining basic assessment capabilities even when some sensors fail. The user interface provides visualization tools to intuitively display the behavioral consistency change trends at each monitoring point.

[0031] The maintenance and calibration mechanism periodically performs automatic checks on the behavior recording device, including sensor accuracy verification, positioning marker status checks, and communication link testing. During calibration, the system automatically records the baseline parameters of each sensor for standardized processing of subsequent monitoring data. When a device performance deviation exceeds the allowable range, the system triggers a maintenance alarm and automatically adjusts relevant algorithm parameters to ensure monitoring continuity.

[0032] Example 2: See Figure 3 The behavior consistency assessment unit employs a multi-dimensional time series analysis method when calculating the behavior consistency index of the behavior capture unit. This process first establishes a data point sampling mechanism on the time axis for the optimized behavior data corresponding to each behavior sensor. The selection of sampling times follows an adaptive principle; the system dynamically adjusts the sampling density based on the data change rate, automatically increasing the number of sampling points during periods of rapid change in behavior characteristics and appropriately reducing the sampling frequency during relatively stable periods. The data points at each sampling time contain the synchronous measurement values ​​of each sensor at that time, forming a time-aligned multi-sensor dataset.

[0033] For calculating behavioral consistency at a single sampling moment, the system employs an evaluation method based on statistical distribution similarity. First, the data values ​​from each behavior sensor at that moment are normalized to eliminate the influence of differences in sensor dimensions. The normalized data constitutes a multi-dimensional feature vector, and the system calculates the dispersion of this vector in the feature space as a measure of single-point behavioral consistency. The dispersion assessment comprehensively considers the standard deviation and covariance matrix characteristics of the data across various dimensions, reflecting the collaborative change patterns among multi-sensor data. When the data from each sensor exhibits high consistency at a given moment, the dispersion of the feature vector is low; when significant differences occur, the dispersion increases accordingly.

[0034] The single-point behavioral consistency measure on the time series is smoothed using a moving average method to eliminate fluctuations caused by instantaneous disturbances. The width of the smoothing window matches the typical time scale of behavioral feature changes, preserving meaningful pattern changes while filtering out random noise. The system performs integral calculations on the smoothed single-point measure series and generates a comprehensive behavioral consistency index by combining it with time weighting coefficients. This index not only reflects the overall consistency level but also, through the setting of time weighting coefficients, pays more attention to the recent trend of behavioral feature changes.

[0035] The pain recognition module establishes a multi-level decision-making mechanism when determining whether to perform pain behavior capture. The first level compares the differences in behavioral consistency indicators between any two groups of behavior capture units. The difference calculation uses a relative rate of change assessment method to avoid misjudgments caused by absolute value differences. The system maintains a dynamically updated difference matrix, recording the real-time comparison results between all monitoring points. The second level calculates a global behavioral consistency fluctuation coefficient based on the difference matrix. This coefficient comprehensively reflects the degree of coordination of behavioral patterns throughout the entire monitoring area.

[0036] The calculation of the fluctuation coefficient incorporates a spatial correlation correction factor, taking into account the physical distribution of monitoring points within the radiotherapy room. Differences between adjacent points are given higher weight, as spatially close points should theoretically exhibit stronger behavioral consistency. The system continuously tracks the fluctuation coefficient's trajectory, establishing short-term and long-term reference baselines. When the real-time fluctuation coefficient falls below a preset threshold and the duration exceeds the minimum judgment interval, an anomaly detection process is triggered.

[0037] In the anomaly detection process, the system checks the independent consistency index of each behavior capture unit one by one. These indicators are compared with preset standard values ​​using fuzzy logic to avoid decision instability caused by rigid thresholds. The system constructs a multi-dimensional decision space, jointly evaluating parameters such as the fluctuation coefficient, independent consistency index, and their changing trends. When evidence from multiple dimensions jointly supports the existence of an abnormal behavior pattern, the system determines that pain behavior capture operation needs to be performed.

[0038] A confidence accumulation mechanism is employed during the decision-making process to avoid false triggers caused by occasional interference. Each suspected anomaly is recorded and assigned an initial confidence score. As the observation period continues, the system dynamically adjusts this score based on newly acquired evidence. Only when the accumulated confidence score exceeds the action threshold will the capture operation be finally confirmed and executed. This mechanism effectively balances the relationship between detection sensitivity and false alarm rate.

[0039] In terms of hardware implementation, the behavior consistency assessment function is distributed between edge computing nodes and the central processing unit. Time-sensitive single-point metric calculations are performed by edge nodes closer to the sensors, reducing data transmission latency. Global metric calculations and decision-making logic run in the central processing unit, leveraging its greater computing power for complex analysis. The system employs a dual-channel communication architecture, with critical data transmissions using redundant paths to ensure reliability.

[0040] The software algorithm adopts a modular design, decoupling the core evaluation logic from specific sensor types and adapting to different monitoring scenarios through parameter configuration. The system maintains a behavioral pattern knowledge base, recording historical evaluation results and decision-making cases, supporting case-based reasoning assistance. Evaluation parameters are adjustable online, allowing medical staff to fine-tune key parameters such as sensitivity based on actual observations.

[0041] The system continuously monitors itself during operation, including checking the rationality of the input and output of the evaluation logic, monitoring computing resource usage, and diagnosing communication status. When a potential problem is detected, it automatically switches to degraded mode and issues a maintenance reminder, ensuring basic functionality while avoiding erroneous decisions. A complete audit log is maintained for the data collection and evaluation process, supporting post-event analysis and process traceability.

[0042] Through multi-level time-series analysis and spatial correlation modeling, accurate assessment of patient behavioral consistency is achieved in complex radiotherapy environments. A multi-level decision-making mechanism combined with confidence management enables the system to reliably identify genuine pain behavior characteristics. The co-design of hardware and software ensures that the assessment process meets real-time requirements while maintaining sufficient analytical depth. The system possesses good adaptability and maintainability, and its assessment performance can be continuously optimized with accumulated user experience.

[0043] Example 3: See Figure 4 The behavior recording device employs an asymmetric spatial distribution of multiple calibration points. The coordinate relationships of these points in three-dimensional space form the basic framework for the system's geometric feature analysis. Each calibration point includes both active emission and passive reflection modules. The active emission module periodically emits a positioning signal in a specific frequency band, while the passive reflection module amplifies and reflects detection signals emitted by other devices in the environment. This dual-mode design ensures that the calibration points can be reliably identified under various observation conditions, maintaining their positioning function even when strong electromagnetic interference is generated by radiotherapy equipment. The spatial arrangement of the calibration points is optimized to ensure that clear signals from at least three calibration points can be captured simultaneously from any observation direction.

[0044] When collecting behavioral feature information from multiple observation locations, the pain recognition module employs a time-division multiplexing scanning strategy. The system acquires data sequentially from different angles according to a preset location switching sequence, with the dwell time at each location dynamically adjusted based on the information richness of that location. During acquisition at a single location, the system simultaneously records the spatial coordinates of key behavioral points and the response signals of calibration points, establishing a time-aligned observation dataset. Data acquisition at each location encompasses a complete workflow; the entire process from signal transmission, propagation, reflection to reception is precisely timed to calculate key parameters such as signal flight time.

[0045] The geometric correlation feature analysis employs a hierarchical processing architecture. At the first level, the system calculates the relative distances and angular relationships between key behavioral points and calibration points for data from a single observation azimuth. Distance calculations are based on signal propagation time measurements, while angular relationships are obtained through multi-receiver phase difference analysis. These raw geometric parameters are then unified into a local coordinate system based on the current behavior recording device after coordinate transformation. The second level of processing fuses these local geometric features with the absolute spatial location information of the calibration points to construct the constraint relationships between the key behavioral points in the global coordinate system.

[0046] The global geometric model is constructed using an iterative optimization method. During system initialization, an initial model framework is established based on the known spatial locations of calibration points. As observation data accumulates, the model parameters are progressively optimized by minimizing the reprojection error. The reprojection error is defined as the degree of difference between the geometric relationships predicted by the model and the actual observed values. During optimization, weights are assigned to different observation orientations, with orientations exhibiting better signal quality receiving higher confidence levels. The model update frequency is matched to the data acquisition rate to ensure timely reflection of behavioral changes while avoiding overfitting to noise.

[0047] When deriving the optimal pain intervention plan, the system combines a global geometric model with the patient's physiological state data. Physiological state data is acquired in real time through medical monitoring equipment, including but not limited to indicators such as heart rate, respiratory rate, and skin conductance. After time alignment processing, this data is jointly analyzed with the motion trajectories of key behavioral points to identify correlation patterns between the two. Environmental condition data mainly records parameters such as treatment room temperature, humidity, and equipment operating status, which may indirectly affect the patient's behavioral performance.

[0048] Potential pain intervention plans are generated using a rule-based reasoning approach. The system maintains an intervention knowledge base, where each intervention rule is associated with geometric feature patterns and physiological state conditions. When the match between real-time data and the preconditions of a rule exceeds an activation threshold, the corresponding intervention plan is added to the candidate list. Each candidate plan is accompanied by a multi-dimensional evaluation vector describing its expected performance in terms of pain relief effect, feasibility, and patient comfort.

[0049] The optimization screening process is based on a multi-objective decision-making framework. The system defines the following utility function to evaluate the overall suitability of each candidate solution: in, The overall utility value of the representative scheme, This represents the weight coefficient of the i-th evaluation dimension. It is a plan The standardized scoring function is used for the i-th dimension, where n is the total number of evaluation dimensions. Weighting coefficients are dynamically adjusted based on the treatment stage and individual patient differences; for example, during critical treatment periods, more emphasis is placed on pain relief, while during the recovery period, more attention is paid to long-term comfort. The scoring function employs a fuzzy membership method to avoid decision-making inconsistencies caused by rigid thresholds.

[0050] During the implementation of the intervention, the system continuously monitors feedback on the intervention's effectiveness. This feedback data includes both direct changes in patient behavior and responses to physiological indicators. Feedback analysis employs a change detection algorithm to identify significant differences before and after the intervention. When the actual results deviate significantly from the expected outcomes, the system initiates a program adjustment process, reassessing the suitability of each candidate program based on the latest data. This closed-loop adjustment mechanism enables the intervention program to adapt to dynamic changes in the patient's condition.

[0051] The hardware system employs a distributed architecture, with geometric computation tasks distributed to dedicated processors for parallel execution. Each processing node is responsible for data processing at a specific observation azimuth, and intermediate results are exchanged between nodes via a high-speed bus. A central coordinator integrates the outputs of all nodes to maintain the consistency of the global geometric model. The system is equipped with redundant computing resources, with backup instances for critical processing modules to ensure basic functionality is maintained even in the event of partial hardware failure.

[0052] In terms of software implementation, the geometric modeling algorithm adopts a plug-in design, with the core framework providing basic computational services and specific modeling methods selectable through configuration. The system supports dynamic loading of multiple coordinate transformation algorithms to adapt to application scenarios with varying accuracy requirements. The intervention knowledge base employs version-based management, allowing medical staff to review and adjust rule content, with the system automatically recording all modification history. The user interface provides 3D visualization tools to intuitively display the spatial relationship changes between key behavioral points and calibration points.

[0053] The system employs multiple data verification mechanisms during runtime. Raw observation data undergoes integrity checks before entering the processing flow, intermediate calculation results are timestamped and verified, and manual confirmation is implemented at key decision points. These measures effectively prevent erroneous data from influencing system judgments. The performance monitoring module continuously tracks the latency and resource usage of each processing stage, automatically triggering load balancing strategies when potential bottlenecks are detected.

[0054] In terms of maintenance and upgrades, the system supports remote diagnostics and parameter adjustment. The functional status of calibration points can be periodically self-checked, and abnormal situations are automatically reported. The parameters of the geometric modeling algorithm have online adjustment capabilities, optimizing model accuracy without system downtime. System updates utilize hot deployment technology, allowing for gradual switching to new software versions without affecting the current treatment process. A data backup mechanism ensures that all critical information has multiple copies to prevent accidental loss.

[0055] Example 4: See Figure 5The alarm activation module operates based on multi-source data fusion, triggering alarms by analyzing the dynamic relationship between the patient's physiological state, environmental conditions, and pain intervention plans. During operation, the system generates a series of structured data, organized into a timestamped sequence for easy analysis and decision-making. Below is a typical data recording example demonstrating key parameters collected by the system over a specific time period: This form records the patient's status changes over five minutes during radiotherapy. The system collects basic physiological parameters at 30-second intervals, while also recording environmental conditions and the operational status of the treatment equipment. Pain scores are periodically assessed and input by medical staff using a standard scale, and intervention protocol codes correspond to the pain management strategies currently being implemented by the system. The consistency index reflects the degree of data consistency from the behavior capture units; lower values ​​indicate a greater likelihood of abnormal patient behavior. The alarm status field displays the system's current judgment, categorized into four levels: normal, attentive, warning, and alarm.

[0056] The storage of historical pain intervention records adopts a hierarchical structure, comprising three main parts: a protocol summary, implementation parameters, and effect feedback. The protocol summary records the intervention type (such as medication adjustment, postural change, treatment pause, etc.) and a basic description; the implementation parameters detail the specific operational details, including quantitative indicators such as dosage, angle, and duration; and the effect feedback collects patient response data after the intervention. These records are organized along a timeline to form a complete treatment history map.

[0057] When the system detects a significant deviation between the optimal pain intervention plan and the preset monitoring parameter values, it initiates a parameter update process. This process first analyzes the nature and extent of the deviation, distinguishing between temporary fluctuations and persistent changes. For persistent deviations, the system assesses the applicability of the existing parameter strategy and generates adjustment suggestions. Parameter updates employ a gradual strategy, first testing the effectiveness of new parameters on a small scale, and then applying them comprehensively after confirming their safety and effectiveness. Each parameter adjustment is recorded with detailed decision-making basis and operation logs, forming a traceable change history.

[0058] The alarm mechanism employs a multi-condition combination judgment logic for triggering. The system continuously monitors three key dimensions: abnormality of physiological indicators, environmental risk level, and deviation of intervention effect. Each dimension has a set basic threshold and dynamic adjustment range to avoid false alarms caused by sudden changes in a single indicator. Alarm levels progress according to severity, from suggestive notifications to emergency interruption of treatment. High-level alarms will simultaneously activate multiple response channels, including visual alerts, audible prompts, and remote notifications.

[0059] In specific applications within a radiotherapy room, the system may encounter the following typical situations: When a patient experiences limb twitching due to treatment pain, the consistency index of the behavior capture unit will significantly decrease, while physiological parameters will show an increased heart rate and elevated skin conductance. The system compares historical data and finds a significant deviation between the current state and preset parameters. After optimization calculations, it recommends adjusting the pressure parameters of the body positioning device. Once confirmed by medical staff, the system automatically updates the parameters and continuously monitors the improvement. If no expected improvement is observed within a certain period, the alarm level is escalated and further measures are recommended.

[0060] The maintenance interface for monitoring parameter configurations provides an intuitive user experience. Healthcare professionals can view current parameter values, historical trends, and system suggestions. Parameter adjustments support both manual input and system recommendations; important modifications require double confirmation. The interface uses color coding to distinguish different parameter categories, and critical safety parameters are protected with locks to prevent accidental operations. All parameter changes are recorded with operator information and the reason for modification, complying with medical quality management standards.

[0061] The system is integrated with the radiotherapy equipment through a standardized communication protocol. Monitoring parameters can be synchronized to the treatment control system in real time, automatically adjusting the equipment's operating mode when necessary. This deep integration ensures rapid response to emergencies while maintaining the independent security of each system. Data exchange employs encryption and verification mechanisms to prevent information tampering or loss.

[0062] Maintenance personnel can view the system's operational status through a dedicated diagnostic interface. This interface displays the working logs, resource usage, and error records of each module. Regularly generated system health reports summarize key performance indicators to help identify potential problems promptly. Software updates employ differentiated version management, ensuring that different components can be upgraded independently without affecting overall functionality.

[0063] The data backup strategy employs a three-tier storage architecture. Real-time data is stored in a high-speed cache for current treatment use; daily summary data is transferred to online storage, supporting mid-term queries and analysis; and long-term archived data is stored offline to ensure the permanent preservation of important records. The backup process is executed automatically and its integrity is verified, with multiple copies of critical data retained for recovery options.

[0064] The system demonstrated excellent adaptability in actual operation. The monitoring parameters could automatically switch between preset protocols to suit the characteristics of different treatment stages. For example, a more sensitive detection threshold was used in the early stages of treatment to detect discomfort early; during the stable phase, the standards were appropriately relaxed to reduce unnecessary intervention. This dynamic adjustment capability enabled the system to optimize the treatment process while ensuring safety.

[0065] Example 5: The interference suppression module is designed to address the complex electromagnetic environment generated during the operation of radiotherapy equipment. Multi-level signal processing technology ensures the data acquisition quality of the behavior sensor. This module is directly connected to the behavior recording device, forming a closed-loop electromagnetic interference management system that automatically activates the protection mechanism when a radiotherapy equipment start-up signal is detected.

[0066] The system employs a pre-triggered monitoring mode to establish a baseline reference before the radiotherapy equipment is officially started. When a equipment readiness signal is detected, the interference suppression module immediately activates a preset data acquisition window, continuously recording the raw outputs of each sensor. The duration of this baseline acquisition phase is configured based on the specific electromagnetic environment characteristics of the treatment room, typically covering the complete transition from standby to full operation. During this period, the system analyzes the fluctuation range and spectral characteristics of each sensor signal, establishing a personalized noise characteristic profile. This baseline data is stored in a buffer as a reference standard for subsequent interference assessment.

[0067] After the radiotherapy equipment is started, the interference suppression module enters real-time monitoring mode. The system adopts a parallel processing architecture, simultaneously tracking changes in multiple key parameters. The instantaneous offset amplitude of the sensor signal is sampled at high speed through a dedicated hardware channel, with the sampling frequency automatically adjusted according to the signal type to ensure the capture of transient characteristics of electromagnetic interference. The data at each sampling point is compared with a reference fluctuation range to calculate the relative rate of change. The system not only focuses on amplitude changes but also analyzes the degree of signal waveform distortion to comprehensively assess the impact of interference.

[0068] When an interference signal exceeding a threshold is detected, the system initiates a multi-level suppression strategy. The primary suppression measure is to inject a reverse calibration signal into the affected sensor. These calibration signals are dynamically generated based on the interference characteristics and contain components with the opposite phase to the original interference. The signal injection process employs a gradual adjustment method, with the initial intensity set as a fixed proportion of the interference amplitude, and then gradually adjusted based on feedback. The system continuously monitors the signal improvement after injection until the offset amplitude returns to an acceptable range.

[0069] The dynamic noise reduction algorithm employs different processing strategies based on the type of interference. For broadband random noise, the system increases the number of sampling points and smooths the data using digital filtering; for periodic pulse interference, adaptive notch filtering is used to eliminate specific frequency components; for transient strong interference, the sampling frequency is temporarily reduced to avoid data saturation. These techniques can be applied individually or in combination to form targeted solutions. The adjustment parameters are dynamically optimized based on real-time monitoring results, seeking a balance between noise suppression and signal fidelity.

[0070] The hardware implementation of the electromagnetic shielding mechanism comprises several key components. At the physical level, the behavior recording device employs a layered shielding structure, with an outer layer of highly conductive metal and an inner layer of high-permeability material. Key circuit boards utilize an independent shielding chamber design, and sensitive signal lines use twisted-pair cables coupled with filters. In terms of circuit design, differential signal transmission is employed to enhance common-mode rejection. The power supply system is equipped with a multi-stage filtering network to block the propagation path of conducted interference.

[0071] The system software architecture comprises three main functional blocks: interference feature identification, suppression strategy selection, and effect evaluation. The feature identification module analyzes the time and frequency domain characteristics of the sensor signal, matching the current interference pattern with typical cases in the knowledge base. The strategy selection module generates a processing plan based on the identification results and predicts the possible effects. The effect evaluation module tracks the execution results of suppression measures in real time and triggers strategy adjustments when necessary. These three modules form a closed-loop control system, achieving intelligent interference management.

[0072] The maintenance interface provides detailed recording and statistical analysis functions for interference events. Each interference event is categorized and stored, including information such as the time of occurrence, duration, degree of impact, and measures taken. The system can generate temporal distribution charts of interference patterns to help technicians understand the changing patterns of the electromagnetic environment in the treatment room. The calibration log records the parameters and effects of all reverse signal injections, providing a reference for subsequent optimization.

[0073] The system performs continuous self-testing procedures during operation. Noise floor levels of each sensor are periodically tested to verify the integrity of the shielding structure. The operating status of critical circuits is monitored in real time, and any abnormalities are immediately reported. Communication links undergo periodic quality checks to ensure reliable data transmission. These self-testing measures operate as background tasks and do not affect normal monitoring functions.

[0074] Upgradeability and scalability are key considerations in the system design. The interference signature library supports online updates, allowing for the addition of new interference handling schemes. The suppression algorithm employs a modular design, facilitating the replacement or enhancement of specific functions. Sufficient margins are reserved in the hardware interfaces to support the integration of future new sensors. System version upgrades utilize a seamless switching mechanism, ensuring uninterrupted treatment.

[0075] In real-world radiotherapy scenarios, the system demonstrates stable anti-interference performance. When devices such as linear accelerators generate strong electromagnetic pulses, the interference suppression module responds quickly, maintaining the validity of behavioral monitoring data. Sensor signals remain available during interference, preventing data loss from affecting pain assessment. The system exhibits good adaptability to radiotherapy equipment from different manufacturers and can identify interference patterns with various characteristics.

[0076] Healthcare professionals can understand the system's operational status through a simple interface. Interference events are presented in a non-technical manner, highlighting the extent of their impact on monitoring functions. Important alarms are accompanied by clear handling suggestions, reducing operational complexity. Most system operating parameters are automatically adjusted, requiring manual confirmation only when necessary.

[0077] Data security measures are implemented throughout the entire processing flow. Raw sensor data is encrypted at the acquisition point, and verification mechanisms are used during transmission to prevent tampering. Interference event logs contain complete contextual information, supporting post-event analysis. System logs are archived regularly to meet compliance requirements for medical data management.

[0078] By combining hardware shielding, signal processing, and intelligent algorithms, the electromagnetic interference problem in the radiotherapy environment is effectively solved. The system design emphasizes practicality and reliability, improving the quality of monitoring data without increasing operational complexity.

[0079] 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.

[0080] 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.

Claims

1. An intelligent pain behavior monitoring system for radiotherapy patients, applied in a radiotherapy environment, wherein, The radiotherapy environment includes multiple patient monitoring points, characterized in that the system includes: The behavior monitoring module includes multiple behavior capture units and a behavior consistency assessment unit. The multiple behavior capture units are respectively arranged at the multiple patient monitoring points. The behavior consistency assessment unit is used to receive the output signal of the behavior capture unit and calculate the behavior consistency index of the behavior capture unit accordingly. The pain recognition module is connected to the behavior consistency assessment unit and is used to determine whether a pain behavior capture operation needs to be performed based on the behavior consistency index of each behavior capture unit, and to collect the pain behavior feature information of the behavior capture unit when it is determined that a pain behavior capture operation should be performed. A behavior control module, connected to the pain recognition module, is used to process the pain behavior characteristic information to deduce the optimal pain intervention plan for the patient, and adjust the monitoring parameter configuration of the radiotherapy environment according to the optimal pain intervention plan.

2. The intelligent radiotherapy patient pain behavior monitoring system according to claim 1, characterized in that, The behavior capture unit includes a behavior recording device installed around the patient monitoring point. The behavior recording device maintains a fixed relative position with the patient monitoring point. The behavior recording device integrates multiple behavior sensors and is also equipped with a positioning marker.

3. The intelligent radiotherapy patient pain behavior monitoring system according to claim 2, characterized in that, The behavior consistency evaluation unit calculates the behavior consistency index of the behavior capture unit based on the output signal of the behavior capture unit. The specific process is as follows: For each behavior sensor, the original behavior data acquired by the behavior sensor in the continuous time series is processed by time window aggregation to generate aggregated behavior data and discrete behavior components. The aggregated behavior data and discrete behavior components are combined using a filtering algorithm to output the optimized behavior data in the continuous time series. Based on the optimized behavior data in a continuous time series corresponding to each behavior sensor, the behavior consistency index of the behavior capture unit is calculated.

4. The intelligent radiotherapy patient pain behavior monitoring system according to claim 3, characterized in that, The behavior consistency evaluation unit calculates the behavior consistency index of the behavior capture unit based on the optimized behavior data in the continuous time series corresponding to each behavior sensor. The specific process is as follows: for each behavior sensor, multiple data points at multiple sampling times are selected from the optimized behavior data in the continuous time series corresponding to the behavior sensor, and the behavior data values ​​of the behavior sensor at the multiple sampling times are obtained; for each sampling time, based on the behavior data values ​​of all the behavior sensors at the sampling time, a single-point behavior consistency metric corresponding to that sampling time is calculated; and the behavior consistency index of the behavior capture unit is determined by combining the single-point behavior consistency metrics corresponding to all the sampling times.

5. The intelligent radiotherapy patient pain behavior monitoring system according to claim 4, characterized in that, The pain recognition module determines whether to perform a pain behavior capture operation based on the behavior consistency index of each behavior capture unit. The specific process is as follows: select any two groups of behavior capture units and calculate the difference in behavior consistency index between the two groups of behavior capture units; calculate the behavior consistency fluctuation coefficient based on the difference in behavior consistency index between all behavior capture units; when the behavior consistency fluctuation coefficient is less than a preset fluctuation threshold and the behavior consistency index of any behavior capture unit exceeds a preset consistency standard, it is determined that a pain behavior capture operation should be performed.

6. The intelligent radiotherapy patient pain behavior monitoring system according to any one of claims 2 to 5, characterized in that, The behavior recording device is also equipped with multiple calibration points; the pain recognition module collects pain behavior feature information of the behavior capture unit, including: collecting behavior feature information of the behavior capture unit from multiple observation positions; the behavior control module derives the optimal pain intervention plan for the patient based on the pain behavior feature information, including: for each observation position, analyzing the geometric correlation features between the key behavior points and the multiple calibration points based on the behavior feature information collected from that observation position; integrating the geometric correlation features between the key behavior points and the multiple calibration points corresponding to all observation positions to construct a global geometric model between the key behavior points and the multiple calibration points; and deriving the optimal pain intervention plan for the patient based on the global geometric model between the key behavior points and the multiple calibration points.

7. The intelligent radiotherapy patient pain behavior monitoring system according to claim 6, characterized in that, The behavior control module derives the optimal pain intervention plan for the patient based on a global geometric model between the key behavioral points and multiple calibration points. The specific process is as follows: generating multiple potential pain intervention plans based on the global geometric model; acquiring the patient's physiological state data at continuous time points; and acquiring environmental condition data at continuous time points. By combining the patient's physiological state data and environmental condition data at consecutive time points, the multiple potential pain intervention programs are optimized and screened to determine the optimal pain intervention program.

8. The intelligent radiotherapy patient pain behavior monitoring system according to claim 7, characterized in that, It also includes an alarm activation module, which is connected to the behavior control module and is used to trigger an alarm mechanism in the radiotherapy environment by using the patient's physiological state data at continuous time points, the environmental condition data at continuous time points, and the optimal pain intervention plan.

9. The intelligent radiotherapy patient pain behavior monitoring system according to claim 8, characterized in that, The behavior control module adjusts the monitoring parameter configuration of the radiotherapy environment according to the optimal pain intervention plan. The specific process is as follows: a preset monitoring parameter strategy is generated by referring to historical pain intervention records. The preset monitoring parameter strategy includes preset monitoring parameter values ​​for different time periods. When the deviation between the optimal pain intervention plan and the preset monitoring parameter values ​​for the corresponding time period exceeds a preset deviation limit, the monitoring parameter configuration and the preset monitoring parameter strategy are updated according to the optimal pain intervention plan.

10. The intelligent radiotherapy patient pain behavior monitoring system according to claim 2, characterized in that, It also includes an interference suppression module, which is connected to the behavior recording device and is used to activate an electromagnetic shielding mechanism when a radiotherapy device start signal is detected. The specific execution process is as follows: acquiring the original data fluctuation range of the behavior sensor within a preset period before the radiotherapy device is started; monitoring the instantaneous data offset amplitude of the behavior sensor in real time after the radiotherapy device is started; if the instantaneous data offset amplitude exceeds a preset proportional threshold of the original data fluctuation range, injecting a reverse calibration signal into the behavior sensor; and dynamically adjusting the acquisition frequency of the behavior sensor based on the reverse calibration signal until the instantaneous data offset amplitude returns to the original data fluctuation range.

Images