Cardiovascular rehabilitation power vehicle control system based on electrocardiosignals

Abstract

The invention provides a cardiovascular rehabilitation power vehicle control system based on electrocardiosignals, the power vehicle control system comprises a closed-loop heart rate control system combining heart rate control with a power vehicle, and the safety and effectiveness of training are ensured by enabling the actual heart rate in the rehabilitation training process to be matched with the expected target heart rate; the control system establishes a heart rate change identification model according to the heart rate dynamic characteristics, determines parameters of the heart rate change identification model through a system identification experiment, and then researches and designs a heart rate feedback controller according to the heart rate change identification model; an electrocardio signal processing and analysis algorithm including an electrocardio quality evaluation algorithm and an atrial fibrillation detection algorithm is added into the control system; according to the system, a more personalized, safer and more efficient training scheme can be provided for cardiovascular patients, the functions of real-time collection and storage of electrocardiosignals, electrocardiogram quality evaluation, atrial fibrillation detection, real-time heart rate monitoring and control and the like can be achieved, and safety and effectiveness in the cardiovascular rehabilitation training process are ensured.

Description

Technical Field

[0001] This invention relates to the field of rehabilitation training equipment technology, and in particular to a cardiovascular rehabilitation power vehicle control system based on electrocardiogram signals. Background Technology

[0002] Cardiovascular disease (CVD) is the most common non-communicable disease worldwide and a leading cause of death, accounting for approximately one-third of all deaths globally, posing a significant health and economic burden globally. With an aging population and the prevalence of unhealthy lifestyles, the burden of CVD continues to increase. Therefore, the prevention, diagnosis, and treatment of CVD have become crucial global issues. Cardiac rehabilitation training is an important component of CVD treatment, significantly improving patients' exercise capacity, mitigating cardiovascular risk factors, enhancing quality of life, and reducing mortality. Regular aerobic exercise has significant benefits in improving the health and quality of life of healthy adults and some CVD patients. Exercise-based cardiac rehabilitation can effectively improve cardiopulmonary function in CVD patients. Cardiopulmonary endurance refers to the body's ability to sustain activity, comprehensively reflecting the body's capacity to absorb, transport, and utilize oxygen. Studies have found that cardiopulmonary endurance (CRF) helps reduce the risk of cardiovascular and other diseases. Good cardiopulmonary endurance not only prevents cardiovascular and other diseases but also ensures the body's ability to work effectively for extended periods and recover bodily functions.

[0003] Stationary bike testing is a scientific method for assessing cardiorespiratory endurance. The user's exercise load can be adjusted through small changes in power intensity. Stationary bike exercise has been widely used in cardiac rehabilitation for patients and physical training for healthy subjects. Heart rate is used as a measure of exercise intensity in exercise prescriptions. To match the actual heart rate during exercise with the target heart rate, the workload (e.g., exercise intensity) needs to be dynamically adjusted. If the heart rate during exercise is lower than the target heart rate, the workload should be increased; if the heart rate exceeds the target heart rate, the workload should be decreased. Compared to exercise methods such as treadmills, stationary bikes are a form of active exercise. If the subject experiences discomfort or an excessively high heart rate during the exercise test, they can stop the test voluntarily. Therefore, stationary bike testing is characterized by high safety. Furthermore, stationary bikes are relatively less expensive and easier to transport than treadmills, and it is easy to monitor indicators such as ECG and blood pressure during the test. Because stationary bikes provide a stable seated position, they are the preferred equipment in exercise testing and prescriptions for cardiac rehabilitation. Cardiovascular rehabilitation stationary bike control systems that combine heart rate control with stationary bike functionality demonstrate significant advantages in cardiac rehabilitation. Through real-time heart rate monitoring and dynamic power adjustment, more personalized, safe, and efficient training programs can be provided for cardiovascular patients.

[0004] The most common approach to heart rate control during exercise involves real-time monitoring of heart rate changes using heart rate monitoring devices. Subjects then adjust their exercise intensity to adapt to these changes, increasing the exercise load when the heart rate is low and decreasing it when it is high, thus achieving a target heart rate and maintaining stability. Racinais et al. manually adjusted the exercise load every 30 seconds in their experiments to keep the heart rate stable at the target level. However, this subjective control method is susceptible to individual physiological differences and sensitivity during practical application, leading to insufficient control precision and unsatisfactory results. Other studies have employed open-loop heart rate control methods. Zhou Xu et al. proposed an open-loop heart rate control method based on a linear heart rate-load model established during exercise. This method calculates the corresponding exercise load based on a preset target heart rate and then guides the user through training by setting the appropriate exercise load. However, the models constructed by this type of open-loop control strategy have limitations. They struggle to accurately reflect the quantitative relationship between heart rate and exercise load and are also affected by the volatility and variability of heart rate response, resulting in significant heart rate control errors for some users.

[0005] Based on the above problems, some researchers have attempted to use closed-loop control strategies to achieve precise heart rate control during exercise. This involves real-time monitoring of the current heart rate and feedback to automatically adjust the exercise load, achieving the goal of heart rate tracking. In academic research, nonlinear controller structures and control strategies are frequently employed. Cheng et al. established a second-order nonlinear model of the heart rate response during exercise and achieved precise heart rate control based on an optimal control strategy combining linear quadratic form and H∞. Nguyen et al. studied nonlinear neural network control methods, achieving good heart rate tracking performance during exercise. However, the model parameters of these nonlinear control techniques are difficult to estimate accurately, and stability analysis is complex, making them unsuitable for practical use in real-world exercise. Most commercial training devices are based on linear time-invariant (LTI) controllers. Kawada et al. experimentally verified a simple single linear time-invariant PI controller applied to a power bike, providing accurate and stable heart rate control in both healthy individuals and patients with heart disease. Argha et al. combined an LTI-based PID controller with auditory biofeedback signals for heart rate control during power bike exercise. Hunt et al. experimentally demonstrated that a very simple LTI controller can achieve high-precision, stable, and robust heart rate control. However, these studies did not consider the issue of electrocardiogram monitoring during training, and there was relatively little description of the actual training system.

[0006] To address the shortcomings of the aforementioned methods, this patent constructs a cardiovascular rehabilitation power vehicle control system based on electrocardiogram (ECG) signals. This system enables real-time acquisition and storage of ECG signals, ECG quality assessment and atrial fibrillation detection, and real-time heart rate monitoring and control, ensuring the safety and effectiveness of cardiovascular rehabilitation training. Summary of the Invention

[0007] This invention proposes a cardiovascular rehabilitation power vehicle control system based on electrocardiogram (ECG) signals, which can realize functions such as real-time acquisition and storage of ECG signals, ECG quality assessment and atrial fibrillation detection, and real-time heart rate monitoring and control, ensuring the safety and effectiveness of cardiovascular rehabilitation training.

[0008] The present invention adopts the following technical solution.

[0009] A cardiovascular rehabilitation trolley control system based on electrocardiogram (ECG) signals is disclosed. The trolley control system includes a closed-loop heart rate control system that integrates heart rate control with the trolley for cardiac rehabilitation training. The system studies closed-loop heart rate control methods through trolley motion to ensure that the actual heart rate during rehabilitation training matches the expected target heart rate, thereby ensuring the safety and effectiveness of the training. Based on heart rate dynamics characteristics, the control system establishes a heart rate variability identification model and determines the parameters of the model through system identification experiments. Then, a heart rate feedback controller is designed based on the heart rate variability identification model. Furthermore, ECG signal processing and analysis algorithms, including ECG quality assessment algorithms and atrial fibrillation detection algorithms, are incorporated into the control system.

[0010] The power vehicle control system includes an electrocardiogram acquisition device (Polar H10 heart rate monitor), a host computer (PC), and the power vehicle;

[0011] When the power vehicle is used for rehabilitation training, its control system collects the subject's electrocardiogram (ECG) signals in real time and performs quality assessments, removing low-quality signals with significant noise interference and retaining high-quality signals. An atrial fibrillation detection algorithm is used to monitor for potential abnormal ECG problems during training. The overall system architecture is as follows: Figure 1 As shown in the diagram. The host computer comprises a Vue-based web frontend and a Flask backend service. It interacts with the ECG acquisition device and the power vehicle via Bluetooth. Specifically, the host computer receives ECG and heart rate data from the ECG acquisition device, processes and analyzes the ECG signals, calculates the output of the heart rate controller, and interacts with the power vehicle via Bluetooth serial communication to achieve closed-loop heart rate control. Simultaneously, it displays the ECG waveform and analysis results in real-time through a browser interface.

[0012] The specific steps for using the power vehicle control system include:

[0013] Step 1: The subject wears the heart rate strap of the ECG acquisition device properly in the center of the chest and adjusts the tightness of the strap to ensure wearing comfort and signal acquisition stability.

[0014] Step 2: Open the browser interface of the power vehicle control system and set the rehabilitation training duration and target heart rate value;

[0015] Step 3: Click the Connect Device button. The host computer of the power vehicle control system will start scanning for nearby Bluetooth devices. When the heart rate belt of the corresponding subject is detected, the scanning will stop and the device name will be displayed on the interface.

[0016] Step 4: Click the "Receive Data" button. The host computer will start receiving the real-time ECG signal collected by the heart rate belt and dynamically draw the ECG waveform from left to right on the interface.

[0017] Step 5: Click the Start Training button. The subject uses the exercise bike to exercise. The front-end interface of the exercise bike accesses the back-end server of the host computer. The back-end server runs the heart rate control algorithm, outputs the target power value and sends it to the exercise bike. It also receives the exercise status response from the exercise bike and returns the target power and exercise status to the front-end interface for display. At the same time, the back-end server also performs analysis and processing including ECG signal preprocessing, quality assessment and atrial fibrillation detection, and returns the analysis results to the front-end interface for display.

[0018] Step 6: After the rehabilitation training is completed, the power vehicle control system saves the training information and displays the training results to the subject.

[0019] A closed-loop heart rate control system includes a system identification method and a heart rate controller, using output power as the control variable and the current actual heart rate (HR). actual With target heart rate HR target The comparison allows the heart rate controller to calculate a target power P. target The data is then sent to the power vehicle, and when the subject uses the power vehicle for rehabilitation training exercises, different target power levels P are recorded. target Changing the current exercise state, thereby changing the heart rate; and under the action of the heart rate feedback controller, the HR is adjusted. actual Maintain in HR target Fluctuations.

[0020] In the system identification method, the target power and actual heart rate (P) are used. target →HR actual This constitutes the identification model; the direct impact of power changes is considered as a change in motion load.

[0021] First, establish a kinetic model for the dynamic change of heart rate during exercise with respect to exercise load; based on the phase of exercise physiology, model the kinetic model of heart rate change as a first-order linear time-invariant transfer function, as shown in Equation 1.

[0022] Equation 1;

[0023] Where is the steady-state gain, is the time constant.

[0024] The identification model of system identification works in open-loop mode, and the open-loop structure of the identification model is as Figure 2 shown. Specifically, during the system identification process, first give a target power curve to the ergometer, and then the internal controller of the ergometer automatically adjusts the actual output power to fit the target power. When the subject uses the adjusted ergometer for rehabilitation training exercise, record the heart rate change in real time during the exercise; after the system identification experiment is completed, calculate the relevant parameters of the P target →HR actual identification model offline through the collected input target power P target and output target heart rate HR actual data.

[0025] The system identification target power curve is set in the form of a square wave, as Figure 3 shown. The subject performs ergometer exercise and records the heart rate change with a sampling period T s = 5s. The whole test process includes a 3-minute warm-up stage at the beginning, an intermediate stage, and a 3-minute relaxation stage at the end. During the intermediate stage, the target power alternates between low power (P low ) and high power (P high ) every 3 minutes. After the test, select the P target and HR actual data in the time period of 3min < t < 18min as the input and output data for system identification respectively.

[0026] During the system identification process, the target power in the intermediate stage alternates between low power P low and high power P high ; the set values of low power P low and high power P high are set for different subjects and obtained through preliminary experiments.

[0027] The specific preliminary experiment is as follows: during the process of the subject performing ergometer exercise, manually adjust the set value of the target power and observe the change of the actual heart rate; when the actual heart rate of the subject stabilizes at HR target , the corresponding target power is used as the set value of the intermediate power P m , Plow and P high With P m The relationship is shown in Equations 2 and 3 below.

[0028] Formula 2;

[0029] Formula 3;

[0030] When HR target Set to maximum heart rate (HR) max When it is 60%, the specific calculation method is shown in Equations 4 and 5 below.

[0031] Formula 4;

[0032] Formula 5.

[0033] During the system identification process, the input and output data of the original data are first obtained with a sampling period of 5 seconds. Then, the MATLAB system identification toolbox is used for data processing and analysis. The specific process is as follows: the original data is detrended, the parameters of the identification model are obtained by using the least squares method, and the original data is input into the identification model to obtain the simulation output results. The actual output and simulation results are compared, and the effectiveness of the model is verified by two indicators: goodness-of-fit (%fit) and root mean square error (RMSE).

[0034] After completing system identification and obtaining the dynamic response model of the subject's heart rate to power input, the closed-loop heart rate control system operates in closed-loop mode based on the negative feedback mechanism. The heart rate feedback closed-loop control structure adopts the framework of a PI controller.

[0035] Specifically, an electrocardiogram (ECG) acquisition device is used to receive the subject's actual heart rate data in real time, and this data is used as a feedback signal. This data is compared with a preset target heart rate, and the error between the two is used as the input to a heart rate controller. The controller outputs an updated target power P every 5 seconds, a preset interval. target The command is sent to the power vehicle, thereby dynamically adjusting the output power P of the power vehicle. actual The subjects performed power cycling exercises, and their actual heart rate changed accordingly and was continuously monitored, thus forming a closed loop to achieve closed-loop tracking and control of heart rate.

[0036] like Figure 4As shown, the negative feedback mechanism uses proportional-integral control, also known as PI control. It is one of the most classic and widely used feedback control strategies in industrial control and biomedical engineering. While ensuring the elimination of steady-state error, it takes into account simplicity, robustness and safety. It is particularly suitable for physiological closed-loop control tasks such as heart rate, which are slow-dynamic, have low noise tolerance and have continuous disturbances.

[0037] The PI controller effectively improves the dynamic response performance of the system and eliminates steady-state deviation by combining the proportional effect of the current error with the integral effect of the historical error.

[0038] Let e(t) represent the heart rate tracking error at time t, as defined in Formula 6 below; where r(t) represents the target heart rate input and y(t) represents the actual heart rate feedback, both in bpm.

[0039] The output u(t) of the PI controller represents the target power of the power vehicle at time t, in W, and is calculated as shown in Formula 7 below.

[0040] Formula 6;

[0041] Formula 7;

[0042] Among them, K p K represents the proportional gain. i These two parameters represent the integral gain. The PI controller uses these two parameters to adjust the system output so that it tracks the set value.

[0043] Methods for setting the parameters of a heart rate controller also include: numerically simulating the heart rate controller by building a Simulink simulation system, within a preset parameter search range (K). p ∈[0.5, 2.0], K i ∈[0.01, 0.4]), traverse multiple groups (K) p , K i The parameters were combined and tested to evaluate their response performance to a step change in target heart rate (from 80 bpm to 100 bpm). The step response and the corresponding target power change of different parameter groups were compared. With the constraints of step response overshoot <5 bpm and settling time (±5 bpm) <180 s, the parameter group with the smallest steady-state error of heart rate step response and the smoothest fluctuation of target power output was selected as the parameters of the final heart rate controller.

[0044] After the subject begins exercising on the power bike, the system calculates and dynamically updates the target power command in real time based on the aforementioned PI controller, thereby enabling the actual heart rate to track the target heart rate and forming a closed-loop control.

[0045] In existing cardiac rehabilitation research, HR target It is usually set to the individual's maximum heart rate (HR). max The target heart rate is 60%–80%, which aligns with the American Heart Association (AHA) and American Association for Cardiopulmonary Rehabilitation (AACVPR) recommendations for moderate-intensity aerobic training, aiming to balance training effectiveness and safety. The heart rate closed-loop control system proposed in this invention targets the heart rate HR... target Set as the subject's maximum heart rate (HR) max This intensity is 60% of the recommended low to moderate exercise load for cardiac rehabilitation and is suitable for the early rehabilitation training phase of patients with stable cardiovascular disease.

[0046] The cardiovascular rehabilitation scooter control system includes a Bluetooth communication module, a network transmission module, a data processing module, and a data display module; the system functional structure is as follows: Figure 5 As shown,

[0047] The Bluetooth communication module is one of the key components of the system, mainly responsible for two tasks: First, it pairs and connects the host computer with the ECG acquisition device based on the Bluetooth Low Energy protocol, ensuring that the web front-end can receive ECG data from the device in real time and parse the data according to a specific data packet format, thereby ensuring the accuracy and integrity of the data; Second, it establishes a virtual serial connection between the host computer and the power vehicle based on the Bluetooth serial port pass-through protocol, realizing the reliable issuance of control commands and the real-time feedback of the power vehicle's operating status, including power, speed, and torque information.

[0048] The network transmission module is used to establish a data transmission channel between the Web front-end and the Flask back-end. The front-end can upload the collected ECG data and heart rate data to the back-end for processing and analysis, and receive the results returned by the back-end.

[0049] The data processing module is used by the backend server to perform quality assessment, anomaly detection, processing, and analysis on the received ECG data, and at the same time input the received heart rate data into the heart rate controller for calculation to obtain the control quantity;

[0050] The data display module has the functions of real-time ECG waveform drawing, ECG indicator display, and power vehicle operation status display. It can also pop up a reminder to the user when an abnormal ECG is detected.

[0051] The training system combining heart rate control and power vehicle proposed in this invention shows significant advantages in cardiac rehabilitation. Through real-time heart rate monitoring and dynamic power adjustment, it can provide cardiovascular patients with more personalized, safe and efficient training programs.

[0052] This invention patent describes a closed-loop heart rate control method based on power bike exercise research, ensuring that the actual heart rate during training matches the expected target heart rate, thereby guaranteeing the safety and effectiveness of training. First, a heart rate change identification model is established based on heart rate dynamics characteristics, and the model parameters are determined through system identification experiments. Then, a heart rate feedback controller is designed based on this model. Simultaneously, electrocardiogram (ECG) signal processing and analysis algorithms are incorporated into the system, including ECG quality assessment and atrial fibrillation detection. During the user's power bike exercise, ECG signals are collected in real time and their quality is assessed, removing low-quality signals with significant noise interference and retaining high-quality signals. The atrial fibrillation detection algorithm monitors for potential abnormal ECG problems during training in real time.

[0053] This invention patent also has the following advantages and uses:

[0054] (1) The cardiovascular rehabilitation power vehicle control system proposed in this patent integrates electrocardiogram quality assessment and atrial fibrillation detection algorithms. It can process and analyze the electrocardiogram signals of the subjects in real time during cardiovascular rehabilitation training, and remind the subjects when risks occur, so as to ensure safety during the training process.

[0055] (2) This patent adopts a closed-loop heart rate control strategy. The heart rate dynamics model is obtained through system identification experiments. Based on this, the heart rate controller is designed to provide accurate and stable heart rate control.

[0056] (3) This patented system adopts a front-end and back-end separation architecture for independent development. The back-end is responsible for ECG signal processing and analysis, heart rate control algorithm execution and device communication, while the front-end focuses on user interaction, data visualization and training status monitoring, which improves the maintainability and scalability of the system. Attached Figure Description

[0057] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments:

[0058] Appendix Figure 1 This is a schematic diagram of the overall system architecture in an embodiment of the present invention;

[0059] Appendix Figure 2 This is a schematic diagram of the open-loop structure of the identification model in an embodiment of the present invention;

[0060] Appendix Figure 3 This is a schematic diagram of the target power curve identified by the system in an embodiment of the present invention;

[0061] Appendix Figure 4 This is a schematic diagram of the heart rate feedback closed-loop control structure in an embodiment of the present invention;

[0062] Appendix Figure 5 This is a schematic diagram of the system functional structure in an embodiment of the present invention;

[0063] Appendix Figure 6 This is a schematic diagram of the electrocardiogram signal data processing flow in an embodiment of the present invention;

[0064] Appendix Figure 7 This is a schematic diagram of the heart rate controller output calculation process in an embodiment of the present invention. Detailed Implementation

[0065] As shown in the figure, a cardiovascular rehabilitation power vehicle control system based on electrocardiogram (ECG) signals is presented. The power vehicle control system includes a closed-loop heart rate control system that integrates heart rate control with the power vehicle for cardiac rehabilitation training. The system studies closed-loop heart rate control methods through power vehicle motion to ensure that the actual heart rate during rehabilitation training matches the expected target heart rate, thereby ensuring the safety and effectiveness of the training. Based on heart rate dynamics characteristics, the control system establishes a heart rate change identification model and determines the parameters of the model through system identification experiments. Then, a heart rate feedback controller is designed based on the heart rate change identification model. Furthermore, ECG signal processing and analysis algorithms, including ECG quality assessment algorithms and atrial fibrillation detection algorithms, are incorporated into the control system.

[0066] The power vehicle control system includes an electrocardiogram acquisition device (Polar H10 heart rate monitor), a host computer (PC), and the power vehicle;

[0067] When the power vehicle is used for rehabilitation training, its control system collects the subject's electrocardiogram (ECG) signals in real time and performs quality assessments, removing low-quality signals with significant noise interference and retaining high-quality signals. An atrial fibrillation detection algorithm is used to monitor for potential abnormal ECG problems during training. The overall system architecture is as follows: Figure 1 As shown in the diagram. The host computer comprises a Vue-based web frontend and a Flask backend service. It interacts with the ECG acquisition device and the power vehicle via Bluetooth. Specifically, the host computer receives ECG and heart rate data from the ECG acquisition device, processes and analyzes the ECG signals, calculates the output of the heart rate controller, and interacts with the power vehicle via Bluetooth serial communication to achieve closed-loop heart rate control. Simultaneously, it displays the ECG waveform and analysis results in real-time through a browser interface.

[0068] The specific steps for using the power vehicle control system include:

[0069] Step 1: The subject wears the heart rate strap of the ECG acquisition device properly in the center of the chest and adjusts the tightness of the strap to ensure wearing comfort and signal acquisition stability.

[0070] Step 2: Open the browser interface of the power vehicle control system and set the rehabilitation training duration and target heart rate value;

[0071] Step 3: Click the Connect Device button. The host computer of the power vehicle control system will start scanning for nearby Bluetooth devices. When the heart rate belt of the corresponding subject is detected, the scanning will stop and the device name will be displayed on the interface.

[0072] Step 4: Click the "Receive Data" button. The host computer will start receiving the real-time ECG signal collected by the heart rate belt and dynamically draw the ECG waveform from left to right on the interface.

[0073] Step 5: Click the Start Training button. The subject uses the exercise bike to exercise. The front-end interface of the exercise bike accesses the back-end server of the host computer. The back-end server runs the heart rate control algorithm, outputs the target power value and sends it to the exercise bike. It also receives the exercise status response from the exercise bike and returns the target power and exercise status to the front-end interface for display. At the same time, the back-end server also performs analysis and processing including ECG signal preprocessing, quality assessment and atrial fibrillation detection, and returns the analysis results to the front-end interface for display.

[0074] Step 6: After the rehabilitation training is completed, the power vehicle control system saves the training information and displays the training results to the subject.

[0075] A closed-loop heart rate control system includes a system identification method and a heart rate controller, using output power as the control variable and the current actual heart rate (HR). actual With target heart rate HR target The comparison allows the heart rate controller to calculate a target power P. target The data is then sent to the power vehicle, and when the subject uses the power vehicle for rehabilitation training exercises, different target power levels P are recorded. target Changing the current exercise state, thereby changing the heart rate; and under the action of the heart rate feedback controller, the HR is adjusted. actual Maintain in HR target Fluctuations.

[0076] System identification is an important method for establishing mathematical models of dynamic systems based on experimental data, and it is widely used in fields such as control engineering, signal processing, biomedical engineering, and sports science. Its basic principle is: under known excitation (input) applied to the system, the system's response (output) data is simultaneously collected, and based on this, a parameterized model that best fits the input-output relationship is estimated from a pre-defined model structure. The advantage of this method is that it does not rely entirely on the system's internal physical mechanisms, making it particularly suitable for physiological systems with complex structures, unknown parameters, or those difficult to model accurately.

[0077] In this example, the system identification method uses target power and actual heart rate (P) as the basis for identification. target →HR actual This constitutes the identification model; the direct impact of power changes is considered as a change in motion load.

[0078] First, establish a kinetic model of the dynamic change of heart rate with exercise load during exercise; based on the phase of exercise physiology, the kinetic model of heart rate change is modeled as a first-order linear time-invariant transfer function, as shown in Equation 1.

[0079] Equation 1;

[0080] Where is the steady-state gain, is the time constant.

[0081] The identification model of system identification works in the open-loop mode, and the open-loop structure of the identification model is as Figure 2 shown. Specifically, during the system identification process, first give a target power curve to the ergometer, and then the internal controller of the ergometer automatically adjusts the actual output power to fit the target power. When the subject uses the adjusted ergometer for rehabilitation training exercise, the heart rate change is recorded in real time during the exercise; after the system identification experiment is completed, through the collected input target power P target and the output target heart rate HR actual data, calculate the relevant parameters of the P target →HR actual identification model offline.

[0082] The system identification target power curve is set in the form of a square wave, as Figure 3 shown. The subject performs ergometer exercise, and records the heart rate change with a sampling period T s = 5 s. The whole test process includes a 3-minute warm-up stage at the beginning, an intermediate stage, and a 3-minute relaxation stage at the end. During the intermediate stage, the target power alternates every 3 minutes according to the low power (P low ) and the high power (P high ). After the test, select the P target and HR actual in the time period of 3 min < t < 18 min as the input and output data of system identification respectively.

[0083] During the system identification process, the target power in the intermediate stage alternates between the low power P low and the high power P high ; the set values of the low power P low and the high power P high are set for different subjects and obtained through preliminary experiments.

[0084] The specific preliminary experiment is as follows: during the process of the subject performing ergometer exercise, manually adjust the set value of the target power and observe the change of the actual heart rate; when the actual heart rate of the subject stabilizes at HR target , the corresponding target power is used as the intermediate power P mThe setting value, P low and P high With P m The relationship is shown in Equations 2 and 3 below.

[0085] Formula 2;

[0086] Formula 3;

[0087] When HR target Set to maximum heart rate (HR) max When it is 60%, the specific calculation method is shown in Equations 4 and 5 below.

[0088] Formula 4;

[0089] Formula 5.

[0090] During the system identification process, the input and output data of the original data are first obtained with a sampling period of 5 seconds. Then, the MATLAB system identification toolbox is used for data processing and analysis. The specific process is as follows: the original data is detrended, the parameters of the identification model are obtained by using the least squares method, and the original data is input into the identification model to obtain the simulation output results. The actual output and simulation results are compared, and the effectiveness of the model is verified by two indicators: goodness-of-fit (%fit) and root mean square error (RMSE).

[0091] After completing system identification and obtaining the dynamic response model of the subject's heart rate to power input, the closed-loop heart rate control system operates in closed-loop mode based on the negative feedback mechanism. The heart rate feedback closed-loop control structure adopts the framework of a PI controller.

[0092] Specifically, an electrocardiogram (ECG) acquisition device is used to receive the subject's actual heart rate data in real time, and this data is used as a feedback signal. This data is compared with a preset target heart rate, and the error between the two is used as the input to a heart rate controller. The controller outputs an updated target power P every 5 seconds, a preset interval. target The command is sent to the power vehicle, thereby dynamically adjusting the output power P of the power vehicle. actual The subjects performed power cycling exercises, and their actual heart rate changed accordingly and was continuously monitored, thus forming a closed loop to achieve closed-loop tracking and control of heart rate.

[0093] like Figure 4As shown, the negative feedback mechanism uses proportional-integral control, also known as PI control. It is one of the most classic and widely used feedback control strategies in industrial control and biomedical engineering. While ensuring the elimination of steady-state error, it takes into account simplicity, robustness and safety. It is particularly suitable for physiological closed-loop control tasks such as heart rate, which are slow-dynamic, have low noise tolerance and have continuous disturbances.

[0094] The PI controller effectively improves the dynamic response performance of the system and eliminates steady-state deviation by combining the proportional effect of the current error with the integral effect of the historical error.

[0095] Let e(t) represent the heart rate tracking error at time t, as defined in Formula 6 below; where r(t) represents the target heart rate input and y(t) represents the actual heart rate feedback, both in bpm.

[0096] The output u(t) of the PI controller represents the target power of the power vehicle at time t, in W, and is calculated as shown in Formula 7 below.

[0097] Formula 6;

[0098] Formula 7;

[0099] Among them, K p K represents the proportional gain. i These two parameters represent the integral gain. The PI controller uses these two parameters to adjust the system output so that it tracks the set value.

[0100] Methods for setting the parameters of a heart rate controller also include: numerically simulating the heart rate controller by building a Simulink simulation system, within a preset parameter search range (K). p ∈[0.5, 2.0], K i ∈[0.01, 0.4]), traverse multiple groups (K) p , K i The parameters were combined and tested to evaluate their response performance to a step change in target heart rate (from 80 bpm to 100 bpm). The step response and the corresponding target power change of different parameter groups were compared. With the constraints of step response overshoot <5 bpm and settling time (±5 bpm) <180 s, the parameter group with the smallest steady-state error of heart rate step response and the smoothest fluctuation of target power output was selected as the parameters of the final heart rate controller.

[0101] After the subject begins exercising on the power bike, the system calculates and dynamically updates the target power command in real time based on the aforementioned PI controller, thereby enabling the actual heart rate to track the target heart rate and forming a closed-loop control.

[0102] In existing cardiac rehabilitation research, HR target It is usually set to the individual's maximum heart rate (HR). max The target heart rate is 60%–80%, which aligns with the American Heart Association (AHA) and American Association for Cardiopulmonary Rehabilitation (AACVPR) recommendations for moderate-intensity aerobic training, aiming to balance training effectiveness and safety. The heart rate closed-loop control system proposed in this invention targets the heart rate HR... target Set as the subject's maximum heart rate (HR) max This intensity is 60% of the recommended low to moderate exercise load for cardiac rehabilitation and is suitable for the early rehabilitation training phase of patients with stable cardiovascular disease.

[0103] The cardiovascular rehabilitation scooter control system includes a Bluetooth communication module, a network transmission module, a data processing module, and a data display module; the system functional structure is as follows: Figure 5 As shown,

[0104] The Bluetooth communication module is one of the key components of the system, mainly responsible for two tasks: First, it pairs and connects the host computer with the ECG acquisition device based on the Bluetooth Low Energy protocol, ensuring that the web front-end can receive ECG data from the device in real time and parse the data according to a specific data packet format, thereby ensuring the accuracy and integrity of the data; Second, it establishes a virtual serial connection between the host computer and the power vehicle based on the Bluetooth serial port pass-through protocol, realizing the reliable issuance of control commands and the real-time feedback of the power vehicle's operating status, including power, speed, and torque information.

[0105] The network transmission module is used to establish a data transmission channel between the Web front-end and the Flask back-end. The front-end can upload the collected ECG data and heart rate data to the back-end for processing and analysis, and receive the results returned by the back-end.

[0106] The data processing module is used by the backend server to perform quality assessment, anomaly detection, processing, and analysis on the received ECG data, and at the same time input the received heart rate data into the heart rate controller for calculation to obtain the control quantity;

[0107] The data display module has the functions of real-time ECG waveform drawing, ECG indicator display, and power vehicle operation status display. It can also pop up a reminder to the user when an abnormal ECG is detected.

[0108] The cardiovascular rehabilitation trolley control system proposed in this example mainly consists of three parts: an electrocardiogram acquisition device (Polar H10 heart rate monitor), a host computer (PC), and the trolley. The overall system architecture is as follows: Figure 1As shown in the diagram, the host computer consists of a Vue-based web front-end and a Flask back-end service. The host computer interacts with the ECG acquisition device and the power vehicle via Bluetooth. Specifically, the host computer receives ECG and heart rate data from the ECG acquisition device, processes and analyzes the ECG signals, calculates the output of the heart rate controller, and interacts with the power vehicle via Bluetooth serial communication to achieve closed-loop heart rate control. Simultaneously, the browser interface displays the ECG waveform and analysis results in real time.

[0109] This example demonstrates the ability to promptly identify and warn of cardiovascular risks during exercise. Considering the need for real-time processing, a 10-second ECG signal was selected for ECG processing and analysis, following the steps of preprocessing, quality assessment, and atrial fibrillation detection. The flowchart is as follows: Figure 6 As shown, the first step is to preprocess the input 10-second segment of ECG data, including resampling, bandpass filtering, and Z-score normalization. This step directly calls the corresponding Python functions. Next, the data is input into a quality assessment model to determine whether the signal segment is acceptable or unacceptable. If it is acceptable, the next step, atrial fibrillation detection, is performed; otherwise, the analysis stops. The atrial fibrillation detection model determines whether atrial fibrillation exists in the signal segment. The ECG signal quality assessment and atrial fibrillation detection algorithms use PyTorch-based deep learning algorithms. This patent integrates these algorithms into the Flask backend service and ensures efficient operation when receiving data in real time. Specifically, this patent uses the ".pth" weight files of two pre-trained deep learning models, in PyTorch state dictionary format. The model architecture is loaded and weight parameters are fused using a Python script, and finally exported as a deployable TorchScript format. The pre-trained TorchScript model is then loaded for real-time inference to obtain classification results.

[0110] In this example, for heart rate data, the error between the heart rate and a pre-set target heart rate is used as input to the heart rate controller every 5 seconds. After calculation, the target power is output and sent to the power vehicle. The flowchart is as follows. Figure 7 As shown. The heart rate controller used in this step is a Python-based PI controller.

Claims

1. A cardiovascular rehabilitation power vehicle control system based on electrocardiogram signals, characterized in that: The power vehicle control system includes a closed-loop heart rate control system that integrates heart rate control with the power vehicle. By ensuring that the actual heart rate during rehabilitation training matches the expected target heart rate, the safety and effectiveness of the training are guaranteed. Based on the heart rate dynamics characteristics, the control system establishes a heart rate change identification model and determines the parameters of the heart rate change identification model through system identification experiments. Then, a heart rate feedback controller is designed based on the heart rate change identification model. Furthermore, the control system incorporates electrocardiogram (ECG) signal processing and analysis algorithms, including ECG quality assessment algorithms and atrial fibrillation detection algorithms.

2. The cardiovascular rehabilitation power vehicle control system based on electrocardiogram signals according to claim 1, characterized in that: The power vehicle control system includes an electrocardiogram (ECG) acquisition device, a host computer, and the power vehicle itself. When the power vehicle is used for rehabilitation training, its control system collects the subject's electrocardiogram (ECG) signals in real time and assesses their quality. It uses an atrial fibrillation detection algorithm to monitor for any abnormal ECG issues that may occur during training. The host computer consists of a front-end and a back-end. The host computer interacts with the ECG acquisition equipment and the power vehicle. Specifically, the host computer receives ECG and heart rate data from the ECG acquisition equipment, processes and analyzes the ECG signals, calculates the output of the heart rate controller, and interacts with the power vehicle via Bluetooth to achieve closed-loop heart rate control. Simultaneously, the interface displays the ECG waveform and analysis results in real time.

3. The cardiovascular rehabilitation power vehicle control system based on electrocardiogram signals according to claim 2, characterized in that: The specific steps for using the power vehicle control system include: Step 1: The subject wears the heart rate strap of the ECG acquisition device properly in the center of the chest and adjusts the tightness of the strap to ensure wearing comfort and signal acquisition stability. Step 2: Open the browser interface of the power vehicle control system and set the rehabilitation training duration and target heart rate value; Step 3: The host computer of the power vehicle control system scans the surrounding devices; when the heart rate belt of the corresponding subject is scanned, the device name is displayed on the interface; Step 4: The host computer receives the real-time ECG signal collected by the heart rate belt and dynamically draws the ECG waveform from left to right on the interface; Step 5: The subject uses the exercise vehicle to exercise. The front-end interface of the exercise vehicle accesses the back-end server of the host computer. The back-end server runs the heart rate control algorithm, outputs the target power value and sends it to the exercise vehicle, and receives the exercise status response from the exercise vehicle. It returns the target power and exercise status to the front-end interface for display. At the same time, the back-end server also performs analysis and processing including ECG signal preprocessing, quality assessment and atrial fibrillation detection, and returns the analysis results to the front-end interface for display. Step 6: After the rehabilitation training is completed, the power vehicle control system saves the training information and displays the training results to the subject.

4. The cardiovascular rehabilitation power vehicle control system based on electrocardiogram signals according to claim 2, characterized in that: A closed-loop heart rate control system includes a system identification method and a heart rate controller, using output power as the control variable and the current actual heart rate (HR). actual With target heart rate HR target The comparison allows the heart rate controller to calculate a target power P. target The data is then sent to the power vehicle, and when the subject uses the power vehicle for rehabilitation training exercises, different target power levels P are recorded. target Changing the current state of exercise causes a change in heart rate; And under the action of the heart rate feedback controller, HR actual Maintain in HR target Fluctuations.

5. A cardiovascular rehabilitation power vehicle control system based on electrocardiogram signals according to claim 4, characterized in that: In the system identification method, the identification model is composed of target power and actual heart rate; the direct impact of power changes is considered as a change in exercise load. First, a dynamic model of the dynamic change of heart rate with exercise load during exercise is established. Based on the phase of exercise physiology, the dynamic model of heart rate change is modeled as a first-order linear time-invariant transfer function, as shown in Equation 1. Official 1; in For steady-state gain, is the time constant. The identification model of the system identification works in open-loop mode. During the system identification process, a target power curve is first given to the power vehicle. Then, the internal controller of the power vehicle automatically adjusts the actual output power to fit the target power. When the subject uses the adjusted power vehicle for rehabilitation training, the heart rate changes are recorded in real time during the exercise. After the system completes the identification, it uses the collected input target power P target and the target heart rate HR output actual Data, offline calculation of relevant parameters of the identification model.

6. The cardiovascular rehabilitation power vehicle control system based on electrocardiogram signals according to claim 5, characterized in that: During the system identification process, the target power in the intermediate stage is based on the low power P. low and high power P high Alternating changes; low power P low and high power P high The set values ​​were determined for different subjects and obtained through preliminary experiments. The preliminary experiment involved manually adjusting the target power setting during the exercise on a stationary bike and observing changes in the actual heart rate. The experiment concluded when the subject's actual heart rate stabilized at HR. target At that time, the corresponding target power is taken as the intermediate power P. m The setting value, P low and P high With P m The relationship is shown in Equations 2 and 3 below. Official 2; Official 3; When HR target Set to maximum heart rate (HR) max When it is 60%, the specific calculation method is shown in Equations 4 and 5 below. Official 4; Official 5.

7. The cardiovascular rehabilitation power vehicle control system based on electrocardiogram signals according to claim 6, characterized in that: During the system identification process, the input and output data of the original data are first obtained with a sampling period of preset duration. Then, the MATLAB system identification toolbox is used for data processing and analysis. The specific process is as follows: the original data is detrended, the parameters of the identification model are obtained by using the least squares method, and the original data is input into the identification model to obtain the simulation output results. The actual output and simulation results are compared, and the effectiveness of the model is verified by two indicators: goodness of fit and root mean square error.

8. The cardiovascular rehabilitation power vehicle control system based on electrocardiogram signals according to claim 7, characterized in that: After completing system identification and obtaining the dynamic response model of the subject's heart rate to power input, the closed-loop heart rate control system operates in closed-loop mode based on the negative feedback mechanism. The heart rate feedback closed-loop control structure adopts the framework of a PI controller. Specifically, an electrocardiogram (ECG) acquisition device is used to receive the subject's actual heart rate data in real time, and this data is used as a feedback signal. This data is compared with a preset target heart rate, and the error between the two is used as the input to a heart rate controller. The controller outputs an updated target power P every preset time interval. target The command is sent to the power vehicle, thereby dynamically adjusting the output power P of the power vehicle. actual The subjects performed power cycling exercises, and their actual heart rate changed accordingly and was continuously monitored, thus forming a closed loop to achieve closed-loop tracking and control of heart rate. The PI controller effectively improves the dynamic response performance of the system and eliminates steady-state deviation by combining the proportional effect of the current error with the integral effect of the historical error. Let e(t) represent the heart rate tracking error at time t, as defined in Formula 6 below; where r(t) represents the target heart rate input and y(t) represents the actual heart rate feedback, both in bpm. The output u(t) of the PI controller represents the target power of the power vehicle at time t, in W, and is calculated as shown in Formula 7 below. Official 6; Official 7; Among them, K p K represents the proportional gain. i These two parameters represent the integral gain. The PI controller uses these two parameters to adjust the system output so that it tracks the set value.

9. A cardiovascular rehabilitation power vehicle control system based on electrocardiogram signals according to claim 8, characterized in that: Methods for setting the parameters of a heart rate controller also include: numerically simulating the heart rate controller by building a Simulink simulation system, within a preset parameter search range (K). p ∈[0.5, 2.0], K i ∈[0.01, 0.4]), traverse multiple groups (K) p ,K i The parameters were combined and their response performance to the target heart rate step change was tested. The heart rate step response and the corresponding target power change of different parameter groups were compared. With the constraints of step response overshoot <5bpm and settling time (±5bpm) <180s, the parameter group with the smallest steady-state error of heart rate step response and the smoothest fluctuation of target power output was selected as the parameters of the final heart rate controller. After the subject begins exercising on the power bike, the system calculates and dynamically updates the target power command in real time based on the aforementioned PI controller, thereby enabling the actual heart rate to track the target heart rate and forming a closed-loop control.

10. A cardiovascular rehabilitation power vehicle control system based on electrocardiogram signals according to claim 8, characterized in that: The cardiovascular rehabilitation power vehicle control system includes a Bluetooth communication module, a network transmission module, a data processing module, and a data display module; The Bluetooth communication module is responsible for two tasks: First, it pairs and connects the host computer with the ECG acquisition device based on the Bluetooth Low Energy protocol, ensuring that the web front-end can receive ECG data from the device in real time and parse the data according to a specific data packet format, thereby ensuring the accuracy and integrity of the data. Secondly, a virtual serial connection is established between the host computer and the power vehicle based on the Bluetooth serial port transparent transmission protocol, so as to realize the reliable issuance of control commands and the real-time transmission of the power vehicle's operating status, including power, speed and torque information. The network transmission module is used to establish a data transmission channel between the Web front-end and the Flask back-end. The front-end can upload the collected ECG data and heart rate data to the back-end for processing and analysis, and receive the results returned by the back-end. The data processing module is used by the backend server to perform quality assessment, anomaly detection, processing, and analysis on the received ECG data, and at the same time input the received heart rate data into the heart rate controller for calculation to obtain the control quantity; The data display module has the functions of real-time ECG waveform drawing, ECG indicator display, and power vehicle operation status display. It can also pop up a reminder to the user when an abnormal ECG is detected.

Images