Pain monitoring and rehabilitation equipment control method, device and equipment

Abstract

The invention discloses a pain monitoring and rehabilitation equipment control method, device and equipment and a readable storage medium, and relates to the technical field of brain-computer interaction. Comprising the following steps: acquiring a cerebral cortex blood oxygen signal of a target user, and preprocessing the cerebral cortex blood oxygen signal to obtain a preprocessed blood oxygen signal; extracting multi-dimensional features from the preprocessed blood oxygen signals based on a sliding time window; selecting features of which the contribution degree scores are greater than a preset score threshold value from the multi-dimensional features as a target feature set; inputting the target feature set into a pre-trained pain recognition model, and outputting a pain recognition result of the target user; inputting the target feature set into a pre-trained gait recognition model, and outputting a gait recognition result of the target user; and controlling rehabilitation equipment based on the pain recognition result and the gait recognition result. According to the method, secondary injury of the rehabilitation equipment to the user is avoided.

Description

Technical Field

[0001] This invention relates to the field of brain-computer interface technology, and more specifically to a method, apparatus, device, and readable storage medium for controlling pain monitoring and rehabilitation equipment. Background Technology

[0002] In recent years, brain-computer interface (BCI)-based intelligent motor rehabilitation technology has provided a new rehabilitation pathway for patients with motor dysfunction. By decoding brain signals to identify the user's movement intentions and driving rehabilitation devices such as exoskeletons, it is possible to help patients engage in active rehabilitation training. However, existing BCI rehabilitation technologies still have significant bottlenecks. Current rehabilitation devices generally lack effective monitoring and feedback mechanisms for the patient's physiological state during training. The training process is often mechanical and repetitive, ignoring potential adverse reactions such as pain and fatigue. This not only affects the rehabilitation effect but may also lead to secondary injuries such as muscle strains and joint wear due to delayed pain feedback, posing serious safety hazards.

[0003] To address these issues, some studies have attempted to utilize brain signals for pain recognition. For example, Fernandez et al. investigated a method for pain assessment using functional near-infrared spectroscopy (fNIRS) and a deep learning model, achieving high accuracy in distinguishing different levels of thermal stimulation pain. However, its drawbacks include being based on offline data analysis, which prevents real-time recognition and feedback of pain states, thus failing to meet the immediacy requirements of rehabilitation training; furthermore, the experimental paradigm is based on static thermal stimulation and has not validated its effectiveness in dynamic and complex walking tasks, where brain signals are susceptible to interference from motion artifacts.

[0004] For example, Khan et al. studied a method combining fNIRS and machine learning to classify painlessness, low pain, and high pain. However, its drawbacks are that the recognition accuracy is relatively limited and it is also an offline analysis mode. More importantly, this study, as well as other similar studies, only stay at the level of pain state recognition and do not propose how to effectively link the recognition results with the motion control of rehabilitation equipment to achieve adaptive adjustment of motion parameters.

[0005] Current research primarily relies on offline analysis, resulting in significant time delays and failing to meet the real-time monitoring and immediate intervention needs for pain in clinical rehabilitation. Furthermore, most experiments are conducted in static environments, raising questions about the effectiveness and robustness of these models in real-world dynamic walking tasks. Existing technological approaches are limited to pain identification and fail to provide a complete technical solution for real-time, closed-loop, and adaptive adjustment of pain monitoring results and rehabilitation device motion parameters.

[0006] Therefore, there is an urgent need for a pain monitoring and rehabilitation device control method that can overcome the above-mentioned shortcomings. Summary of the Invention

[0007] The purpose of this invention is to provide a method, apparatus, device, and readable storage medium for controlling pain monitoring and rehabilitation equipment. By acquiring the cerebral cortex blood oxygenation signal of the target user, the method can determine the pain recognition result and gait recognition result of the target user based on the cerebral cortex blood oxygenation signal, and then control the rehabilitation equipment based on the pain recognition result and gait recognition result. This avoids ignoring the user's pain state when controlling the rehabilitation equipment, and thus avoids the rehabilitation equipment causing secondary harm to the user.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] In a first aspect, the present invention provides a method for controlling a pain monitoring and rehabilitation device, the method comprising:

[0010] Obtain the blood oxygenation signal of the target user's cerebral cortex, and preprocess the blood oxygenation signal of the cerebral cortex to obtain the preprocessed blood oxygenation signal;

[0011] Multidimensional features are extracted from the preprocessed blood oxygen signal based on a sliding time window; the multidimensional features include time domain features, frequency domain features, and spatial features.

[0012] From the multidimensional features, features with contribution scores greater than a preset score threshold are selected as the target feature set;

[0013] The target feature set is input into a pre-trained pain recognition model, which outputs the pain recognition result of the target user.

[0014] The target feature set is input into a pre-trained gait recognition model, and the gait recognition result of the target user is output.

[0015] The rehabilitation equipment is controlled based on the pain recognition results and gait recognition results.

[0016] In some embodiments, the cerebral cortex blood oxygenation signal is preprocessed to obtain a preprocessed blood oxygenation signal, including:

[0017] The abnormal values ​​in the blood oxygenation signal of the cerebral cortex were eliminated by using a Hampshire filter to obtain the initial blood oxygenation signal;

[0018] The initial blood oxygen signal is decomposed into different physiological frequency bands by a second-order Chebyshev bandpass filter to obtain the preprocessed blood oxygen signal; the different physiological frequency bands include the endothelial cell metabolic activity frequency band and the neural activity frequency band.

[0019] In some embodiments, selecting features from the multidimensional features whose contribution scores are greater than a preset score threshold as the target feature set includes:

[0020] The Boruta algorithm is used to filter candidate features from the multidimensional features;

[0021] Calculate the mutual information value of each candidate feature and determine the importance score of each candidate feature based on the random forest model;

[0022] The mutual information value and the importance score are weighted and summed to obtain the contribution score of each candidate feature;

[0023] Candidate features whose contribution scores are greater than a preset score threshold are selected from each of the importance scores as the target feature set.

[0024] In some embodiments, controlling the rehabilitation device based on the pain recognition results and gait recognition results includes:

[0025] If the pain recognition result is a pain-free state, then the rehabilitation device is controlled based on the gait recognition result;

[0026] If the pain recognition result indicates a painful state, the rehabilitation device is controlled based on the pain recognition result and the attenuation coefficient.

[0027] In some embodiments, the gait recognition result includes initial stride length and initial cadence, and controlling the rehabilitation device based on the gait recognition result includes:

[0028] Control commands are generated based on the initial stride length and initial step frequency;

[0029] The rehabilitation equipment is controlled to move according to the initial stride length and initial stride frequency based on the control commands.

[0030] In some embodiments, controlling the rehabilitation device based on pain recognition results and attenuation coefficients includes:

[0031] The target stride and target step frequency are calculated by multiplying the initial stride length and initial step frequency by the attenuation coefficient, respectively.

[0032] Control commands are generated based on the target stride length and target step frequency;

[0033] The control commands control the rehabilitation equipment to move according to the target stride length and target cadence.

[0034] In a second aspect, the present invention also provides a pain monitoring and rehabilitation equipment control device, the device comprising:

[0035] The signal acquisition module is used to acquire the blood oxygenation signal of the cerebral cortex of the target user and preprocess the blood oxygenation signal of the cerebral cortex to obtain a preprocessed blood oxygenation signal.

[0036] The feature extraction module is used to extract multidimensional features from the preprocessed blood oxygen signal based on a sliding time window; the multidimensional features include time domain features, frequency domain features, and spatial features;

[0037] The feature selection module is used to select features from the multidimensional features whose contribution scores are greater than a preset score threshold as the target feature set.

[0038] The first recognition module is used to input the target feature set into a pre-trained pain recognition model and output the pain recognition result of the target user.

[0039] The second recognition module is used to input the target feature set into a pre-trained gait recognition model and output the gait recognition result of the target user.

[0040] The device control module is used to control the rehabilitation device based on the pain recognition results and gait recognition results.

[0041] Thirdly, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the pain monitoring and rehabilitation device control method provided in the first aspect.

[0042] Fourthly, the present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the pain monitoring and rehabilitation device control method provided in the first aspect.

[0043] Fifthly, the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the pain monitoring and rehabilitation device control method provided in the first aspect.

[0044] The beneficial effects of this invention are as follows: The pain monitoring and rehabilitation equipment control method of this application first acquires the cortical blood oxygen signal of the target user and preprocesses the cortical blood oxygen signal to obtain a preprocessed blood oxygen signal; then, based on a sliding time window, it extracts multidimensional features from the preprocessed blood oxygen signal; the multidimensional features include time-domain features, frequency-domain features, and spatial features; then, it selects features with contribution scores greater than a preset score threshold from the multidimensional features as a target feature set; then, it inputs the target feature set into a pre-trained pain recognition model and outputs the pain recognition result of the target user; simultaneously, it inputs the target feature set into a pre-trained gait recognition model and outputs the gait recognition result of the target user; finally, it controls the rehabilitation equipment based on the pain recognition result and the gait recognition result. By acquiring the cortical blood oxygen signal of the target user, it is possible to determine the pain recognition result and gait recognition result of the target user based on the cortical blood oxygen signal, and then control the rehabilitation equipment based on the pain recognition result and gait recognition result, thus avoiding ignoring the user's pain state when controlling the rehabilitation equipment, and thus avoiding secondary harm to the user caused by the rehabilitation equipment.

[0045] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, the preferred embodiments of the present invention are described in detail below with reference to the accompanying drawings. Attached Figure Description

[0046] Figure 1 This is a flowchart illustrating a pain monitoring and rehabilitation device control method according to an embodiment of the present invention;

[0047] Figure 2 This is a schematic diagram of the experimental results of a pain recognition model according to an embodiment of the present invention;

[0048] Figure 3 This is a schematic diagram of the predicted probability results of a pain recognition model according to an embodiment of the present invention;

[0049] Figure 4 This is a schematic diagram of stride prediction results of a gait recognition model according to an embodiment of the present invention;

[0050] Figure 5 This is a schematic diagram of the gait frequency prediction result of a gait recognition model according to an embodiment of the present invention;

[0051] Figure 6 This is a schematic diagram of a gait curve according to an embodiment of the present invention;

[0052] Figure 7 This is a flowchart illustrating another method for controlling a pain monitoring and rehabilitation device according to an embodiment of the present invention;

[0053] Figure 8 This is a schematic diagram of the structure of a pain monitoring and rehabilitation equipment control device according to an embodiment of the present invention;

[0054] Figure 9 This is a schematic diagram of an electronic device structure provided in an embodiment of this application. Detailed Implementation

[0055] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. 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.

[0056] It should be noted that references to "an embodiment," "embodiment," "example embodiment," etc., in this specification refer to the described embodiment including specific features, structures, or characteristics; however, not every embodiment must include these specific features, structures, or characteristics. Furthermore, such expressions do not refer to the same embodiment. Moreover, when describing specific features, structures, or characteristics in conjunction with embodiments, whether or not explicitly described, it is indicated that incorporating such features, structures, or characteristics into other embodiments is within the knowledge of those skilled in the art.

[0057] Furthermore, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0058] In some embodiments, such as Figure 1 As shown, a method for controlling a pain monitoring and rehabilitation device is provided, including:

[0059] S101, acquire the blood oxygenation signal of the target user's cerebral cortex, and preprocess the blood oxygenation signal of the cerebral cortex to obtain the preprocessed blood oxygenation signal.

[0060] Specifically, the blood oxygenation signal of the target user's cerebral cortex can be collected using a functional near-infrared spectroscopy (fNIRS) system (the collection area is selected from the target user's prefrontal cortex, supplementary motor area, premotor area, primary motor cortex, and primary somatosensory cortex). Since the collected cerebral cortex blood oxygenation signal may contain other noise signals, it is necessary to preprocess the cerebral cortex blood oxygenation signal to remove noise from the cerebral cortex signal and obtain a preprocessed blood oxygenation signal.

[0061] Specifically, the preprocessing of the cerebral cortex blood oxygen signal includes: using a Hanpour filter to eliminate abnormal values ​​in the cerebral cortex blood oxygen signal to obtain an initial blood oxygen signal; and using a second-order Chebyshev bandpass filter to decompose the initial blood oxygen signal into different physiological frequency bands to obtain a preprocessed blood oxygen signal; the different physiological frequency bands include the endothelial cell metabolic activity frequency band and the neural activity frequency band.

[0062] Because spikes or abnormal data points may occur during the target user's walking process due to momentary head shaking or poor contact between the probe and the skin, a Hanpur filter is needed to eliminate these abnormal values ​​to obtain the initial blood oxygen signal. Then, a second-order Chebyshev bandpass filter is used to decompose the initial blood oxygen signal into different physiological frequency bands, thereby separating signal components from different physiological sources, thus obtaining the preprocessed blood oxygen signal.

[0063] S102 extracts multidimensional features from preprocessed blood oxygen signals based on a sliding time window.

[0064] Among them, multidimensional features include time-domain features, frequency-domain features, and spatial features.

[0065] Specifically, the blood oxygenation signal collected during the cerebral cortex is a continuous signal. Therefore, a sliding time window can be used to extract features from the preprocessed blood oxygenation signal. These features include time domain features, frequency domain features, and spatial features. Extracting multiple feature types can increase the accuracy of subsequent pain recognition and gait recognition results.

[0066] S103: Select features from the multidimensional features whose contribution scores are greater than the preset score threshold as the target feature set.

[0067] Specifically, among the extracted multidimensional features, there may be some invalid features, that is, features that contribute little to pain recognition and gait recognition. Therefore, it is necessary to calculate the contribution score of each feature in the multidimensional features and set a preset score threshold. Features with contribution scores lower than the preset score threshold are removed, and the remaining features are retained as the target feature set, thereby reducing the computational load of the pain recognition model and gait recognition model and increasing recognition efficiency.

[0068] Specifically, the method for selecting features with contribution scores greater than a preset score threshold from multidimensional features as the target feature set can be as follows: use the Boruta algorithm to filter candidate features from multidimensional features; calculate the mutual information value of each candidate feature and determine the importance score of each candidate feature based on the random forest model; perform a weighted summation of the mutual information value and the importance score to obtain the contribution score of each candidate feature; and select candidate features with contribution scores greater than a preset score threshold from each importance score as the target feature set.

[0069] The Boruta algorithm is a feature selection algorithm that aims to initially screen multidimensional features, selecting features with higher contributions as candidate features. It then calculates the mutual information value of each candidate feature and determines the importance score of each candidate feature based on a random forest model. The mutual information value and importance score are weighted and summed to obtain the contribution score of each candidate feature. Based on the contribution score, the candidate features are screened a second time, selecting candidate features with contribution scores greater than a preset score threshold as the target feature set. This reduces the computational load of the model and thus increases the model's recognition efficiency.

[0070] S104: Input the target feature set into the pre-trained pain recognition model and output the pain recognition result of the target user.

[0071] Specifically, by inputting the target feature set into a pre-trained pain recognition model, the pain recognition model can output the pain recognition results of the target user, including the pain state and the non-pain state.

[0072] The pain recognition model is pre-trained. During the training phase, the user's sample feature set (obtained in the same way as the target feature set) is used as input, and real pain feedback is used as a reference to train the pain recognition model until the accuracy of the pain recognition model reaches the preset requirements.

[0073] S105 inputs the target feature set into the pre-trained gait recognition model and outputs the gait recognition result of the target user.

[0074] Specifically, the target feature set is input into a pre-trained gait recognition model, which can then output the gait recognition results of the target user, including the initial stride length and initial stride frequency.

[0075] The gait recognition model is also pre-trained. During the training phase, the user's sample feature set (obtained using the same method as the target feature set) is used as input, and the user's stepped gait parameters during walking are used as a reference to train the gait recognition model until its accuracy reaches the preset requirements. It should be noted that the process of obtaining the user's stepped gait parameters during walking is as follows: first, the user's gait parameters during walking are collected using an inertial sensor; then, these gait parameters are processed into stepped parameters to obtain the user's stepped gait parameters during walking.

[0076] S106 controls rehabilitation equipment based on pain recognition and gait recognition results.

[0077] Specifically, control commands can be generated based on pain recognition results and gait recognition results, and then the rehabilitation equipment can be controlled according to the control commands to help the target user move.

[0078] Optionally, if the pain recognition result is a pain-free state, the rehabilitation device can be controlled based on the gait recognition result.

[0079] Specifically, if the pain recognition result is a pain-free state, control commands are generated based on the initial stride length and initial cadence; based on the control commands, the rehabilitation equipment is controlled to move according to the initial stride length and initial cadence.

[0080] When the recognition result is a pain-free state, it means that the target user is in good condition and can move according to the initial stride and initial cadence identified by the gait recognition model. At this time, the rehabilitation equipment is controlled to assist the target user in moving according to the initial stride and initial cadence.

[0081] Optionally, if the pain recognition result is a pain state, the rehabilitation device can be controlled based on the pain recognition result and the attenuation coefficient.

[0082] Specifically, if the pain recognition result is a pain state, the initial stride length and initial cadence are multiplied by the attenuation coefficient to calculate the target stride length and target cadence; control commands are generated based on the target stride length and target cadence; and the rehabilitation equipment is controlled to move according to the target stride length and target cadence based on the control commands.

[0083] When the recognition result indicates a state of pain, it means that the target user is not in a good condition. If the user moves according to the initial stride length and initial cadence identified by the gait recognition model, it may cause secondary injury to the target user due to moving too fast. In this case, the user or instructor should determine the attenuation coefficient (e.g., 0.7), and then multiply the attenuation coefficient by the initial stride length and initial cadence to obtain the target stride length and target cadence. This will reduce the target user's stride length and cadence, alleviate the target user's pain, and avoid causing secondary injury to the target user.

[0084] The pain monitoring and rehabilitation equipment control method in the above embodiments first acquires the cortical blood oxygenation signal of the target user and preprocesses it to obtain a preprocessed blood oxygenation signal. Then, based on a sliding time window, multidimensional features are extracted from the preprocessed blood oxygenation signal. These multidimensional features include time-domain features, frequency-domain features, and spatial features. Next, features with contribution scores greater than a preset score threshold are selected from the multidimensional features as the target feature set. The target feature set is then input into a pre-trained pain recognition model, which outputs the target user's pain recognition result. Simultaneously, the target feature set is input into a pre-trained gait recognition model, which outputs the target user's gait recognition result. Finally, the rehabilitation equipment is controlled based on the pain recognition result and the gait recognition result. By acquiring the cortical blood oxygenation signal of the target user, the method can determine the target user's pain recognition result and gait recognition result based on the cortical blood oxygenation signal, and then control the rehabilitation equipment based on these results. This avoids ignoring the user's pain state when controlling the rehabilitation equipment, thus preventing secondary harm to the user from the rehabilitation equipment.

[0085] In another embodiment, experiments were conducted on multiple users based on the pain recognition model of this application, and the experimental results are as follows: Figure 2 As shown, the results indicate that the average accuracy of the test set is 85.6%, precision is 86.5%, recall is 84.2%, F1 score is 85.3%, and average decision lag is 1.1 seconds. The prediction probability results for one of the test sets are shown below. Figure 3 The model identifies a pain state when the predicted probability value is greater than 0.51. The results from this sample demonstrate that the model can accurately and promptly identify pain states. This data proves that the method of this invention can accurately and quickly identify pain states in dynamic walking tasks, achieving performance levels suitable for real-time applications and solving the problem that existing technologies are mostly offline and static analyses.

[0086] In another embodiment, the gait recognition model in this application was validated, and the validation results are referenced. Figure 4 and Figure 5 , Figure 4 For stride prediction results, Figure 5 The results show the stride frequency prediction, with the mean absolute percentage errors of the stride length parameter fitting and the stride frequency parameter fitting being 3.73% and 3.94%, respectively.

[0087] In another embodiment, a gait curve is also determined based on stride length, cadence, and pain status, with reference to... Figure 6The curves represent the normal output curve when there is no pain and the output curve after adding the attenuation coefficient when a pain state is detected, as well as the comparison with the original gait curve. This shows that when a pain state is detected, the method in this application can smoothly reduce the amplitude and speed of movement according to the control command, which is consistent with the expectation, demonstrating the effectiveness and feasibility of the method in this application.

[0088] To more comprehensively demonstrate this solution, this embodiment presents an optional approach to pain monitoring and rehabilitation equipment control, such as... Figure 7 As shown:

[0089] S201, acquire the target user's cerebral cortex blood oxygenation signal.

[0090] S202 uses a Hample filter to eliminate abnormal values ​​in the blood oxygenation signal of the cerebral cortex to obtain the initial blood oxygenation signal.

[0091] S203 decomposes the initial blood oxygen signal into different physiological frequency bands through a second-order Chebyshev bandpass filter to obtain a preprocessed blood oxygen signal.

[0092] Among them, different physiological frequency bands include the frequency band of endothelial cell metabolic activity and the frequency band of neural activity.

[0093] S204 extracts multidimensional features from preprocessed blood oxygen signals based on a sliding time window.

[0094] Among them, multidimensional features include time-domain features, frequency-domain features, and spatial features.

[0095] S205 uses the Boruta algorithm to select candidate features from multidimensional features.

[0096] S206, calculate the mutual information value of each candidate feature, and determine the importance score of each candidate feature based on the random forest model.

[0097] S207, the mutual information value and importance score are weighted and summed to obtain the contribution score of each candidate feature.

[0098] S208: Select candidate features whose contribution scores are greater than a preset score threshold from each importance score as the target feature set.

[0099] S209: Input the target feature set into the pre-trained pain recognition model and output the pain recognition result of the target user.

[0100] S210 inputs the target feature set into the pre-trained gait recognition model and outputs the gait recognition result of the target user.

[0101] S211, if the pain recognition result is no pain state, then generate control instructions based on the initial stride length and initial cadence.

[0102] S212, based on control commands, controls the rehabilitation equipment to move according to the initial stride length and initial stride frequency.

[0103] S213, if the pain recognition result is a pain state, then the initial stride length and initial cadence are multiplied by the attenuation coefficient to calculate the target stride length and target cadence.

[0104] S214 generates control commands based on target stride length and target stride frequency.

[0105] S215 controls the rehabilitation equipment to move according to the target stride length and target stride frequency based on control commands.

[0106] The specific processes of S201-S215 described above can be found in the description of the above method embodiments. Their implementation principles and technical effects are similar, and will not be repeated here.

[0107] Based on the same inventive concept, this application also provides a pain monitoring and rehabilitation equipment control device for implementing the pain monitoring and rehabilitation equipment control method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations of one or more embodiments of the pain monitoring and rehabilitation equipment control device provided below can be found in the limitations of the pain monitoring and rehabilitation equipment control method described above, and will not be repeated here.

[0108] In one embodiment, such as Figure 8 As shown, a pain monitoring and rehabilitation equipment control device is provided, the device comprising:

[0109] The signal acquisition module 30 is used to acquire the blood oxygenation signal of the cerebral cortex of the target user and preprocess the blood oxygenation signal of the cerebral cortex to obtain a preprocessed blood oxygenation signal.

[0110] Feature extraction module 31 is used to extract multidimensional features from the preprocessed blood oxygen signal based on a sliding time window; the multidimensional features include time domain features, frequency domain features and spatial features;

[0111] Feature selection module 32 is used to select features with contribution scores greater than a preset score threshold from the multidimensional features as the target feature set;

[0112] The first recognition module 33 is used to input the target feature set into a pre-trained pain recognition model and output the pain recognition result of the target user.

[0113] The second recognition module 34 is used to input the target feature set into a pre-trained gait recognition model and output the gait recognition result of the target user.

[0114] The device control module 35 is used to control the rehabilitation device based on the pain recognition results and gait recognition results.

[0115] In another embodiment, the above Figure 8 The signal acquisition module 30 is specifically used to: eliminate abnormal values ​​in the blood oxygen signal of the cerebral cortex using a Hanpu filter to obtain an initial blood oxygen signal; decompose the initial blood oxygen signal into different physiological frequency bands using a second-order Chebyshev bandpass filter to obtain the preprocessed blood oxygen signal; the different physiological frequency bands include the endothelial cell metabolic activity frequency band and the neural activity frequency band.

[0116] In another embodiment, the above Figure 8 The feature selection module 32 is specifically used for: using the Boruta algorithm to filter candidate features from the multidimensional features; calculating the mutual information value of each candidate feature and determining the importance score of each candidate feature based on the random forest model; performing a weighted summation of the mutual information value and the importance score to obtain the contribution score of each candidate feature; and selecting candidate features with a contribution score greater than a preset score threshold from each importance score as the target feature set.

[0117] In another embodiment, the above Figure 8 The device control module 35 includes:

[0118] A first control unit is configured to control the rehabilitation device based on the gait recognition result if the pain recognition result indicates a pain-free state.

[0119] The second control unit is used to control the rehabilitation device based on the pain recognition result and the attenuation coefficient if the pain recognition result indicates a pain state.

[0120] In another embodiment, the gait recognition result includes an initial stride length and an initial cadence. The first control unit in the above embodiment is specifically used to: generate control commands based on the initial stride length and initial cadence; and control the rehabilitation device to move according to the initial stride length and initial cadence based on the control commands.

[0121] In another embodiment, the second control unit in the above embodiment is specifically used to: multiply the initial stride length and initial cadence by the attenuation coefficient to obtain the target stride length and target cadence; generate control commands based on the target stride length and target cadence; and control the rehabilitation device to move according to the target stride length and target cadence based on the control commands.

[0122] This application also provides an electronic device, in some embodiments, referring to... Figure 9As shown, the electronic device 700 includes an input unit 710, a memory 720, a processor 730, and an output unit 740. The memory 720 stores program instructions that can be executed on the processor 730. The processor 730 can execute the pain monitoring and rehabilitation device control method and / or technical solution based on the foregoing embodiments by calling the program instructions. The electronic device 700 can be a mobile terminal device such as a mobile phone or computer.

[0123] Furthermore, embodiments of this application also provide a computer-readable storage medium for storing a computer program that performs a method for controlling a pain monitoring and rehabilitation device. For example, computer program instructions, when executed by a computer, can invoke or provide the methods and / or technical solutions according to this application through the operation of the computer. The program instructions that invoke the methods of this application may be stored in a fixed or removable storage medium, and / or transmitted via data streams in broadcast or other signal carrying media, and / or stored in a storage medium that operates according to the program instructions.

[0124] Obviously, those skilled in the art should understand that the modules or steps of this application described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby storing them in a storage device for execution by a computing device, or fabricating them separately as individual integrated circuit modules, or fabricating multiple modules or steps as a single integrated circuit module. Thus, this application is not limited to any particular combination of hardware and software.

[0125] The technical features of the above embodiments can be arbitrarily integrated. For the sake of brevity, not all possible integrations of the technical features in the above embodiments are described. However, as long as the integration of these technical features does not contradict each other, they should be considered to be within the scope of this specification.

[0126] The above embodiments merely illustrate several implementation methods of the present invention, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A method for controlling a pain monitoring and rehabilitation device, characterized in that, The method includes: Obtain the blood oxygenation signal of the target user's cerebral cortex, and preprocess the blood oxygenation signal of the cerebral cortex to obtain the preprocessed blood oxygenation signal; Multidimensional features are extracted from the preprocessed blood oxygen signal based on a sliding time window; the multidimensional features include time domain features, frequency domain features, and spatial features. From the multidimensional features, features with contribution scores greater than a preset score threshold are selected as the target feature set; The target feature set is input into a pre-trained pain recognition model, which outputs the pain recognition result of the target user. The target feature set is input into a pre-trained gait recognition model, and the gait recognition result of the target user is output. The rehabilitation equipment is controlled based on the pain recognition results and gait recognition results.

2. The pain monitoring and rehabilitation device control method as described in claim 1, characterized in that, The cerebral cortex blood oxygenation signal is preprocessed to obtain a preprocessed blood oxygenation signal, including: The abnormal values ​​in the blood oxygenation signal of the cerebral cortex were eliminated by using a Hampshire filter to obtain the initial blood oxygenation signal; The initial blood oxygen signal is decomposed into different physiological frequency bands by a second-order Chebyshev bandpass filter to obtain the preprocessed blood oxygen signal; the different physiological frequency bands include the endothelial cell metabolic activity frequency band and the neural activity frequency band.

3. The pain monitoring and rehabilitation device control method as described in claim 1, characterized in that, Features with contribution scores greater than a preset score threshold are selected from the multidimensional features as the target feature set, including: The Boruta algorithm is used to filter candidate features from the multidimensional features; Calculate the mutual information value of each candidate feature and determine the importance score of each candidate feature based on the random forest model; The mutual information value and the importance score are weighted and summed to obtain the contribution score of each candidate feature; Candidate features whose contribution scores are greater than a preset score threshold are selected from each of the importance scores as the target feature set.

4. The pain monitoring and rehabilitation device control method as described in claim 1, characterized in that, Controlling the rehabilitation device based on the pain recognition results and gait recognition results includes: If the pain recognition result is a pain-free state, then the rehabilitation device is controlled based on the gait recognition result; If the pain recognition result indicates a painful state, the rehabilitation device is controlled based on the pain recognition result and the attenuation coefficient.

5. The pain monitoring and rehabilitation device control method as described in claim 4, characterized in that, The gait recognition result includes initial stride length and initial gait frequency. Controlling the rehabilitation device based on the gait recognition result includes: Control commands are generated based on the initial stride length and initial step frequency; The rehabilitation equipment is controlled to move according to the initial stride length and initial stride frequency based on the control commands.

6. The pain monitoring and rehabilitation device control method as described in claim 5, characterized in that, The rehabilitation device is controlled based on pain recognition results and attenuation coefficients, including: The target stride and target step frequency are calculated by multiplying the initial stride length and initial step frequency by the attenuation coefficient, respectively. Control commands are generated based on the target stride length and target step frequency; The control commands control the rehabilitation equipment to move according to the target stride length and target cadence.

7. A control device for pain monitoring and rehabilitation equipment, characterized in that, The device includes: The signal acquisition module is used to acquire the blood oxygenation signal of the cerebral cortex of the target user and preprocess the blood oxygenation signal of the cerebral cortex to obtain a preprocessed blood oxygenation signal. The feature extraction module is used to extract multidimensional features from the preprocessed blood oxygen signal based on a sliding time window; the multidimensional features include time domain features, frequency domain features, and spatial features; The feature selection module is used to select features from the multidimensional features whose contribution scores are greater than a preset score threshold as the target feature set. The first recognition module is used to input the target feature set into a pre-trained pain recognition model and output the pain recognition result of the target user. The second recognition module is used to input the target feature set into a pre-trained gait recognition model and output the gait recognition result of the target user. The device control module is used to control the rehabilitation device based on the pain recognition results and gait recognition results.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the pain monitoring and rehabilitation device control method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the pain monitoring and rehabilitation device control method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the pain monitoring and rehabilitation device control method as described in any one of claims 1 to 6.

Images