A Multi-Sensor Wearable Tremor Detection System Based on an Adaptive Model

Abstract

A multi-sensor wearable tremor detection system based on an adaptive model includes a tremor detection module, an adaptive model diagnosis module, and a model database. The tremor detection module includes an electromyographic signal sensor and an acceleration signal sensor for collecting forearm muscle electromyographic signals and hand acceleration signals. The adaptive model diagnosis module includes forearm muscle-bone coupling tremor dynamics model construction, multi-source tremor parameter theoretical calculation, multi-sensor data combination, model adaptive update, and disease diagnosis steps. It constructs a forearm muscle-bone coupling tremor dynamics model with adaptive capabilities and diagnoses data collected by the tremor detection module. The model database provides basic data. The present invention combines the forearm muscle-bone coupling tremor dynamics model with adaptive update capabilities with multi-sensor data, utilizes multiple sensors to enrich the data, and simultaneously links human physiological signals with the tremor state of the upper limb, thereby improving the rationality and interpretability of the diagnostic results.

Description

Technical Field

[0001] The present invention belongs to the technical field of medical detection equipment, and in particular relates to a multi-sensor wearable tremor detection system based on an adaptive model. Background Art

[0002] Parkinson's disease and essential tremor are both common diseases that can cause involuntary tremors in the human body. The tremor forms of the two diseases are different. Parkinson's disease causes resting tremor, which is a 4-6Hz tremor when the patient is at rest and not moving. In addition to resting tremor, Parkinson's patients also experience common symptoms such as slow movement, muscle rigidity, and difficulty maintaining posture. Essential tremor has two symptoms: postural tremor and action tremor. The tremor frequency at the time of onset is 8-12Hz. Postural tremor is mainly manifested as obvious tremor when the hand maintains a certain posture; action tremor is the most characteristic manifestation of essential tremor, mainly manifested as obvious tremor when the hand completes a certain movement. As the disease worsens, the patient's tremor amplitude will become larger and larger, causing serious impact on the patient's normal work and life.

[0003] Early diagnosis and treatment of tremor disorders can effectively prevent the progression of tremor diseases. Currently, the mainstream diagnostic method for tremor disorders in hospitals relies on large-scale medical equipment combined with the doctor's subjective judgment of the patient's tremor severity to determine the patient's condition. This method's diagnostic accuracy relies heavily on the doctor's experience, which is easily influenced by subjective factors and lacks objectivity. Using wearable tremor detection devices to diagnose tremor disorders can effectively solve this problem.

[0004] Wearable tremor detection devices are devices used to detect tremor disorders. With the development of sensors and information technology, a variety of sensors have been incorporated into tremor detection devices, diversifying the types of data they can capture. Data-driven diagnostic methods (application number: 201911375827.X, title: "Parkinson's Resting-State Tremor Assessment Method Based on Wearable Somatosensory Network"; application number: 202110891734.3, title: "A Modeling Method for Tremor Detection, Hand Tremor Detection Device, and Method") do not conduct in-depth analysis of the mechanisms and patterns of human degeneration. Instead, they use machine learning algorithms to process data to diagnose tremor disorders. However, data-driven methods rely too much on training data, lack necessary empirical guidance, and have poor interpretability. Summary of the Invention

[0005] To overcome the shortcomings of the above-mentioned prior art, the present invention provides a multi-sensor wearable tremor detection system based on an adaptive model. Starting from the perspective of human body mechanisms, this system describes the degenerative behavior of the human body using mathematical function expressions. By establishing simulation models of human body parts, relatively reliable and interpretable results are obtained. By combining a forearm muscle-bone coupling tremor dynamics model with adaptive updating capabilities with multi-sensor data, the data collected by multiple sensors is fully utilized to guide the diagnosis of tremor diseases, achieving high-accuracy diagnosis of tremor diseases.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] A multi-sensor wearable tremor detection system based on an adaptive model, comprising a tremor detection module, an adaptive model diagnosis module, and a model database;

[0008] The tremor detection module includes an electromyographic signal sensor and an acceleration signal sensor, which respectively collect the electromyographic signal of the muscles that control the movement of the forearm and the acceleration signal of the hand;

[0009] The adaptive model diagnosis module includes five steps: forearm muscle-bone coupling tremor dynamics model construction, multi-source tremor parameter theoretical calculation, multi-sensor data combination, model adaptive update, and disease diagnosis; the forearm muscle-bone coupling tremor dynamics model construction step constructs a forearm muscle-bone coupling tremor dynamics model based on the common space of multi-source tremor characteristics; the multi-source tremor parameter theoretical calculation step calculates the unknown parameters of the forearm muscle-bone coupling tremor dynamics model based on the sample observation data of tremor patients in the model database; the multi-sensor data combination step selects the optimal data from the multi-sensor data collected by the tremor detection module based on the sensitivity of the multi-sensor data of tremor patients in the model database to the health status, and sends the combined data to the forearm muscle-bone coupling tremor dynamics model; the model adaptive update step updates the parameters and status of the forearm muscle-bone coupling tremor dynamics model based on the data obtained in the multi-sensor data combination step; the disease diagnosis step uses the updated forearm muscle-bone coupling tremor dynamics model to diagnose the subject, obtain the diagnosis result and verify it;

[0010] The model database provides basic data for the construction of forearm muscle-bone coupling tremor dynamics model, theoretical calculation of multi-source tremor parameters, and multi-sensor data combination.

[0011] The tremor detection module adopts a wearable design and is worn on the forearm of the subject. Wireless transmission is used to transmit multi-sensor signals from the tremor detection module to the host computer.

[0012] The forearm muscle-bone coupled tremor dynamics model construction link describes the relationship between the tremor feature degradation function and the degradation process of the subject's upper limb tremor symptoms, and the relationship between the tremor symptom multi-source characterization function and multi-sensor data by constructing a multi-source tremor feature common space model; the tremor feature degradation function is combined with the tremor symptom multi-source characterization function to form a forearm muscle-bone coupled tremor dynamics model based on the multi-source tremor feature common space.

[0013] The multi-source tremor parameter theoretical calculation link calculates unknown parameters such as tremor characteristic parameters and forearm muscle-bone coupling tremor dynamics parameters of the forearm muscle-bone coupling tremor dynamics model based on the sample observation data of tremor subjects in the model database, and inputs the parameters into the forearm muscle-bone coupling tremor dynamics model.

[0014] The multi-sensor data combination link selects the optimal data from the real-time multi-sensor data of the tremor patients collected by the tremor detection module based on the sensitivity of the multi-sensor data of the tremor patients in the model database to the health status, and combines them to eliminate signal redundancy between different sensor data and improve the system's computing efficiency. The combined data is then sent to the forearm muscle-bone coupling tremor dynamics model.

[0015] In the model adaptive updating link, the constructed forearm muscle-bone coupling tremor dynamics model is adaptively updated using the optimal real-time multi-sensor data combination selected in the multi-sensor data combination link, so that the updated model has stronger pertinence to the subject.

[0016] In the disease diagnosis process, the updated forearm muscle-bone coupling tremor dynamics model is used to diagnose tremor diseases in the subject, obtain the diagnosis result and verify it; if the diagnosis result is correct, it is output; if the diagnosis result is incorrect, the forearm muscle-bone coupling tremor dynamics model construction process is feedback optimized, and the optimized forearm muscle-bone coupling tremor dynamics model is used to re-diagnose the subject's multi-sensor data.

[0017] The model database provides basic data for constructing a forearm muscle-bone coupling tremor dynamics model. It is composed of multi-sensor data from multiple groups of subjects whose tremor disease status has been determined by experts, and the model database is dynamically updated by incorporating available data collected during the diagnosis process.

[0018] Compared with the prior art, the present invention has the following beneficial technical effects:

[0019] The present invention uses a diagnostic approach that combines a forearm muscle-bone coupling tremor dynamics model with adaptive updating capabilities with multi-sensor data to diagnose tremor disorders. This allows the forearm muscle-bone coupling tremor dynamics model to model and analyze data collected by multiple sensors, effectively utilizing the rich data collected by multiple sensors. The forearm muscle-bone coupling tremor dynamics model, based on the shared space of multi-source tremor features, follows human physiological laws, links human physiological signals with upper limb tremor status, and improves the rationality and interpretability of diagnostic results. The multi-sensor data combination process combines real-time sensor data based on their varying sensitivity to the tremor patient's health status, reducing data redundancy and improving system computational efficiency. The model adaptive update process updates the forearm muscle-bone coupling tremor dynamics model based on the combined real-time sensor data, adjusting model parameters for subjects in different states, thereby achieving high specificity for each condition. The disease diagnosis process uses the updated forearm muscle-bone coupling tremor dynamics model to diagnose tremor disorders in subjects and verifies the diagnostic results, ensuring their reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 This is a block diagram of the system composition of the present invention.

[0021] Figure 2 This is the operating block diagram of the adaptive model diagnosis module of the present invention. DETAILED DESCRIPTION

[0022] The present invention is described in detail below with reference to examples and drawings.

[0023] Reference Figure 1 ,A multi-sensor wearable tremor detection system based on an adaptive model, including a tremor detection module, an adaptive model diagnosis module, and a model database;

[0024] The tremor detection module includes an electromyographic signal sensor and an acceleration signal sensor, which respectively collect the electromyographic signal of the forearm's main muscles that control movement and the hand's acceleration signal. The collected data is sent to the digital-analog linkage diagnosis module via wireless transmission. The signal acquisition process is as follows: the electromyographic signal sensor and the acceleration sensor are respectively fixed to the corresponding positions of the subject's upper limb. After the subject's state stabilizes, the subject is instructed to perform a specified movement, and the subject's electromyographic signal and acceleration signal are sampled during the movement.

[0025] The adaptive model diagnosis module uses the constructed forearm muscle-bone coupling tremor dynamics model to diagnose the data to be diagnosed collected by the tremor detection module and outputs a diagnostic result. To ensure the reliability of the output diagnostic result, the correctness of the diagnosis is judged. If the diagnostic result is incorrect, feedback optimization is performed on the construction process of the forearm muscle-bone coupling tremor dynamics model, and the data to be diagnosed is diagnosed again using the optimized model. If the diagnostic result is correct, the result is output as the final diagnosis result, and the diagnostic data is incorporated into the model database, which is dynamically updated.

[0026] The model database provides basic data for the construction of a forearm muscle-bone coupling tremor dynamics model, the theoretical calculation of multi-source tremor parameters, and the combination of multi-sensor data. The model database construction process is as follows: experts diagnose the tremor disease status of the subjects, classify the subjects according to whether they are ill and their disease status, then use the tremor detection module to collect signals from the subjects according to the signal acquisition process, and summarize the collected available data to form a model database. During the actual diagnosis process, the model database is dynamically updated by incorporating the available data collected during the diagnosis process.

[0027] Reference Figure 2 The adaptive model diagnosis module includes five steps: construction of forearm muscle-bone coupling tremor dynamics model, theoretical calculation of multi-source tremor parameters, combination of multi-sensor data, model adaptive update, and disease diagnosis;

[0028] The forearm muscle-bone coupled tremor dynamics model construction link describes the relationship between the tremor feature degradation function and the degradation process of the subject's upper limb tremor symptoms, the tremor symptom multi-source characterization function and multi-sensor data by constructing a multi-source tremor feature common space model; in this link, the tremor feature degradation function is combined with the tremor symptom multi-source characterization function to form a forearm muscle-bone coupled tremor dynamics model based on the multi-source tremor feature common space.

[0029] The multi-source tremor parameter theoretical calculation link calculates unknown parameters such as tremor characteristic parameters and forearm muscle-bone coupling tremor dynamics parameters of the forearm muscle-bone coupling tremor dynamics model based on the sample observation data of tremor subjects in the model database, and inputs the parameters into the forearm muscle-bone coupling tremor dynamics model.

[0030] The multi-sensor data combination link selects the optimal data from the real-time multi-sensor data of the tremor patients collected by the tremor detection module based on the sensitivity of the multi-sensor data of the tremor patients in the model database to the health status, and combines them to eliminate signal redundancy between different sensor data and improve the system's computing efficiency. The combined data is then sent to the forearm muscle-bone coupling tremor dynamics model.

[0031] In the model adaptive updating link, the constructed forearm muscle-bone coupling tremor dynamics model is adaptively updated using the optimal real-time multi-sensor data combination selected in the multi-sensor data combination link, so that the updated model has stronger pertinence to the subject.

[0032] In the disease diagnosis process, the updated forearm muscle-bone coupling tremor dynamics model is used to diagnose tremor diseases in the subject, obtain the diagnosis result and verify it; if the diagnosis result is correct, it is output; if the diagnosis result is incorrect, the forearm muscle-bone coupling tremor dynamics model construction process is feedback optimized, and the subject data is re-diagnosed using the optimized forearm muscle-bone coupling tremor dynamics model.

Claims

1. A multi-sensor wearable tremor detection system based on an adaptive model, comprising a tremor detection module, an adaptive model diagnosis module, and a model database; The tremor detection module includes an electromyographic signal sensor and an acceleration signal sensor, which respectively collect the electromyographic signal of the muscles that control the movement of the forearm and the acceleration signal of the hand; The adaptive model diagnosis module includes five steps: forearm muscle-bone coupling tremor dynamics model construction, multi-source tremor parameter theoretical calculation, multi-sensor data combination, model adaptive update, and disease diagnosis. The forearm muscle-bone coupling tremor dynamics model construction step constructs a forearm muscle-bone coupling tremor dynamics model based on the common space of multi-source tremor characteristics. The multi-source tremor parameter theoretical calculation step calculates the unknown parameters of the forearm muscle-bone coupling tremor dynamics model based on the tremor patient sample observation data in the model database. The multi-sensor data combination step selects the optimal data from the multi-sensor data collected by the tremor detection module based on the sensitivity of the multi-sensor data of the tremor patients in the model database to the health status, and sends the combined data to the forearm muscle-bone coupling tremor dynamics model. The model adaptive update link updates the parameters and states of the forearm muscle-bone coupling tremor dynamics model based on the data obtained in the multi-sensor data combination link; In the disease diagnosis phase, the updated forearm muscle-bone coupling tremor dynamics model was used to diagnose the subjects, obtain the diagnostic results, and verify them; The model database provides basic data for the construction of forearm muscle-bone coupling tremor dynamics model, theoretical calculation of multi-source tremor parameters, and multi-sensor data combination.

2. The system according to claim 1, wherein: The tremor detection module adopts a wearable design and is worn on the forearm of the subject. Wireless transmission is used to transmit signals from the tremor detection module to the host computer.

3. The system according to claim 1, wherein: The forearm muscle-bone coupled tremor dynamics model construction process describes the relationship between the tremor feature degradation function and the degradation process of the subject's upper limb tremor symptoms, the tremor symptom multi-source characterization function and multi-sensor data through a multi-source tremor feature common space model; the tremor feature degradation function is combined with the tremor symptom multi-source characterization function to form a forearm muscle-bone coupled tremor dynamics model based on the multi-source tremor feature common space.

4. The system according to claim 1, wherein: The multi-source tremor parameter theoretical calculation link calculates the tremor characteristic parameters of the forearm muscle-bone coupling tremor dynamics model and the unknown parameters of the forearm muscle-bone coupling tremor dynamics parameters based on the tremor subject sample observation data in the model database, and inputs the parameters into the forearm muscle-bone coupling tremor dynamics model.

5. The system according to claim 1, wherein: The multi-sensor data combination step selects the optimal data from the real-time multi-sensor data of the tremor patients collected by the tremor detection module based on the sensitivity of the multi-sensor data of the tremor patients in the model database to the health status, and sends the combined data to the forearm muscle-bone coupling tremor dynamics model.

6. The system according to claim 1, wherein: In the model adaptive updating link, the constructed forearm muscle-bone coupling tremor dynamics model is adaptively updated using the optimal real-time multi-sensor data combination selected in the multi-sensor data combination link, so that the updated model has stronger pertinence to the subject.

7. The system according to claim 1, wherein: In the disease diagnosis step, the updated forearm muscle-bone coupling tremor dynamics model is used to diagnose tremor diseases in the subject, obtain the diagnosis results and verify them; If the diagnosis result is correct, it is output; if the diagnosis result is incorrect, the process of constructing the forearm muscle-bone coupling tremor dynamics model is feedback optimized, and the subject data is re-diagnosed using the optimized forearm muscle-bone coupling tremor dynamics model.

8. The system according to claim 1, wherein: The model database provides basic data for the construction of the forearm muscle-bone coupling tremor dynamics model, the theoretical calculation of multi-source tremor parameters, and the combination of multi-sensor data. It is composed of multi-sensor data from multiple groups of subjects whose tremor disease status has been determined by experts, and the model database is dynamically updated by incorporating available data collected during the diagnosis process.

Images