TW201720373A - Method and system for measuring spasticity - Google Patents

Abstract

This invention mainly discloses a method for measuring spasticity which is used to solve a problem of the known measuring manner easy to bring about erroneous judgment. The method is apply to a measuring spasticity system having a sensing end electrically connected to a measuring end, and comprises steps of placing the sensing end at a side of a joint of a limb to sense a strength exerting at the side and physics parameters of the side movement, and, placing the measuring end at another side of the joint of the limb to sense the environment temperature, physics parameters of the side movement and collecting the temperature, the strength and the physics parameter of the sides based on a sampling rate during a measurement period. Furthermore, the measuring spasticity system is disclosed. Thus, it can actually resolve the said problem.

Description

Muscle tension sensing method and system

The present invention relates to a muscle force sensing method and system; and more particularly to a muscle tension sensing method and system that can be used to assist a medical professional in estimating the degree of abnormal muscle tone of a rehabilitation person.

With the advancement of medical technology, the success rate of disease treatment has increased, resulting in an increase in life expectancy. However, some patients have not fully recovered after treatment. For example, if a patient with brain or spinal cord injury is treated, if there is still limb hemiplegia, then It is usually necessary to cooperate with the follow-up rehabilitation process, and the period can be fully restored. With the advent of the aging society, the care and medical needs in the rehabilitation process will also increase.

In the rehabilitation process of patients with hemiplegia, cerebral palsy, Parkinson's disease or spinal cord injury, they will experience abnormal muscle tension, leading to clinical symptoms such as muscle pumping and stiffness. The limbs are applied with force, and the relevant scales are used to diagnose whether the muscle tension of the patients is abnormal according to subjective experience. The diagnosis results of different doctors may be different, which is difficult to be objective.

Moreover, the conventional muscle force sensing method mostly uses instruments (such as a constant velocity muscle force tester or a miniature muscle force sensor) to sense quantitative data of a limb movement in a certain part, but cannot rely on data of a single part. Knowing the angle and tension of the limbs on both sides of the joint when the limb is bent, it is easy to cause the doctor (or rehabilitation) to misjudge the situation. Moreover, the conventional constant velocity muscle tester is mainly used for athletes to test, which is not only bulky and expensive; the conventional micro-muscle force sensor can only be used for simple muscle strength test, unable to measure muscle abnormality level, so that the physician or The rehabilitation engineer cannot effectively assess the condition of the patient and immediately dispose of it.

In view of this, it is necessary to improve the shortcomings of the above prior art to meet the actual needs. Seek to improve its practicality.

The present invention provides a muscle tension sensing method which can effectively sense data on the degree of abnormal muscle tone of a patient.

The present invention further provides a muscle tension sensing system capable of effectively sensing data on the degree of abnormal muscle tone of a patient.

The muscle tension sensing system of the present invention may include: a sensing end configured to be disposed on a side of a joint of a limb, the sensing end comprising a pressure sensor and a first inertia sensor, a pressure sensor for sensing a resistance applied to the side limb, the first inertia sensor for sensing a physical quantity when the side limb moves; and a measuring end for being disposed on the other side of the joint of the limb The measuring end includes a temperature sensor, a second inertia sensor and a processing unit, the temperature sensor is configured to sense a temperature of the environment, and the second inertia sensor is configured to sense the other side limb The processing unit is electrically connected to the temperature sensor, the second inertia sensor, the pressure sensor, and the first inertia sensor, and the processing unit is configured according to a measurement time. The sampling rate collects output signals of the temperature sensor, the pressure sensor, the first inertia sensor, and the second inertia sensor.

The muscle tension sensing method disclosed in the present invention can be applied to the muscle tension sensing system, and the system can include a sensing end and a measuring end, the sensing end is electrically connected to the measuring end, and the measuring end is coupled In addition to the mobile computing device, the method may include: arranging the sensing end on one side of the joint of the limb, so that the sensing end senses the resistance applied to the side limb and the physical quantity when the side limb moves; And configuring the measuring end on the other side of the joint of the limb, so that the measuring end senses the temperature of the environment and the physical quantity when the other side limb moves, collecting the temperature according to a sampling rate within a measuring time, Resistance and physical quantity when the two limbs move.

The measuring terminal can include a communication unit, the communication unit is electrically connected to the processing unit, and the communication unit can be coupled to a mobile computing device, and the processing unit of the measuring terminal can The temperature sensor, the pressure sensor, the first inertia sensor and the second inertia sensor output temperature, the resistance and the physical quantity when the two limbs move, constitute a discrete time data, the processing unit or the action The computing device may generate an abnormal level by a machine learning prediction method according to the discrete time data and the two weight matrix of the completed machine learning training; the machine learning prediction method may be a neural network method or a support vector machine method; The action computing device or the processing unit may calculate the angle value and the velocity value of the joint activity according to the output signals of the first inertia sensor and the second inertia sensor, and according to the temperature value, the resistance value, and the angle value of each discrete time And the velocity value and the two weight matrix are calculated to generate the abnormal level. Thereby, the abnormal degree of muscle tension can be determined actively according to the above sensing signal, as an objective basis for determining the muscle tension of the patient, which can assist the physician in diagnosis and avoid or misrepresent the situation of the rehabilitation teacher.

The first inertia sensor and the second inertia sensor respectively comprise an accelerometer, a gyroscope and a direction meter; the measuring terminal is provided with a display electrically connected to the processing unit. Since the components of the system are not expensive parts, it is advantageous to reduce the manufacturing cost.

The above-mentioned muscle tension sensing system and method can automatically generate the above-mentioned abnormality level in the process of patient measurement, and the abnormal level is predicted based on the weight matrix learned from a large amount of training data, which can greatly improve the prediction accuracy, as Determine the objective basis of the patient's muscle tension, avoid the misjudgment caused by the subjective factors of the physician or rehabilitation teacher; the components of the system are not expensive parts, which is beneficial to reduce the manufacturing cost, and the system is easy to operate and can be extended to the family or medical care place. It can achieve the effects of "improving objective prediction" and "avoiding subjective factors misjudging".

1‧‧‧Sense end

11‧‧‧ Pressure Sensor

12‧‧‧First inertia sensor

2‧‧‧Measurement end

21‧‧‧Temperature Sensor

22‧‧‧Second Inertia Sensor

23‧‧‧Processing unit

24‧‧‧ display

25‧‧‧Communication unit

3‧‧‧Mobile computing device

J‧‧‧ Elbow joint

S1‧‧‧Sensing steps

S2‧‧‧Measurement steps

S3‧‧‧ Estimation steps

v, w‧‧‧ weight matrix

x‧‧‧Discrete time data

y, y’‧‧‧痉挛

z‧‧‧Exception series

Figure 1 is a system block diagram of an embodiment of a muscle tension sensing system of the present invention.

Fig. 2 is a schematic enlarged view showing the appearance of the sensing end and the measuring end of the present invention.

Figure 3 is a schematic representation of the use of an embodiment of the muscle tone sensing system of the present invention.

Fig. 4 is a flow chart showing the operation of the embodiment of the muscle tension sensing method of the present invention.

Figure 5 is a schematic illustration of the machine learning training and prediction process of the present invention.

The above and other objects, features and advantages of the present invention will become more <RTIgt; "coupled connection" means that the two devices communicate with each other via electromagnetic coupling technology, such as: using Bluetooth, Wi-Fi or 4G-LTE to transmit data, but not limited thereto. Those of ordinary skill in the art will understand.

Please refer to FIG. 1, which is a system block diagram of an embodiment of the muscle tension sensing system of the present invention. The system embodiment may include a sensing end 1 and a measuring end 2 (the appearance of which may be as shown in FIG. 2 ), and the measuring end 2 is electrically connected to the sensing end 1 , and the sensing end 1 and the measuring end End 2 is used to record data on the force and activity of both sides of the joint of the patient, which can be used as a basis for assessing the degree of abnormal muscle tone.

Referring to FIG. 1 again, the sensing end 1 can be a device containing a sensing function of pressure and movement for arranging on the side of a joint of a patient's limb (eg, wrist, elbow, knee or ankle). (As shown in Fig. 3, the wrist portion on the side of the elbow joint J), the resistance applied to the side limb and the physical quantity when the side limb is moved are sensed. In this embodiment, the sensing end 1 can include a pressure sensor 11 and a first inertia sensor 12, and the pressure sensor 12 can be a piezoelectric sensing component or the like for sensing a The first inertia sensor 13 may include a nine-axis accelerometer, a gyroscope, a directional gauge, etc. for sensing the resistance of the side limbs, such as a physicist or a physician. The physical quantity (such as acceleration, gravity magnetic angle and pointing 3D vector signal), but not limited to this.

Please refer to FIG. 1 again. The measuring end 2 can be a device containing temperature, movement sensing, data processing and communication functions, and is arranged on the other side of the joint of the limb (as shown in FIG. 3). The upper arm of the other side of the elbow joint J) to sense the temperature of the environment, the other limb The physical quantity when the body moves, and collects the temperature, the resistance, and the movement amount of the two limbs according to a sampling rate within a measuring time to form a discrete time data (discrete-time) Data), as a basis for subsequent assessment of abnormalities in muscle tone. In this embodiment, the measuring terminal 2 can include a temperature sensor 21, a second inertia sensor 22, and a processing unit 23, and the temperature sensor 21 can be a sensing of a conventional positive and negative temperature coefficient. The temperature sensing component is configured to sense the temperature of the environment, but the temperature sensing component can also be disposed at the sensing terminal 1 , and is not limited thereto; the second inertia sensor 22 can include an accelerometer The gyroscope and the direction meter are used to sense the physical quantity (such as acceleration, gravity magnetic quantity and pointing three-dimensional vector signals) when the other limb is moved; the processing unit 23 can be microprocessor or digital signal processing. The processing unit 23 is electrically connected to the temperature sensor 21, the second inertia sensor 22, the pressure sensor 11 and the first inertia sensor 12, and the processing unit 23 can perform a process. The processing unit 23 collects the pressure sensor 11 according to the sampling rate (for example, 1 to 30 times/second) during the measurement time (for example, 1 to 40 seconds), and the first inertia sense. The output signals of the detector 12, the temperature sensor 21 and the second inertia sensor 22, such as: the temperature of the first, second, third, ... , The resistance value, acceleration, gravity and magnetic quantities and other points, but not limited to the composition of the discrete-time information.

In addition, as shown in the first and second figures, the measuring terminal 2 can also be provided with a display 24 (such as a seven-segment display or a liquid crystal display, etc.), and the display 24 is electrically connected to the processing unit 23 for displaying the evaluation process. Parameters such as: a measurement time, a sampling rate, or a resistance value output by the pressure sensor 11, a movement amount output by the first inertia sensor 12, a temperature value output by the temperature sensor 21, and a second inertia sensor 22 The amount of movement of the output, etc., but not limited to this.

Referring to FIG. 1 again, the system embodiment may further include a mobile computing device 3 (such as a notebook computer, a tablet computer, a smart phone, or a cloud computing platform) coupled to the measuring terminal 2, and the measuring terminal 2 or the mobile computing device 3 can execute an application according to the discrete time data and the two weight matrix of the completed machine learning training. The machine learning prediction algorithm produces an abnormality degree that can be used by medical personnel (eg, a rehabilitation engineer or physician) as a basis for assessing the degree of abnormal muscle tone in a patient. In this embodiment, the measuring terminal 2 may further include a communication unit 25 (such as a wireless communication unit such as Bluetooth, Wi-Fi or 4G-LTE), and the communication unit 25 is electrically connected to the processing unit 23, and the communication unit is The mobile computing device 3 is coupled to the mobile computing device 3, and the processing unit 23 or the mobile computing device 3 of the measuring terminal 2 can use the machine learning prediction method according to the discrete time data and the two weight matrix (eg, a neural network method or support) The vector machine method, etc.) generates the abnormality level, for example, according to the values in the output signals of the first inertia sensor 12 and the second inertia sensor 22 (eg, acceleration, gravity magnetic quantity, pointing, etc.) The angle value and the velocity value of the joint activity, for example, the angle calculation method can calculate the attitude angle of the sensor by coordinate transformation of the gravity magnetic quantity value, wherein the coordinate conversion method can include: a direction cosine matrix, a quaternion method or The conversion method, such as the Euler angle, can be understood by those having ordinary knowledge in the technical field, and will not be described here; further, according to the temperature value, the resistance value, the angle value and the velocity value of each discrete time and the two weights Generating the abnormality prediction series array (e.g.: 1 to 6), but is not limited thereto.

Please refer to FIG. 4, which is a flow chart showing the operation of the embodiment of the muscle tension sensing method of the present invention. The embodiment of the method can be applied to the embodiment of the muscle tension sensing system. The method embodiment mainly includes a sensing step S1 and a measuring step S2. Please refer to FIGS. 1 and 2 together.

In the sensing step S1, the sensing end 1 can be disposed on one side of the joint of the limb (such as the elbow joint shown in FIG. 2), so that the sensing end 1 senses the resistance applied to the side limb and the The physical quantity when the side limbs move. In this embodiment, the sensing end 1 can be used to sense the resistance applied to the side limb and the physical quantity (such as acceleration, gravity magnetic quantity and pointing three-dimensional vector signal) when the side limb moves. As mentioned above.

In the measuring step S2, the measuring end 2 can be disposed on the other side of the joint of the limb, so that the measuring end 2 senses the temperature of the environment and the physical quantity when the other side limb moves. The temperature, the resistance, and the physical quantity when the two limbs move are collected according to the sampling rate described above. In this embodiment, the measuring end 2 can be used to sense the temperature of the environment, the physical quantity of the other side of the limb when moving (eg, acceleration, gravity magnetic quantity, and three-dimensional vector signals such as pointing), and can be used in the above measuring time. The temperature, the resistance, and the physical quantity when the two limbs move are collected according to the above sampling rate to form the discrete time data, and the description thereof has been described in detail above.

Referring to FIG. 3 again, the method embodiment may further include an estimating step S3, and the measuring end 2 may form the temperature, the resistance, and the physical quantity when the two limbs move to form a discrete time data, and the measuring end 2 or the mobile computing device 3 may generate the abnormal level according to the machine learning prediction method according to the discrete time data and the two weight matrix obtained by completing the machine learning training. In this embodiment, the measuring terminal 2 or the mobile computing device 3 can generate the abnormal level according to the discrete time data and the two weight matrix by a neural network method or a support vector machine method (for example, 1 to 6 levels). ), the description of which has been detailed above. The machine learning training and prediction process of the muscle tension sensing method embodiment is exemplified below, but is not limited thereto.

For example, as shown in FIG. 5, it is a schematic diagram of the machine learning training and prediction process of the present invention. Wherein, the machine learning training method can collect the angle value, the speed value, the pressure value and the temperature value of each time point in different testing processes in the process of applying the muscle tension sensing system in the process of the doctor testing the different patients. And the physician subjective judgment level numerical results (a sample value can be as shown in Table 1 below, the measurement time is 20 seconds, the sampling rate is 1 time / second).

The values in Table 1 above may be arranged into a vector according to a row or a column to form a vector having a dimension n of 80 as the discrete time data x , as shown in the following formula (1) or (1'). Show: x = [30,46,60,77,...,0.1,3.0,2.0,1.0,...,0,0.5,0.5,0.5,...,29.4,29.4,29.3,29.2,...] (1)

x = [30, 0.1, 0, 29.4, 46, 3.0, 0.5, 29.4, ..., 173, 0.0, 1.0, 29.2] (1')

Meanwhile, the physician may be based on muscle tone scale (Ashworth scale as Modified Ashworth scale or shown in the Table II), the performance of muscle strength in patients administered the above-described measurement process determines a spasm series y (three in the following table Shown as 2, and arranged in a vector y=[0,1,0,0,0,0] with a dimension p of 6. Wherein, the above determination process can be performed jointly by a plurality of physicians to reduce errors caused by subjective factors.

In the process of machine learning training (such as: neural network training, training), a large number of discrete time data x and 痉挛 series y can be used for training. The number of intermediate layers of the ANN can be defined as m and m can be defined as 0. The integer value up to 1000, after training, can obtain two sets of weight matrices v , w , the dimension of the weight matrix v is (n+1)×m, and the dimension of the weight matrix w is (m+1)×p, For example, m=12, the dimension of the weight matrix v is 81×12, the dimension of the weight matrix w is 13×6, and the value of each element in the weight matrix v and w is a real value between −1000 and 1000.

Therefore, after the training of the weight matrix v and w is completed, the corresponding machine learning prediction method (for example, neural network prediction method) can be used to predict the 痉挛 y ′ of any discrete time data x , if used The discrete time data x composed of the values described in Table 1 and the two weighting matrices v and w are predicted, and the vector of the 痉挛 痉挛 y′ is obtained as [0.513, 0.667, 0.602, 0.521, 0.379, 0.187], wherein 0.513, 0.667, 0.602, 0.521, 0.379, and 0.187 represent the predicted values of the order of 1, 2, 3, 4, 5, and 6, respectively, so that the maximum number of 0.667 corresponding to the predicted value can be taken from all the predicted values (2) The abnormality level (such as z shown in Fig. 5) is generated, and the abnormality level is 2, which means that the muscle tension is slightly increased. (The affected part is passively flexed and stretched, and suddenly collapses and then relaxes at the end of the joint range of motion. The phenomenon, or the slight resistance in the last paragraph), as a basis for the physician or rehabilitation teacher to determine the muscle tension of the patient.

Therefore, the muscle tension sensing method and system embodiment of the present invention can sense the resistance applied to the side limb and the physical quantity when the side limb moves by using the sensing end disposed on one side of the patient joint; The measuring end of the other side of the joint senses the temperature of the environment and the physical quantity when the other side of the limb moves, and collects the temperature, the resistance, and the movement of the two limbs according to a predetermined sampling rate within a predetermined measuring time. a physical quantity for composing the discrete time data to generate the abnormality level by a machine learning prediction method according to the discrete time data and the two weight matrix of the completed machine learning training, and determining the muscle tension of the patient as a physician or a rehabilitation engineer in accordance with.

In the above, the muscle tension sensing method and system embodiment of the present invention can automatically generate the abnormal level in the process of patient measurement, and the abnormal level is based on a large number of discrete time training. The weight matrix trained by the data and the 痉挛 series is actively predicted, and its prediction accuracy can be greatly improved. It can be used as an objective basis for judging the muscle tension of the patient, avoiding the misjudgment of the subjective factors of the physician or the rehabilitation teacher; the components of the system are not expensive. The parts are beneficial to reduce the manufacturing cost, and the system is easy to operate and can be extended to families or medical care places, and can achieve the effects of "improving objective prediction" and "avoiding human factors as a subjective factor".

While the invention has been described in connection with the preferred embodiments described above, it is not intended to limit the scope of the invention. The technical scope of the invention is protected, and therefore the scope of the invention is defined by the scope of the appended claims.

S1‧‧‧Sensing steps

S2‧‧‧Measurement steps

S3‧‧‧ Estimation steps

Claims (14)

A muscle tension sensing system includes: a sensing end configured to be disposed on a side of a joint of a limb, the sensing end comprising a pressure sensor and a first inertia sensor, the pressure sensing The device is configured to sense a resistance applied to the side limb, the first inertia sensor is configured to sense a physical quantity when the side limb moves; and a measuring end is configured to be disposed on the other side of the joint of the limb, the calculation The end includes a temperature sensor, a second inertia sensor and a processing unit, the temperature sensor is configured to sense the temperature of the environment, and the second inertia sensor is configured to sense the movement of the other side limb The processing unit is electrically connected to the second inertia sensor, the temperature sensor, the pressure sensor and the first inertia sensor, and the processing unit collects according to a sampling rate within a measuring time An output signal of the temperature sensor, the pressure sensor, the first inertia sensor, and the second inertia sensor. The muscle tension sensing system of claim 1, wherein the processing unit outputs the temperature sensor, the pressure sensor, the first inertia sensor, and the second inertia sensor The signals form a discrete time data, and an abnormal progression is generated by a machine learning prediction method according to the discrete time data and the two weight matrix of the completed machine learning training. The muscle tension sensing system of claim 2, wherein the processing unit calculates an angle value and a velocity value of the joint activity according to the output signals of the first inertia sensor and the second inertia sensor, The abnormal level is generated based on the temperature value, the resistance value, the angle value, and the velocity value of each discrete time and the two weight matrix prediction. The muscle tension sensing system of claim 1, wherein the measuring end comprises a communication unit, the communication unit is electrically connected to the processing unit, and the communication unit is coupled to a mobile computing device, and the processing unit The output signals of the temperature sensor, the pressure sensor, the first inertia sensor and the second inertia sensor form a discrete time data, and the mobile computing device performs the machine learning training according to the discrete time data. Two weight matrix The learning prediction method calculates an abnormal series. The muscle tension sensing system according to claim 4, wherein the motion computing device calculates an angle value and a velocity value of the joint activity according to the output signals of the first inertia sensor and the second inertia sensor. The abnormality level is generated according to the temperature value, the resistance value, the angle value, and the velocity value of each discrete time and the two weight matrix prediction. The muscle tension sensing system according to claim 2, wherein the machine learning prediction method is a neural network method or a support vector machine method. The muscle tension sensing system of claim 1, wherein the first inertia sensor and the second inertia sensor respectively comprise an accelerometer, a gyroscope and a directional meter. The muscle tension sensing system of claim 1, wherein the measuring end is provided with a display electrically connected to the processing unit. A muscle tension sensing method is applied to a muscle tension sensing system, the system includes a sensing end and a measuring end, the sensing end is electrically connected to the measuring end, and the measuring end is coupled to a mobile computing device, The step of the method includes: arranging the sensing end on one side of a joint of a limb, so that the sensing end senses a resistance applied to the side limb and a physical quantity when the side limb moves; and configuring the measuring end to The other side of the joint of the limb causes the temperature of the measuring end to sense the temperature of the environment and the physical quantity when the other side of the limb moves, and collects the temperature, the resistance, and the movement of the two limbs according to a sampling rate within a measuring time. The physical quantity of time to form a discrete time data. The muscle tension sensing method according to claim 9, wherein the measuring end combines the temperature, the resistance, and the physical quantity when the two limbs move to form a discrete time data, according to the discrete time data and the completed machine. The two-weight matrix of the learning training is calculated by a machine learning prediction method to generate an abnormal series. The muscle tension sensing method according to claim 10, wherein the measuring end calculates an angle value and a speed value of the joint activity according to a physical quantity when the two limbs move, according to The temperature value, the resistance value, the angle value, and the velocity value of each discrete time are predicted by the two weight matrix to generate the abnormal level. The muscle tension sensing method according to claim 9, wherein the measuring end forms the temperature, the resistance, and the physical quantity when the two limbs move to form a discrete time data, and the mobile computing device is configured according to the discrete time data. The two-weight matrix with the completed machine learning training is calculated by a machine learning prediction method to generate an abnormal series. The muscle tension sensing method according to claim 12, wherein the motion computing device calculates an angle value and a velocity value of the joint activity according to physical quantities when the two limbs move, according to temperature values and resistances of the discrete time. The value, the angle value, and the velocity value are predicted by the two weight matrix to generate the abnormal level. The muscle tension sensing method according to claim 10 or 12, wherein the machine learning prediction method is a type of neural network method or a support vector machine method.