TWI856628B - Automaed risk moment scoring system and its operation method - Google Patents

Abstract

The present disclosure provides an operating method of an automated risk moment scoring system, which includes steps as follows. The data sensed by wearable inertial measurement units are processed to obtain concatenate features; the concatenate features are inputted into a custom neural network model to obtain 3D hip, knee and ankle joint angles and movement angles of 3D torso, foot and pelvis; a risk moment scored is generated on the basis of the 3D hip, knee and ankle joint angles and the movement angles of the 3D torso, foot and pelvis.

Description

Automated real-time risk rating system and its operation method

The present invention relates to a system and an operating method thereof, and in particular to an automated real-time risk scoring system and an operating method thereof.

The Landing Error Scoring System (LESS) is a potential clinical assessment tool for predicting injury risk. In the past, scholars have used it to assess the risk of anterior cruciate ligament injury in young athletes.

However, traditional LESS evaluation has many technical problems, such as the need for post-production synchronization of the evaluation images before scoring, camera angle issues, and difficulty observing the tester's movement changes with only two-axis images. In addition, it is also important to adjust the image resolution during movement. Therefore, the processing is time-consuming and the results are not good.

The present invention proposes an automated real-time risk rating system and its operation method to improve the problems of the prior art.

In one embodiment of the present invention, the automated real-time risk scoring system proposed by the present invention includes a plurality of wearable inertia sensors and a computer device, and the computer device establishes communication with the plurality of wearable inertia sensors. The computer device processes the data sensed by the plurality of wearable inertia sensors to obtain a plurality of sequential features, and inputs the plurality of sequential features into a custom neural network model to obtain three-dimensional hip, knee, and ankle joint angles and three-dimensional trunk, foot, and pelvis activity angles, and performs risk scoring in real time.

In one embodiment of the present invention, the computer device includes a storage device and a processor, and the processor is electrically connected to the storage device. The storage device stores the landing scoring error system, and the processor uses the three-dimensional hip, knee, and ankle joint angles and the three-dimensional trunk, foot, and pelvis activity angles to perform risk scoring in real time based on the landing scoring error system.

In one embodiment of the present invention, each of the plurality of wearable inertia sensors includes a Bluetooth module, the computer device includes a transmission device, and the Bluetooth module establishes communication with the transmission device.

In one embodiment of the present invention, each of the multiple wearable inertial sensors further includes a three-axis accelerometer and a gyroscope, the three-axis accelerometer and the gyroscope are electrically connected to a Bluetooth module, and the computer device further includes a processor, the processor is electrically connected to a transmission device. The processor executes a Butterworth low-pass filter, the Butterworth low-pass filter filters the data of the three-axis accelerometer and the gyroscope to obtain filtered data, the filtered data includes unprocessed data, and the multiple sequential features include the filtered data.

In one embodiment of the present invention, each of the multiple wearable inertial sensors further includes a magnetometer, which is electrically connected to the Bluetooth module. The processor inputs the filtered data together with the data of the magnetometer into the complementary filter to estimate the direction of each of the multiple wearable inertial sensors. The above-mentioned multiple sequential features further include the direction of each of the multiple wearable inertial sensors.

In one embodiment of the present invention, the operating method of the automated real-time risk assessment system proposed by the present invention includes the following steps: (A) processing data sensed by multiple wearable inertial sensors to obtain multiple sequential features, and inputting the multiple sequential features into a custom neural network model to obtain three-dimensional hip, knee, and ankle joint angles and three-dimensional trunk, foot, and pelvis movement angles; (B) based on the three-dimensional hip, knee, and ankle joint angles and three-dimensional trunk, foot, and pelvis movement angles, real-time risk assessment.

In one embodiment of the present invention, each of the multiple wearable inertial sensors includes a three-axis accelerometer and a gyroscope, and step (A) includes: filtering the data of the three-axis accelerometer and the gyroscope through a Butterworth low-pass filter to obtain filtered data, and the filtered data includes unprocessed data.

In one embodiment of the present invention, each of the plurality of wearable inertial sensors further comprises a magnetometer, and step (A) further comprises: estimating the direction of each of the plurality of wearable inertial sensors by inputting the filtered data together with the data of the magnetometer into a complementary filter.

In one embodiment of the present invention, each of the multiple wearable inertial sensors further includes a magnetometer, and step (A) further includes: aggregating the filtered data and the direction of each of the multiple wearable inertial sensors as multiple sequential features, so that the customized neural network model is based on the multiple sequential features to obtain the three-dimensional hip, knee, and ankle joint angles and the three-dimensional trunk, foot, and pelvis activity angles.

In one embodiment of the present invention, step (B) includes: preloading a landing error scoring system; utilizing three-dimensional hip, knee, and ankle joint angles and three-dimensional trunk, foot, and pelvis movement angles to perform risk scoring in real time based on the landing error scoring system.

In summary, the technical solution of the present invention has obvious advantages and beneficial effects compared with the existing technology. Through the automated real-time risk scoring system and its operation method, automated real-time risk scoring is realized, which saves time and has high accuracy.

The following will describe the above description in detail with an implementation method and provide a further explanation of the technical solution of the present invention.

In order to make the description of the present invention more detailed and complete, reference may be made to the attached drawings and various embodiments described below, in which the same numbers represent the same or similar elements. On the other hand, well-known elements and steps are not described in the embodiments to avoid unnecessary limitations on the present invention.

Please refer to FIG. 1. The technical aspect of the present invention is an automated real-time risk scoring system 100, which can be applied in the field of sports or widely used in related technical links. The automated real-time risk scoring system 100 of the present technical aspect can achieve considerable technical progress and has a wide range of industrial utilization value. The following will be used in conjunction with FIG. 1 to illustrate the specific implementation of the automated real-time risk scoring system 100.

It should be understood that various implementations of the automated real-time risk scoring system 100 are described in conjunction with FIG. 1. In the following description, for ease of explanation, many specific details are further set to provide a comprehensive description of one or more implementations. However, the present technology can be implemented without these specific details. In other examples, in order to effectively describe these implementations, known structures and devices are shown in block diagram form. The term "for example" used here means "as an example, instance or illustration." Any embodiment described herein as "for example" need not be construed as better or superior to other embodiments.

In practice, in some embodiments of the present invention, the automated real-time risk scoring system 100 may include a computer device 101. The computer device 101 may be a desktop computer, a portable device (such as a laptop, a tablet computer, a mobile phone, etc.), a computing circuit or other computer equipment.

In practice, in some embodiments of the present invention, the computer device 101 can selectively establish communication with the wearable inertia sensor 190. It should be understood that in the embodiments and the scope of the patent application, the description involving "communication" can generally refer to one component indirectly communicating with another component through other components through wired and/or wireless communication, or one component directly communicating with another component without going through other components. For example, the computer device 101 and the wearable inertia sensor 190 can establish Bluetooth communication.

FIG. 1 is a block diagram of an automated real-time risk assessment system 100 according to an embodiment of the present invention. As shown in FIG. 1 , the automated real-time risk assessment system 100 includes a computer device 101 and a wearable inertia sensor 190 .

In some embodiments of the present invention, the number of wearable inertia sensors 190 may be multiple (e.g., eight), which are respectively worn near the thigh, calf, instep, and sacrum. When in use, the computer device 101 establishes communication with the multiple wearable inertia sensors 190. The computer device 101 processes the data sensed by the multiple wearable inertia sensors 190 to obtain multiple concatenate features, and inputs the multiple concatenate features into the custom neural network model to obtain the three-dimensional hip, knee, and ankle joint angles and the three-dimensional trunk, foot, and pelvis activity angles, so as to perform risk scoring in real time.

Regarding the hardware architecture of the computer device 101, as shown in FIG. 1, the computer device 101 includes a storage device 110, a processor 120, a display 130, and a transmission device 140. For example, the storage device 110 may be a hard disk, a flash storage device, or other storage media, the processor 120 may be a central processing unit, the display 130 may be a built-in display or an external screen, and the transmission device 140 may be a communication device (such as a Bluetooth communication device), a serial port, a transmission line, an access medium, or other communication equipment.

In terms of architecture, the transmission device 140 can selectively establish communication with the wearable inertia sensor 190 . The transmission device 140 is electrically connected to the processor 120 , the storage device 110 is electrically connected to the processor 120 , and the processor 120 is electrically connected to the display 130 .

Regarding the hardware structure of the wearable inertial sensor 190, as shown in FIG. 1, the wearable inertial sensor 190 includes a three-axis accelerometer 191, a gyroscope 192, a magnetometer 193, and a Bluetooth module 194. For example, the Bluetooth module 194 can be a Bluetooth transceiver circuit.

In terms of architecture, the Bluetooth module 194 can selectively establish communication with the transmission device 140, the Bluetooth module 194 is electrically connected to the three-axis accelerometer 191, the Bluetooth module 194 is electrically connected to the gyroscope 192, and the Bluetooth module 194 is electrically connected to the magnetometer 193. It should be understood that in the implementation method and the scope of the patent application, the description involving "electrical connection" can generally refer to one component being indirectly electrically coupled to another component through other components, or one component being directly electrically connected to another component without going through other components. For example, the three-axis accelerometer 191, the gyroscope 192, and the magnetometer 193 may be directly electrically connected to the Bluetooth module 194, or the three-axis accelerometer 191, the gyroscope 192, and the magnetometer 193 may be indirectly connected to the Bluetooth module 194 via a control circuit (eg, a microcontroller).

In some embodiments of the present invention, the storage device 110 stores the landing error scoring system, and the processor 120 uses the above-mentioned three-dimensional hip, knee, and ankle joint angles and the three-dimensional trunk, foot, and pelvis movement angles to perform risk scoring in real time based on the landing error scoring system.

In some embodiments of the present invention, the processor 120 executes a Butterworth low-pass filter, which filters the data of the three-axis accelerometer 191 and the gyroscope 192 to obtain filtered data. The filtered data includes unprocessed data (such as: original data, data that is not further processed after filtering, etc.), so as to facilitate subsequent analysis and calculation of a custom neural network model. In a control experiment, other filters are used to filter the data of the three-axis accelerometer 191 and the gyroscope 192. The filtering effect of the control experiment is not as good as the effect of the present invention using the Butterworth low-pass filter to filter the data of the three-axis accelerometer 191 and the gyroscope 192. In practice, for example, the processor 120 uses a Butterworth low-pass filter with a cutoff frequency of about 12 Hz to filter the data of the three-axis accelerometer 191 and the gyroscope 192, but is not limited to this value.

It should be understood that the terms "about", "approximately" or "substantially" used herein are used to modify any quantity that may vary slightly, but such slight variations do not change its essence. If there is no special explanation in the implementation method, the error range of the value modified by "about", "approximately" or "substantially" is generally allowed within 20%, preferably within 10%, and more preferably within 5%.

In some embodiments of the present invention, the processor 120 inputs the filtered data and the data of the magnetometer 193 into a complementary filter to estimate the direction of each of the multiple wearable inertial sensors 190, such as the pitch, roll, and yaw of the wearable inertial sensors 190. In a control experiment, various other filters were used to process the filtered data and the data of the magnetometer 193, and the filtering effect of the control experiment was not as good as the use of complementary filters to filter the filtered data and the data of the magnetometer 193 in the present invention.

Next, in some embodiments of the present invention, the processor 120 aggregates the filtered data and the direction of each of the multiple wearable inertial sensors 190 as a plurality of serial features, so that the multiple serial features include the filtered data and the direction of each of the multiple wearable inertial sensors 190, and the customized neural network model is based on the multiple serial features to obtain the three-dimensional hip, knee, and ankle joint angles and the three-dimensional trunk, foot, and pelvis activity angles.

Next, in some embodiments of the present invention, the processor 120 has formulated a scoring mechanism for each score (including the size of the valve setting) according to the deduction criteria of each item in the landing scoring error system (LESS) project, and then uses the three-dimensional hip, knee, and ankle joint angles and the three-dimensional trunk, foot, and pelvis activity angles to score. The display 130 can display the score. In practice, for example, the processor 120 is implemented by Python and MATALB version of the LESS program.

In order to further explain the operation method of the above-mentioned automatic real-time risk scoring system 100, please refer to Figures 1 and 2 at the same time. Figure 2 is a flow chart of an operation method 200 of the automatic real-time risk scoring system 100 according to an embodiment of the present invention. As shown in Figure 2, the operation method 200 includes steps S201 to S206 (it should be understood that the steps mentioned in this embodiment, except for those whose order is specifically described, can be adjusted according to actual needs, and can even be executed simultaneously or partially simultaneously).

The operating method 200 may take the form of a computer program product on a non-transitory computer-readable recording medium having a plurality of computer-readable instructions embodied in the medium. Suitable recording media may include any of the following: non-volatile memory, such as read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electronically erasable programmable read-only memory (EEPROM); volatile memory, such as static access memory (SRAM), dynamic access memory (SRAM), double data rate random access memory (DDR-RAM); optical storage devices, such as compact disc-read-only ROM (CD-ROM), digital versatile disc-read-only ROM (DVD-ROM); magnetic storage devices, such as hard disk drives and floppy disk drives.

In step S201, a data source is established. In one embodiment of the present invention, a plurality of wearable inertia sensors 190 are respectively worn near the thigh, calf, instep and sacrum, so that the wearable inertia sensors 190 sense the data of the lower limb movement.

In steps S202 to S205, the data sensed by the multiple wearable inertial sensors 190 are processed to obtain multiple sequential features. In step S205, the multiple sequential features are input into the custom neural network model to obtain the three-dimensional hip, knee, and ankle joint angles and the three-dimensional trunk, foot, and pelvis movement angles. In step S206, based on the three-dimensional hip, knee, and ankle joint angles and the three-dimensional trunk, foot, and pelvis movement angles, risk scoring is performed in real time.

Regarding the specific method of step S202, in some embodiments of the present invention, in step S202, the data of the three-axis accelerometer 191 and the gyroscope 192 are filtered through a Butterworth low-pass filter to obtain filtered data, and the filtered data includes unprocessed data.

Regarding the specific method of step S203, in some embodiments of the present invention, in step S203, the direction of each of the multiple wearable inertial sensors 190 is estimated by inputting the filtered data together with the data of the magnetometer 193 into the complementary filter.

Regarding the specific method of step S204, in some embodiments of the present invention, in step S204, the filtered data and the direction of each of the multiple wearable inertial sensors 190 are aggregated as multiple sequential features, so that the customized neural network model is based on the multiple sequential features to obtain the three-dimensional hip, knee, and ankle joint angles and the three-dimensional trunk, foot, and pelvis activity angles.

Regarding the custom neural network model of step S205, in some embodiments of the present invention, the custom neural network model may be the OpenVINO Inference Engine, but is not limited thereto. In a controlled experiment, using additional human body information (such as joint sound signals, human body electrical signals, and human body images) to replace at least one of the above-mentioned multiple sequential features to train the neural network model, or omitting at least one of the above-mentioned multiple sequential features to train the neural network model will result in poor accuracy of the trained neural network model. The self-defined neural network model of the present invention is trained through the sequential features of a large number of subjects. The trained self-defined neural network model has good accuracy and can obtain more accurate three-dimensional hip, knee, and ankle joint angles and three-dimensional trunk, foot, and pelvis movement angles in step S205.

Regarding the specific method of step S206, in some embodiments of the present invention, in step S206, a landing error scoring system is preloaded; the three-dimensional hip, knee, and ankle joint angles and the three-dimensional trunk, foot, and pelvis movement angles are used to perform risk scoring in real time based on the landing error scoring system.

In summary, the technical solution of the present invention has obvious advantages and beneficial effects compared with the prior art. By means of the automated real-time risk scoring system 100 and the operation method 200 thereof, automated real-time risk scoring is realized, which saves time and has high accuracy.

Although the present invention has been disclosed in the above embodiments, it is not intended to limit the present invention. Anyone skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the scope defined in the attached patent application.

In order to make the above and other objects, features, advantages and embodiments of the present invention more clearly understood, the attached symbols are explained as follows:

100: Automated real-time risk rating system

101:Computer Devices

110: Storage device

120:Processor

130: Display

140: Transmission device

190: Inertial sensor

191: Three-axis accelerometer

192: Gyroscope

193:Magnetometer

194:Bluetooth module

200: How it works

S201～S206: Steps

In order to make the above and other purposes, features, advantages and embodiments of the present invention more clearly understandable, the attached drawings are described as follows: Figure 1 is a block diagram of an automated real-time risk rating system according to an embodiment of the present invention; and Figure 2 is a flow chart of an operation method of an automated real-time risk rating system according to an embodiment of the present invention.

200: How it works

S201~S206: Steps

Claims (10)

An automated real-time risk assessment system includes: a plurality of wearable inertial sensors, which are respectively used to be worn near the thigh, calf, instep and sacrum, so that the wearable inertial sensors sense data of lower limb movements; and a computer device, which establishes communication with the wearable inertial sensors, and processes the data of the lower limb movements sensed by the wearable inertial sensors to obtain a plurality of sequential features, and inputs the plurality of sequential features into a custom neural network model to obtain three-dimensional hip, knee and ankle joint angles and three-dimensional trunk, foot and pelvis activity angles, so as to perform a risk assessment in real time. The automated real-time risk scoring system as described in claim 1, wherein the computer device comprises: a storage device storing a landing scoring error system; and a processor electrically connected to the storage device, the processor utilizing the three-dimensional hip, knee, and ankle joint angles and the three-dimensional trunk, foot, and pelvis movement angles to perform the risk scoring in real time according to the landing scoring error system. An automated real-time risk scoring system as described in claim 1, wherein each of the multiple wearable inertial sensors includes a Bluetooth module, the computer device includes a transmission device, and the Bluetooth module establishes communication with the transmission device. The automated real-time risk scoring system as described in claim 3, wherein each of the multiple wearable inertial sensors further comprises a three-axis accelerometer and a gyroscope, the three-axis accelerometer and the gyroscope are electrically connected to the Bluetooth module, and the computer device further comprises: a processor electrically connected to the transmission device, the processor executes a Butterworth low-pass filter, the Butterworth low-pass filter filters the data of the three-axis accelerometer and the gyroscope to obtain filtered data, the filtered data includes unprocessed data, and the multiple sequential features include the filtered data. An automated real-time risk scoring system as described in claim 4, wherein each of the multiple wearable inertial sensors further comprises a magnetometer, the magnetometer is electrically connected to the Bluetooth module, the processor inputs the filtered data together with the data of the magnetometer into a complementary filter to estimate the direction of each of the multiple wearable inertial sensors, and the multiple sequential features further comprise the direction of each of the multiple wearable inertial sensors. An operating method of an automated real-time risk assessment system, the operating method comprising the following steps: (A) processing data of a lower limb movement sensed by a plurality of wearable inertial sensors to obtain a plurality of sequential features, inputting the plurality of sequential features into a custom neural network model to obtain three-dimensional hip, knee, and ankle joint angles and three-dimensional trunk, foot, and pelvic movement angles, wherein the plurality of wearable inertial sensors are respectively used to be worn near the thigh, calf, instep, and sacrum, so that the plurality of wearable inertial sensors sense the data of the lower limb movement; and (B) performing a risk assessment in real time based on the three-dimensional hip, knee, and ankle joint angles and the three-dimensional trunk, foot, and pelvic movement angles. The operating method as described in claim 6, wherein each of the multiple wearable inertial sensors includes a three-axis accelerometer and a gyroscope, and step (A) includes: filtering the data of the three-axis accelerometer and the gyroscope through a Butterworth low-pass filter to obtain filtered data, and the filtered data includes unprocessed data. The operating method as described in claim 7, wherein each of the multiple wearable inertial sensors further comprises a magnetometer, and step (A) further comprises: estimating the direction of each of the multiple wearable inertial sensors by inputting the filtered data together with the data of the magnetometer into a complementary filter. The operating method as described in claim 8, wherein each of the multiple wearable inertial sensors further comprises a magnetometer, and step (A) further comprises: Aggregating the filtered data and the direction of each of the multiple wearable inertial sensors as the multiple sequential features, so that the customized neural network model can obtain the three-dimensional hip, knee, and ankle joint angles and the three-dimensional trunk, foot, and pelvis activity angles based on the multiple sequential features. The operating method as described in claim 6, wherein step (B) includes: preloading a landing error scoring system; and using the three-dimensional hip, knee, and ankle joint angles and the three-dimensional trunk, foot, and pelvis movement angles to perform the risk scoring in real time based on the landing error scoring system.

Images