AI-BASED PHYSICAL ACTIVITY AND COGNITIVE ABILITY EVALUATION APPARATUS AND METHOD USING sEMG AND MULTIMODAL DATA

Abstract

An AI-based physical activity and cognitive ability evaluation device comprises: a physical activity data processing unit that generates physical activity data by collecting electromyography data through a surface electromyography (sEMG) sensor and deriving exercise intensity and exercise patterns; a cognitive ability data processing unit that generates cognitive ability data by collecting voice data through a voice sensor and deriving voice clarity and voice frequency patterns; a correlation activity analysis unit that inputs the physical activity data and the cognitive ability data into a multimodal body-cognitive performance evaluation model to generate analyzed body-cognitive performance evaluation data based on respective functional evaluations and correlation evaluations; and a body-cognitive evaluation unit that generates a body-cognitive performance evaluation report by calculating a physical activity index and a cognitive ability index through quantitative evaluation of the body-cognitive performance evaluation data, wherein the body-cognitive performance evaluation index report is configured to allow access to a user-customized exercise guide through user interaction.

Description

Technology Field

[0001] The present invention relates to an AI-based physical activity and cognitive ability evaluation device, and more specifically, to a physical activity and cognitive ability evaluation device and method that evaluates physical activity and cognitive ability by analyzing multimodal data using sensors, including surface electromyography (sEMG) and voice-based eye tracking, using AI, and derives a correlation between physical and cognitive performance to provide customized feedback. Background Technology

[0003] With the recent advancement of an aging society, the rise in neurological diseases, and developments in sports science and rehabilitation therapy, the importance of technologies analyzing the correlation between physical activity and cognitive ability is increasing. Most existing assessment systems for physical and cognitive performance have focused solely on evaluating individual physical or cognitive functions, and there is a lack of technologies to quantitatively analyze the interaction between physical and cognitive performance.

[0004] For example, surface electromyography (sEMG) signal analysis is a representative method used to assess physical activity, but it is limited to measuring muscle activity and has limitations in analyzing its correlation with cognitive function. Furthermore, while voice analysis (speech rate, clarity, etc.) and eye tracking (gaze fixation time, movement path, pupil size changes) technologies are utilized for evaluating cognitive ability, methods to comprehensively assess the relationship with physical activity are currently lacking.

[0006] Korean Published Patent No. 10-2022-0106448 (July 29, 2022) provides a personal exercise management system that integrates artificial intelligence and wireless electromyography signal processing to enable more efficient exercise using objective and quantitative data and analysis results by monitoring muscle activity in real time during exercise.

[0007] A personal exercise management system that combines artificial intelligence and wireless electromyography signal processing evaluates the range of motion of fitness and Taekwondo movements in degrees, induces exercise muscle (neuro-motor) training based on the evaluation results, and includes a method to improve the strength, muscle endurance, or flexibility of a single muscle as well as the neuro-muscular coordination ability of muscle groups associated with specific movements to improve functional exercise ability in directions where the human body's sense of balance is lacking, thereby preventing injury that may occur during functional activities and inducing accurate movements. Prior art literature

[0009] Korean Patent Publication No. 10-2022-0106448 (July 29, 2022) The problem to be solved

[0010] One embodiment of the present invention aims to provide an AI-based physical activity and cognitive ability evaluation device and method that evaluates physical activity and cognitive ability by analyzing surface electromyography (sEMG) and voice data using AI, generates quantitative data on body-cognitive performance ability, evaluates body-cognitive performance ability at the individual and group level by analyzing the correlation between these data, and provides customized feedback based on this.

[0011] One embodiment of the present invention aims to provide an AI-based physical activity and cognitive ability evaluation device and method capable of simultaneously collecting and analyzing data on a user's physical activity and cognitive ability, and quantitatively deriving the correlation between physical and cognitive performance abilities by utilizing an AI-based body-cognitive performance evaluation model.

[0012] One embodiment of the present invention aims to provide an AI-based physical activity and cognitive ability evaluation device and method capable of generating a fused embedding vector by projecting physical performance and cognitive performance data onto a Common Latent Space and precisely analyzing the correlation between physical and cognitive performance based on the fused embedding vector.

[0013] One embodiment of the present invention aims to provide an AI-based physical activity and cognitive ability evaluation device and method that monitors a user's physical activity and cognitive performance in real time, detects abnormal patterns using a machine learning-based anomaly detection technique, analyzes the detected patterns to automatically provide exercise intensity adjustment, additional training recommendations, or expert counseling, and generates quantitative data to support experts in using it for evaluation and diagnosis. means of solving the problem

[0015] Among the embodiments, the AI-based physical activity and cognitive ability evaluation device comprises: a physical activity data processing unit that generates physical activity data by collecting electromyography data through a surface electromyography (sEMG) sensor and deriving exercise intensity and exercise patterns; a cognitive ability data processing unit that generates cognitive ability data by collecting voice data through a voice sensor, deriving voice clarity and voice frequency patterns, and analyzing eye movements, movement paths, and pupil size changes through eye tracking; a correlation activity analysis unit that generates physical-cognitive performance evaluation data analyzed based on respective functional evaluations and correlation evaluations by inputting the physical activity data and the cognitive ability data into a multimodal body-cognitive performance evaluation model; and a body-cognitive evaluation unit that generates a body-cognitive performance evaluation report by calculating a physical activity index and a cognitive ability index through a quantitative evaluation of the body-cognitive performance evaluation data, wherein the body-cognitive performance evaluation index report is configured to allow access to a user-customized exercise guide through user interaction.

[0016] The above physical activity data processing unit detects frequency changes in the electromyogram data and, if the frequency change in the electromyogram data exceeds a threshold, determines a time series analysis interval of the electromyogram data to extract the electromyogram data to be analyzed, determines the most similar reference electromyogram data by applying Dynamic Time Warping (DTW) to the electromyogram data to be analyzed from a reference electromyogram database to derive the exercise intensity and exercise pattern, and can generate muscle activity information estimated through the exercise intensity and exercise pattern as the physical activity data.

[0017] The cognitive ability data processing unit can derive the voice clarity by extracting voice features based on linear prediction coefficients (LPC) from the voice data and derive the voice frequency pattern by extracting voice formants from the voice data, and generate cognitive activity information based on energy changes by frequency band estimated through Mel Spectrogram transformation of the voice clarity and the voice frequency pattern as the cognitive ability data.

[0018] The correlation activity analysis unit can generate physical function evaluation data by analyzing the physical activity data through the physical performance evaluation submodel among the multimodal body-cognitive performance evaluation models, generate cognitive function evaluation data by analyzing the cognitive ability data through the cognitive performance evaluation submodel among the multimodal body-cognitive performance evaluation models, and generate body-cognitive correlation function evaluation data by integrating and analyzing the physical activity data and the cognitive ability data through feature fusion between the physical performance evaluation submodel and the cognitive performance evaluation submodel. The correlation activity analysis unit can generate body-cognitive correlation function evaluation data by generating a fusion embedding vector by projecting the physical function evaluation data and the cognitive function evaluation data onto a common latent space formed between the physical performance evaluation submodel and the cognitive performance evaluation submodel, and by modeling the interaction pattern between physical performance and cognitive performance through the fusion embedding vector. The above-mentioned correlation activity analysis unit can detect whether muscle activity is reduced by evaluating the average power of the electromyogram data during the analysis of the physical activity data through the physical performance evaluation submodel, detect whether cognitive ability is reduced by evaluating the average speech rate of the voice data during the analysis of the cognitive ability data through the cognitive performance evaluation submodel, and detect whether there is body-cognitive dissonance during the integrated analysis of the physical activity data and the cognitive ability data through the fusion embedding vector, and reflect this in the body-cognitive correlation function evaluation data.

[0019] The body-cognitive evaluation unit inputs the physical activity index and cognitive ability index into a reference index model to generate a user performance profile, and based on the user performance profile, classifies the user's physical and cognitive performance levels by range to generate the user-customized exercise guide. Based on the reference index model, the body-cognitive evaluation unit compares and analyzes the physical activity index and cognitive ability index with standard indices of the same age group and gender group; if the user's physical and cognitive performance levels fall below a specific threshold, it adjusts the intensity and frequency of the user-customized exercise guide; it continuously monitors the trends in the user's physical and cognitive performance levels; and if the user's physical and cognitive performance levels are determined to be on a continuous downward trend, it recommends expert consultation to the user. Furthermore, the reference index model can be dynamically subdivided based on the trends in the user's physical and cognitive performance levels to increase the accuracy of comparative analysis. Effects of the invention

[0021] The disclosed technology may have the following effects. However, this does not mean that a specific embodiment must include all of the following effects or only the following effects; therefore, the scope of the rights of the disclosed technology should not be understood as being limited by this.

[0022] An AI-based physical activity and cognitive ability evaluation device and method according to one embodiment of the present invention can evaluate physical activity and cognitive ability by analyzing surface electromyography (sEMG) and voice data using AI, and provide customized feedback by deriving a correlation between physical and cognitive performance.

[0023] An AI-based physical activity and cognitive ability evaluation device and method according to one embodiment of the present invention can simultaneously collect and analyze data on a user's physical activity and cognitive ability, and quantitatively derive the correlation between physical and cognitive performance abilities by utilizing an AI-based body-cognitive performance evaluation model.

[0024] An AI-based physical activity and cognitive ability evaluation device and method according to one embodiment of the present invention projects physical performance ability and cognitive performance ability data onto a Common Latent Space to generate a fused embedding vector, and can precisely analyze the correlation between physical and cognitive performance ability based on the fused embedding vector.

[0025] An AI-based physical activity and cognitive ability evaluation device and method according to one embodiment of the present invention monitors a user's physical activity and cognitive performance in real time, detects abnormal patterns using a machine learning-based anomaly detection technique, and analyzes the detected patterns to automatically provide exercise intensity adjustment, additional training recommendations, or expert counseling. Brief explanation of the drawing

[0027] FIG. 1 is a diagram illustrating an AI-based physical activity and cognitive ability evaluation system according to one embodiment of the present invention. FIG. 2 is a diagram illustrating the system configuration of the AI-based physical activity and cognitive ability evaluation device (130) of FIG. 1. Figure 3 is a diagram illustrating the functional configuration of the AI-based physical activity and cognitive ability evaluation device of Figure 1. Figure 4 is a flowchart illustrating an AI-based physical activity and cognitive ability evaluation method performed in the AI-based physical activity and cognitive ability evaluation device of Figure 1. Figure 5 is a flowchart illustrating the process of collecting data through sensors in an AI-based physical activity and cognitive ability evaluation device, and deriving a final evaluation result through filtering and analysis processes. Figure 6 shows the structure of a surface electromyography (sEMG) sensor device for collecting physical activity data. FIG. 7 shows the configuration of a body-cognitive performance evaluation system (700) utilizing eye tracking, voice speech, and wirelessly connected sEMG sensors. Specific details for implementing the invention

[0028] The description of the present invention is merely an example for structural or functional explanation, and therefore the scope of the present invention should not be interpreted as being limited by the examples described in the text. That is, since the examples are subject to various modifications and may take various forms, the scope of the present invention should be understood to include equivalents capable of realizing the technical concept. Furthermore, the objectives or effects presented in the present invention do not imply that a specific example must include all of them or only such effects; therefore, the scope of the present invention should not be understood as being limited by them.

[0029] Meanwhile, the meaning of the terms described in this application should be understood as follows.

[0030] Terms such as “first,” “second,” etc., are intended to distinguish one component from another, and the scope of rights shall not be limited by these terms. For example, the first component may be named the second component, and similarly, the second component may be named the first component.

[0031] When it is stated that one component is “connected” to another component, it should be understood that it may be directly connected to that other component, or that there may be other components in between. Conversely, when it is stated that one component is “directly connected” to another component, it should be understood that there are no other components in between. Meanwhile, other expressions describing the relationships between components, such as “between” and “exactly between” or “adjacent to” and “directly adjacent to,” should be interpreted in the same way.

[0032] A singular expression should be understood to include a plural expression unless the context clearly indicates otherwise, and terms such as “include” or “have” are intended to specify the existence of the implemented features, numbers, steps, actions, components, parts, or combinations thereof, and should be understood not to preclude the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0033] In each step, identifiers (e.g., a, b, c, etc.) are used for convenience of explanation and do not describe the order of the steps; the steps may occur differently from the specified order unless a specific order is clearly indicated in the context. That is, the steps may occur in the same order as specified, may be performed substantially simultaneously, or may be performed in the reverse order.

[0034] The present invention may be implemented as computer-readable code on a computer-readable recording medium, and the computer-readable recording medium includes all types of recording devices in which data that can be read by a computer system is stored. Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage device, etc. Additionally, the computer-readable recording medium may be distributed across networked computer systems, so that computer-readable code can be stored and executed in a distributed manner.

[0035] Unless otherwise defined, all terms used herein have the same meaning as generally understood by those skilled in the art to which this invention pertains. Terms defined in commonly used dictionaries should be interpreted as having meanings consistent with the context of the relevant technology and should not be interpreted as having an ideal or overly formal meaning unless explicitly defined in this application.

[0037] FIG. 1 is a diagram illustrating an AI-based physical activity and cognitive ability evaluation system according to one embodiment of the present invention.

[0038] Referring to FIG. 1, the AI-based physical activity and cognitive ability evaluation system (100) may include a plurality of user terminals (110), an AI-based physical activity and cognitive ability evaluation device (130), and a database (150).

[0039] The user terminal (110) may correspond to a computing device operated by a user. For example, the user terminal (110) may be implemented as a desktop, laptop, tablet PC, or smartphone, but is not necessarily limited thereto and may be implemented as various devices.

[0040] The user terminal (110) collects physical activity and cognitive ability data including a surface electromyography (sEMG) sensor and a voice sensor, and the collected data can be transmitted to an AI-based physical activity and cognitive ability evaluation device (130) via a network.

[0041] The user terminal (110) may include at least one of a first user terminal (110a), a second user terminal (110b), and a third user terminal (110c). For convenience, the first user terminal (110a) may be used by a first user, the second user terminal (110b) may be used by a second user, and the third user terminal (110c) may be used by a third user.

[0042] In an embodiment of the present invention, each of the first to third user terminals (110a to 110c) may be included in one or more user groups. The one or more user groups may be referred to as the first user group, the second user group, the third user group, etc. Meanwhile, a single user may be included in one or more user groups simultaneously.

[0043] In one embodiment, the user terminal (110) may be implemented as a mobile terminal and may be connected to the AI-based physical activity and cognitive ability evaluation device (130) via cellular communication or Wi-Fi communication. In another embodiment, the user terminal (110) may be a desktop and may be connected to the AI-based physical activity and cognitive ability evaluation device (130) via the Internet.

[0044] The AI-based physical activity and cognitive ability evaluation device (130) is a computing device that can be connected to a user terminal (110) via a network, and can perform the role of analyzing physical activity and cognitive ability data through an AI-based evaluation model and evaluating the user's physical and cognitive performance capabilities. In one embodiment, the AI-based physical activity and cognitive ability evaluation device (130) can analyze physical activity and cognitive ability data to calculate a physical activity index and a cognitive ability index, generate a physical-cognitive performance evaluation result, and provide feedback to the user terminal (110). In addition, the AI-based physical activity and cognitive ability evaluation device (130) may not simply evaluate physical and cognitive data, but may include a function to recommend customized exercise guides and training programs by analyzing the correlation between physical activity and cognitive performance capabilities.

[0045] In one embodiment, the AI-based physical activity and cognitive ability evaluation device (130) may be connected to the user terminal (110) through a dedicated agent installed on the user terminal (110). Here, the dedicated agent may correspond to a software program that enables the user terminal (110) and the AI-based physical activity and cognitive ability evaluation device (130) to interact with each other under the user's approval.

[0046] Meanwhile, the connection and combination between the AI-based physical activity and cognitive ability evaluation device (130) and the user terminal (110) described herein may correspond to one embodiment, and it goes without saying that it can be modified and applied in various forms within a normal range depending on various operation and implementation environments.

[0047] The database (150) may correspond to a storage device that stores various information required during the operation of the AI-based physical activity and cognitive ability evaluation device (130). For example, the database (150) may store data such as the user's physical activity and cognitive ability data, analysis results (physical activity index, cognitive ability index), physical-cognitive performance evaluation results, and user-customized exercise guides.

[0048] In addition, in FIG. 1, the database (150) is shown as a logical storage device included in the AI-based physical activity and cognitive ability evaluation device (130), but is not necessarily limited thereto and can be implemented as a device independent of the AI-based physical activity and cognitive ability evaluation device (130).

[0050] FIG. 2 is a diagram illustrating the system configuration of the AI-based physical activity and cognitive ability evaluation device (130) of FIG. 1.

[0051] Referring to FIG. 2, the AI-based physical activity and cognitive ability evaluation device (130) may include a processor (210), memory (230), user input / output unit (250), network input / output unit (270), and communication port unit (290).

[0052] The processor (210) controls the overall operation of the AI-based physical activity and cognitive ability evaluation device (130) and can manage data flow by interacting with the memory (230), user input / output unit (250), network input / output unit (270), and communication port unit (290).

[0053] The processor (210) can execute a data processing procedure for the AI-based physical activity and cognitive ability evaluation device (130) according to an embodiment of the present invention, analyze physical activity and cognitive ability data, and calculate a physical activity index and a cognitive ability index. Additionally, the processor (210) can manage the memory (230) that is read or written during this process and can schedule the synchronization time between the volatile memory and the non-volatile memory in the memory (230). The processor (210) can control the overall operation of the AI-based physical activity and cognitive ability evaluation device (130) and can control the data flow between the memory (230), the user input / output unit (250), the network input / output unit (270), and the communication port unit (290) by being electrically connected to them.

[0054] The processor (210) can be implemented as a CPU (Central Processing Unit) or GPU (Graphics Processing Unit) of the AI-based physical activity and cognitive ability evaluation device (130), and the amount of data processed can be adjusted according to computational performance.

[0055] The memory (230) may include an auxiliary storage device implemented as non-volatile memory such as an SSD (Solid State Disk) or HDD (Hard Disk Drive) and used to store all data required for the AI-based physical activity and cognitive ability evaluation device (130), and may include a main memory implemented as volatile memory such as RAM (Random Access Memory).

[0056] Additionally, the memory (230) can store a set of instructions for performing a data processing procedure of the AI-based physical activity and cognitive ability evaluation device (130) by being executed by an electrically connected processor (210).

[0057] The memory (230) can store physical activity and cognitive ability data, analyzed physical activity index and cognitive ability index, physical-cognitive performance evaluation results, user-customized exercise guide and training program related data.

[0058] The user input / output unit (250) includes an environment for receiving user input and an environment for outputting specific information to the user, and may include an input device including an adapter such as a touch pad, touch screen, virtual keyboard, or pointing device, and an output device including an adapter such as a monitor or touch screen. In one embodiment, the user input / output unit (250) may correspond to a computing device connected via remote access, and in such case, the AI-based physical activity and cognitive ability evaluation device (130) may be performed as an independent server.

[0059] The network input / output unit (270) provides a communication environment for connecting to a user terminal through a network and may include an adapter for communication such as a LAN (Local Area Network), MAN (Metropolitan Area Network), WAN (Wide Area Network), and VAN (Value Added Network). Additionally, the network input / output unit (270) may be implemented to provide short-range communication functions such as WiFi and Bluetooth, or wireless communication functions of 4G or higher for wireless transmission of data.

[0060] The communication port section (290) can be implemented as a port mapping table that performs data routing during the process of transmitting and receiving data through a network. Here, the communication port section (290) can prevent data collisions by distinguishing sessions when transmitting evaluation requests and result data by assigning a unique source port during the process of transmitting and receiving physical activity and cognitive ability data.

[0062] Figure 3 is a diagram illustrating the functional configuration of the user conversation assistant device based on the conversation room of Figure 1.

[0063] Referring to FIG. 3, the AI-based physical activity and cognitive ability evaluation device (130) includes a physical activity data processing unit (310), a cognitive ability data processing unit (320), a correlation activity analysis unit (330), a body-cognition evaluation unit (340), and a control unit (350).

[0064] At this time, the embodiments of the present invention are not required to include all of the above functional configurations simultaneously; depending on each embodiment, some of the above configurations may be omitted, or some or all of the above configurations may be selectively included. Furthermore, one embodiment of the present invention may be implemented as an independent module that selectively includes some of the above configurations, and an AI-based physical activity and cognitive ability evaluation method may be performed through the interoperability between the modules. The operation of each configuration is described in detail below.

[0066] The physical activity data processing unit (310) performs the function of collecting electromyography data through a surface electromyography (sEMG) sensor and generating physical activity data by analyzing exercise intensity and exercise patterns. The physical activity data processing unit (310) can measure electromyography signals generated in muscles in real time while a user performs a specific exercise and evaluate the muscle activation state by analyzing the data.

[0067] The physical activity data processing unit (310) can derive exercise intensity and exercise patterns by analyzing electromyography data. Exercise intensity can be quantitatively evaluated through the amplitude, Root Mean Square (RMS) value, and frequency band analysis of the electromyography signal, and exercise patterns can be derived by identifying specific exercise movements through time-series data analysis. The physical activity data processing unit (310) can evaluate the user's exercise performance ability and generate physical activity data by comprehensively analyzing exercise intensity and exercise patterns.

[0068] The physical activity data processing unit (310) can detect changes in the frequency of the electromyogram data and, if the change is greater than or equal to a set threshold, determine a time series analysis interval to extract the electromyogram data to be analyzed.

[0069] Frequency analysis is a process of analyzing the changing patterns of specific frequency components in electromyography data. When muscle fatigue increases or muscle contraction patterns change, the frequency band of the electromyography signal may show a tendency to decrease or fluctuate. The physical activity data processing unit (310) may apply Fast Fourier Transform (FFT) and Wavelet Transform techniques to analyze frequency changes. FFT is a method of analyzing the intensity of specific frequency components by converting the input signal into the frequency domain, and Wavelet Transform is a method of detecting detailed changes in the signal by analyzing time and frequency simultaneously.

[0070] The physical activity data processing unit (310) can detect when the frequency change of the electromyogram data changes rapidly above a set threshold and perform the function of selecting the time interval in which the frequency change occurred as the target for analysis. The physical activity data processing unit (310) can apply a time series analysis technique to analyze the pattern of muscle activity change in a specific time interval and derive the characteristics of the exercise pattern.

[0071] The physical activity data processing unit (310) can perform the function of evaluating the user's exercise pattern by comparing the analyzed electromyogram data with a reference electromyogram database and determining the most similar exercise pattern.

[0072] The physical activity data processing unit (310) can quantitatively evaluate the similarity between exercise patterns by applying a Dynamic Time Warping (DTW) algorithm to analyze the similarity of exercise patterns. DTW is an algorithm that can evaluate the similarity between data where the time axis changes irregularly, allowing for precise comparison of exercise movements performed at various speeds. While general time series data comparison methods directly compare values ​​at specific points in time, DTW applies a method that deforms the time axis to align similar patterns and minimizes distance values.

[0073] The physical activity data processing unit (310) can evaluate the user's exercise pattern by applying DTW. For example, the physical activity data processing unit (310) can analyze the electromyogram signal generated when the user performs a specific exercise, such as a squat or push-up, based on DTW, and evaluate how similar the analyzed electromyogram signal is to reference data to determine whether the exercise is performed correctly.

[0074] The physical activity data processing unit (310) can perform the function of generating physical activity data by analyzing exercise intensity and exercise patterns and generating estimated muscle activity information.

[0075] The physical activity data processing unit (310) can analyze exercise intensity by utilizing key indicators such as electromyography RMS value, frequency characteristics, and average power. The physical activity data processing unit (310) can reflect the pattern comparison results with applied DTW to analyze exercise patterns and evaluate how similar the user's exercise movements are to an existing reference pattern.

[0076] The physical activity data processing unit (310) can provide real-time feedback to the user when the exercise intensity exceeds a preset intensity limit or when the fatigue level of a specific muscle group increases abnormally. In addition, the physical activity data processing unit (310) can monitor changes in the user's exercise performance ability by analyzing exercise intensity and exercise pattern data over a long period and provide customized exercise guidance.

[0077] In an embodiment according to the claim, the physical activity data processing unit (310) can evaluate exercise performance ability by analyzing the user's body movements. The physical activity data processing unit (310) can collect movement data by utilizing a motion recognition sensor while the user performs a specific exercise.

[0078] When a user performs exercises such as squats or push-ups, the physical activity data processing unit (310) can record changes in joint angles, exercise speed, and whether balance is maintained. The physical activity data processing unit (310) can analyze exercise patterns using a deep learning-based behavior recognition model and evaluate the user's exercise accuracy by comparing it with standard exercise movements.

[0080] The cognitive ability data processing unit (320) collects voice data through a voice sensor and performs the function of generating cognitive ability data by analyzing voice clarity and voice frequency patterns. The cognitive ability data processing unit (320) can measure the user's voice data in real time and quantitatively analyze the state of cognitive ability by evaluating the pronunciation clarity and frequency changes of the voice.

[0081] The cognitive ability data processing unit (320) can derive speech intelligibility by analyzing collected voice data. Speech intelligibility is an indicator that evaluates how clearly a user's pronunciation is conveyed, and the cognitive ability data processing unit (320) can calculate it through voice feature analysis based on Linear Predictive Coding (LPC).

[0082] Linear Predictive Coding (LPC) is a technique for analyzing a voice signal to extract key features. The voice signal has a certain periodicity and pattern, and the cognitive ability data processing unit (320) can apply LPC to predict the signal linearly and extract key spectral features of the voice. The cognitive ability data processing unit (320) can numerically evaluate voice clarity based on LPC coefficients and determine whether the user's speech is being made normally.

[0083] The cognitive ability data processing unit (320) can analyze the frequency components of voice data and extract voice formants to derive a voice frequency pattern. A formant is the resonance frequency of a voice signal and exhibits the characteristic of energy being concentrated in a specific frequency band. The cognitive ability data processing unit (320) can analyze whether a specific formant pattern is formed during the user's normal speech process and detect whether the formant characteristics change when speech disorders or cognitive decline occur.

[0084] The cognitive ability data processing unit (320) can analyze the frequency spectrum of a voice signal by applying a signal processing technique to perform frequency analysis of voice data and detect whether a specific formant frequency is within a normal range or if a change has occurred.

[0085] The cognitive ability data processing unit (320) can perform Mel Spectrogram conversion based on voice clarity and voice frequency patterns to analyze energy changes by frequency band and generate cognitive activity information.

[0086] A Mel spectrogram is a frequency conversion technique that reflects the characteristics of human auditory perception and provides information that is more suitable for frequency changes perceived by humans than a general frequency spectrum. The cognitive ability data processing unit (320) can analyze the energy distribution of voice data by frequency band using the Mel spectrogram conversion and evaluate the user's cognitive performance ability.

[0087] For example, the cognitive ability data processing unit (320) evaluates whether a strong energy distribution is exhibited in a specific frequency band when the user's voice is spoken normally, and can detect a decrease in specific frequency components or an irregular pattern change when cognitive decline occurs.

[0088] The cognitive ability data processing unit (320) can quantitatively evaluate the user's cognitive performance ability and, if necessary, detect early decline in cognitive function. The cognitive ability data processing unit (320) can analyze voice data over a long period to monitor changes in the user's cognitive performance ability and provide continuous cognitive training or customized feedback.

[0090] In an embodiment according to the claim, the cognitive ability data processing unit (320) performs the function of analyzing the user's eye movement pattern, fixation duration, pupil dilation, etc., by utilizing eye tracking technology. Based on the analysis results of the eye movement pattern, fixation duration, pupil dilation, etc., the cognitive ability data processing unit (320) quantitatively evaluates the user's attention, concentration, and visual information processing ability, and can detect early signs of cognitive function decline.

[0091] The cognitive ability data processing unit (320) evaluates changes in the user's visual attention by analyzing data such as the time the gaze stays in a specific screen area, movement paths between screens, and sudden changes in gaze (saccade). The cognitive ability data processing unit (320) can detect the possibility of reduced attention or reduced cognitive function if the fixed gaze time is excessively long or if the movement of the gaze is delayed compared to normal cognitive activity.

[0092] Additionally, the cognitive ability data processing unit (320) analyzes pupil dilation data to measure the user's cognitive load. If continuous pupil dilation is observed while performing a specific task, the cognitive ability data processing unit (320) may interpret this as a state of high cognitive load.

[0093] The cognitive ability data processing unit (320) integrates and analyzes voice data and eye-tracking data to evaluate comprehensive cognitive performance that is difficult to verify through individual data analysis alone.

[0094] For example, the cognitive ability data processing unit (320) analyzes the pattern of gaze fixation on a specific area and the speed of speech utterance together while the user listens to and answers a given question, and evaluates the correlation between language processing ability and visual information processing ability. In addition, the cognitive ability data processing unit (320) analyzes whether gaze movement and speech utterance are synchronized while the user performs a reading or listening task, and quantitatively measures cognitive performance ability related to reading speed, comprehension, and memory.

[0095] For example, the cognitive ability data processing unit (320) can detect the possibility of reduced language comprehension if the speed of eye movement decreases to a certain level or lower while reading a specific text, and the speed of speech utterance also decreases simultaneously. On the other hand, if the speed of speech utterance is maintained while the eye movement is irregular, the cognitive ability data processing unit (320) can determine that there is a possibility that the user is in a state of attention deficit or is having difficulty processing information.

[0096] In an embodiment other than the claims, the cognitive ability data processing unit (320) performs the function of analyzing multimodal data including physical activity data to derive a correlation between physical fatigue and cognitive performance ability. To this end, the rate of change of eye movement speed and voice response speed is compared, and the relationship with physical activity intensity (e.g., heart rate variability) is analyzed to derive a more precise correlation.

[0097] For example, the cognitive ability data processing unit (320) can evaluate the impact of physical activity on cognitive performance by analyzing eye movement patterns and voice response speeds together while the user performs specific physical activities (e.g., maintaining balance, performing arm movements). Additionally, to measure changes in cognitive response speed due to accumulated physical fatigue, it can perform the function of analyzing voice response delay times and eye movement accuracy at regular time intervals. If the eye fixation time increases and voice response speed decreases during specific physical activities, the cognitive ability data processing unit (320) can determine that physical fatigue is likely associated with a decline in cognitive performance.

[0098] The cognitive ability data processing unit (320) performs the function of monitoring changes in the user's cognitive performance ability over the long term based on the results of comprehensively analyzing eye-tracking data, voice data, and physical activity data, and providing customized cognitive training and feedback.

[0099] For example, when a decrease in attention is detected, the cognitive ability data processing unit (320) recommends visual attention training of an appropriate difficulty level by considering the user's eye movement pattern and voice response speed. In addition, when the voice response speed is reduced, it can guide an optimized auditory-language practice program by reflecting the individual's speech pattern and language processing speed.

[0100] Additionally, the cognitive ability data processing unit (320) can provide cognitive training that combines specific physical activities by considering the correlation between physical and cognitive performance abilities. For example, if the eye movement pattern appears abnormal during the performance of a specific movement, the cognitive ability data processing unit (320) recommends combined training that can simultaneously improve physical movement and visual attention.

[0102] The correlation activity analysis unit (330) inputs physical activity data and cognitive ability data into a multimodal body-cognitive performance evaluation model and performs the function of generating body-cognitive performance evaluation data analyzed based on each function evaluation and correlation evaluation. The correlation activity analysis unit (330) includes a physical performance evaluation submodel and a cognitive performance evaluation submodel, and can perform the function of analyzing the relationship between the two data comprehensively along with the process of independently analyzing physical activity data and cognitive ability data. The correlation activity analysis unit (330) evaluates the correlation between physical performance and cognitive performance and can quantitatively analyze the user's body-cognitive performance.

[0103] The physical performance assessment submodel plays the role of generating physical function assessment data by analyzing physical activity data. It evaluates physical function using electromyography (EMG) data and can detect a decrease in muscle activity by analyzing the average power of EMG signals. For example, when a user performs a specific exercise, the physical performance assessment submodel can analyze the signal intensity of the EMG data to assess the accumulation of muscle fatigue and generate physical function assessment data. The physical performance assessment submodel can determine that there is a high likelihood of physical function decline if muscle fatigue increases or the usage of specific muscles decreases.

[0104] The cognitive performance assessment submodel plays the role of generating cognitive function assessment data by analyzing cognitive ability data. It evaluates cognitive ability using voice data and can detect cognitive decline by analyzing the average speech rate of the voice data. Generally, users with normal cognitive function maintain speech at a constant rate, but in cases of cognitive decline, the speech rate may decrease rapidly or exhibit inconsistent patterns. The cognitive performance assessment submodel evaluates the user's cognitive function by analyzing changes in speech rate and irregularities in voice patterns, and can perform additional evaluations if cognitive decline is suspected.

[0105] The correlation activity analysis unit (330) can perform a feature fusion process between the physical performance evaluation submodel and the cognitive performance evaluation submodel to integrately analyze physical activity data and cognitive ability data. The correlation activity analysis unit (330) can perform the evaluation by not analyzing the physical performance evaluation data and the cognitive performance evaluation data individually, but by projecting the two data onto a Common Latent Space to generate a fusion embedding vector. The fusion embedding vector is a vector that quantitatively expresses the relationship between physical performance and cognitive performance, and can be used to model the interaction pattern between the user's physical and cognitive performance. The physical performance evaluation data and the cognitive performance evaluation data generated during the feature fusion process are projected onto the Common Latent Space, and then a fusion embedding vector is generated to be used to more precisely evaluate the association between physical performance and cognitive performance.

[0106] The correlation activity analysis unit (330) can detect whether there is body-cognitive dissonance by fusion-analyzing physical activity data and cognitive ability data. Body-cognitive dissonance refers to a phenomenon in which an imbalance occurs between physical function and cognitive function, and it may occur when there is a low correlation or abnormal pattern between physical performance ability and cognitive performance ability. For example, if a user exhibits a normal movement pattern physically but has a delayed reaction time or unstable speech patterns cognitively, body-cognitive dissonance may be suspected. Conversely, even if the user's cognitive function is normal but the physical reaction speed is reduced or it is difficult to perform specific movements, it may be evaluated as body-cognitive dissonance. The correlation activity analysis unit (330) can detect whether there is body-cognitive dissonance by fusion-analyzing data from a physical performance ability evaluation submodel and a cognitive performance ability evaluation submodel, and reflect this in the body-cognitive performance ability evaluation data.

[0107] The correlation activity analysis unit (330) performs the role of generating body-cognitive performance evaluation data by comprehensively analyzing the relationship between the physical performance evaluation submodel and the cognitive performance evaluation submodel. The body-cognitive performance evaluation data can be used to provide user-customized feedback based on the correlation between physical function and cognitive function. In addition, the body-cognitive performance evaluation data can monitor the user's physical and cognitive health status through long-term pattern analysis and provide customized training guides for continuous improvement.

[0108] In an embodiment according to the claim, the correlational activity analysis unit (330) can analyze physical activity data and cognitive ability data to evaluate the user's physical-cognitive performance ability and provide age-specific customized training. For example, when a 70-year-old user conducts an evaluation to prevent a decline in physical and cognitive functions, the physical activity data processing unit (310) can calculate a physical activity index by analyzing electromyography signals, and the cognitive ability data processing unit (320) can generate a cognitive ability index by analyzing voice data and reaction speed. The correlational activity analysis unit (330) can generate physical-cognitive performance evaluation data after comprehensively analyzing the physical activity index and the cognitive ability index using a physical performance evaluation submodel and a cognitive performance evaluation submodel. As a result of analyzing the physical activity index and the cognitive ability index, the correlational activity analysis unit (330) can evaluate that the physical activity index is higher than the age group average, while the cognitive ability index is 20% lower than the average. The correlational activity analysis unit (330) can recommend customized training to the user to enhance cognitive function based on the results analyzed through the physical performance evaluation submodel and the cognitive performance evaluation submodel. The correlational activity analysis unit (330) can induce improvement in cognitive ability by providing word recall training, attention enhancement games, and auditory training programs. Additionally, if the correlational activity analysis unit (330) detects body-cognitive dissonance through the physical performance evaluation submodel and the cognitive performance evaluation submodel, it can additionally recommend a complex training program that can simultaneously stimulate physical activity and cognitive function.

[0109] In another embodiment according to the claim, the correlation activity analysis unit (330) can analyze physical activity data and cognitive ability data in real time to detect whether there is body-cognitive dissonance and provide a danger warning to the user. For example, while a user in their 60s is performing a designated exercise program, the physical activity data processing unit (310) can detect muscle activity patterns by measuring electromyography signals, and the cognitive ability data processing unit (320) can evaluate cognitive performance by analyzing speech clarity and reaction speed. If a pattern of decreased muscle activity and decreased speech speed is simultaneously detected during exercise, the correlation activity analysis unit (330) can evaluate the possibility of body-cognitive dissonance by fusing data from the physical performance evaluation submodel and the cognitive performance evaluation submodel. Based on the analysis results, the correlation activity analysis unit (330) can provide a body-cognitive dissonance risk warning to the user and immediately send a notification to a guardian or medical staff. Additionally, the correlation activity analysis unit (330) can suggest a method to resolve dissonance by adjusting the user's exercise intensity and providing additional cognitive assistance training. If a state of dissonance is continuously detected, the correlation activity analysis unit (330) may recommend an additional evaluation by a medical professional.

[0110] In another embodiment according to the claim, the correlation activity analysis unit (330) can accumulate physical activity data and cognitive ability evaluation data of multiple users to continuously update the body-cognitive performance model and optimize the customized exercise-cognitive training algorithm. For example, as a result of analyzing user data from various age groups, it may be analyzed that users aged 65 or older show a strong tendency for increased physical activity to have a positive effect on cognitive ability improvement, whereas users aged 45 or younger show a relatively low direct correlation between physical activity and cognitive ability. The correlation activity analysis unit (330) can optimize the body-cognitive performance evaluation model by age group to automatically recommend the most effective customized training guide when a user of a specific age group undergoes physical and cognitive training. Additionally, the correlation activity analysis unit (330) can apply an AI-based recommendation system that provides a customized exercise-cognitive program by reflecting the user's past evaluation data. The AI-based body-cognitive performance recommendation system can analyze individual patterns of change in physical and cognitive performance to suggest long-term improvement plans and effectively monitor the user's health status.

[0111] In an additional embodiment according to the claim, the correlation activity analysis unit (330) can precisely analyze the correlation between physical activity data and cognitive ability data by applying a feature fusion technique. Unlike the conventional method of analyzing physical activity data and cognitive ability data individually, the correlation activity analysis unit (330) can generate a fusion embedding vector by projecting data between a physical performance evaluation submodel and a cognitive performance evaluation submodel onto a Common Latent Space. The correlation activity analysis unit (330) utilizes the fusion embedding vector to quantitatively express the relationship between physical performance and cognitive performance, and can analyze the interaction pattern between physical and cognitive performance more precisely than the method of evaluating physical activity data and cognitive ability data independently. Based on the correlation between physical and cognitive performance, the correlation activity analysis unit (330) can adjust user-customized training intensity and automatically optimize the exercise-cognitive training program. The correlation activity analysis unit (330) comprehensively evaluates changes in the user's physical and cognitive functions based on the fusion embedding vector and can perform the function of identifying users at risk of physical and cognitive function deterioration under specific conditions at an early stage.

[0112] In an additional embodiment, the correlational activity analysis unit (330) can accumulate physical activity data and cognitive ability data over a long period to analyze changes in physical-cognitive performance and perform long-term health monitoring functions. The physical performance evaluation submodel and the cognitive performance evaluation submodel can conduct evaluations by repeatedly collecting data over a certain period to predict the trend of decline in the user's physical and cognitive functions. For example, if a user in their 50s conducts a physical-cognitive performance evaluation three times a week, the correlational activity analysis unit (330) can analyze changes in the user's physical activity index and cognitive ability index based on data collected over a certain period. During the analysis process, if the user's physical activity index remains relatively constant while the cognitive ability index shows a pattern of continuous decrease of 10% or more over the past six months, the correlational activity analysis unit (330) can classify the user as a risk group for decline in cognitive function. Based on the physical-cognitive performance evaluation data, the correlational activity analysis unit (330) can recommend customized preventive measures to prevent decline in physical and cognitive functions. Users who have been confirmed to have a decrease in cognitive ability index as a result of comparing physical activity data and cognitive ability data may need to perform additional cognitive enhancement training programs. The correlation activity analysis unit (330) can consider physical activity data and cognitive ability data together to evaluate whether the user needs to undergo a precise cognitive test and recommend the test.

[0113] The correlation activity analysis unit (330) can support professional diagnosis and counseling by including a function to share the user's health data with medical staff. The correlation activity analysis unit (330) can predict changes in physical and cognitive performance by integratively analyzing the physical performance evaluation submodel and the cognitive performance evaluation submodel, and can suggest customized training for long-term health improvement. Beyond simple evaluation functions, the correlation activity analysis unit (330) can perform the role of preventing the decline of physical and cognitive functions and monitoring long-term health changes.

[0115] The body-cognitive evaluation unit (340) calculates a physical activity index and a cognitive ability index, and performs the function of generating a body-cognitive performance evaluation report based on the evaluation results. The body-cognitive evaluation unit (340) quantitatively analyzes the user's physical and cognitive performance abilities and is configured to provide a customized exercise guide based on the evaluation data.

[0116] The body-cognition evaluation unit (340) can calculate a physical activity index and a cognitive ability index by quantitatively analyzing the physical activity data and cognitive ability data provided by the correlational activity analysis unit (330). The physical activity index quantifies physical performance ability based on the user's electromyography signal, and the cognitive ability index can be generated by analyzing voice data and reaction speed to evaluate cognitive performance ability.

[0117] The body-cognitive evaluation unit (340) can generate a user's performance profile using a reference index model and classify the performance level into intervals. The body-cognitive evaluation unit (340) can automatically generate a customized exercise guide based on the evaluation results and recommend the most appropriate exercise program by comparing the user's physical activity index and cognitive ability index.

[0118] If the physical activity index is higher than the age group average and the cognitive ability index is lower than the average, the body-cognitive evaluation unit (340) can recommend an exercise program to strengthen cognitive performance. Conversely, if both the physical activity index and the cognitive ability index are analyzed to be lower than the average, a combined training program that can improve physical and cognitive functions in a balanced manner can be provided.

[0119] The body-cognitive evaluation unit (340) can evaluate the level of performance by comparing the user's physical activity index and cognitive ability index with standard indices of the same age group and gender group. If the performance evaluation result is found to be below a certain threshold, the body-cognitive evaluation unit (340) can provide effective training by adjusting the intensity and frequency of the exercise guide.

[0120] The body-cognitive evaluation unit (340) continuously monitors changes in the user's performance ability and may recommend expert counseling if physical and cognitive performance ability continues to decline. Additionally, the body-cognitive evaluation unit (340) can adjust customized exercise guides by reflecting changes in performance ability in real time and support additional expert evaluation and intervention if necessary.

[0121] The body-cognitive evaluation unit (340) can generate a body-cognitive performance evaluation report by synthesizing the user's physical and cognitive performance evaluation data. Through the evaluation report, the user can check changes in their performance and set future training directions.

[0122] The body-cognitive evaluation unit (340) can be configured to include user interaction functions in the evaluation report so that the user can directly access a customized exercise guide. Through the evaluation report, the user can visually check the status of their physical and cognitive performance and, as needed, adjust exercise intensity or add specific training.

[0123] The body-cognitive evaluation unit (340) can perform the function of quantitatively evaluating the user's physical and cognitive performance abilities and providing customized exercise guides based on the analysis results. Additionally, the body-cognitive evaluation unit (340) can generate a user's performance ability profile using a reference index model and evaluate the level of performance ability by comparing it with a standard index.

[0124] The body-cognitive evaluation unit (340) can provide a body-cognitive performance evaluation report to support the user in monitoring their condition and selecting appropriate exercise guides. The body-cognitive evaluation unit (340) can also include a function to continuously track changes in performance and recommend expert counseling if necessary.

[0125] In an embodiment according to the claim, the body-cognitive evaluation unit (340) can evaluate the user's physical and cognitive performance capabilities and recommend a customized exercise program. For example, when a 70-year-old user undergoes a body-cognitive performance evaluation to maintain physical and cognitive functions, the physical activity data processing unit (310) can calculate a physical activity index based on the user's electromyography signal. Additionally, the cognitive ability data processing unit (320) can generate a cognitive ability index by analyzing the user's voice data and reaction speed. The body-cognitive evaluation unit (340) can evaluate performance capabilities by comparing the user's physical activity index and cognitive ability index with standard indices of the same age group and gender group. Based on the evaluation results, the body-cognitive evaluation unit (340) can recommend a customized exercise program to enhance the user's cognitive performance capabilities. For example, memory enhancement training, reaction speed improvement training, or cognitive-exercise combined training may be included.

[0126] In another embodiment according to the claim, the body-cognitive evaluation unit (340) can continuously monitor changes in the user's physical and cognitive performance. For example, if a user in their 60s regularly undergoes a body-cognitive performance evaluation for six months, the physical activity index may remain at a certain level, while the cognitive ability index may show a tendency to continuously decrease. If the pattern of decrease in the cognitive ability index exceeds a specific standard, the body-cognitive evaluation unit (340) may recommend expert counseling. Additionally, it may automatically adjust a customized exercise guide to prevent the user's cognitive function from deteriorating. For example, the body-cognitive evaluation unit (340) may increase the training intensity to improve attention or add game-type training to enhance cognitive stimulation.

[0127] In another embodiment according to the claim, the body-cognition evaluation unit (340) can quantitatively evaluate the cognitive response speed by utilizing the user's eye movement data. The body-cognition evaluation unit (340) can measure how quickly the user recognizes and responds to a target point displayed on the screen through an eye-tracking sensor. When the user conducts a test requiring concentration on a specific visual stimulus (e.g., a number or pattern appearing on the screen), the body-cognition evaluation unit (340) can record the time it takes for the gaze to reach the target point and the speed of eye movement. The body-cognition evaluation unit (340) can analyze this data to quantify the user's visual attention and response speed and include them in a body-cognition performance evaluation report.

[0128] The body-cognitive evaluation unit (340) can evaluate whether the user's cognitive response speed is within the standard range for the age group by comparing it with a reference index model. If the evaluation result is lower than the standard value, the body-cognitive evaluation unit (340) can recommend a customized training program to improve the cognitive response speed. For example, the body-cognitive evaluation unit (340) can induce the user to perform visual tracking training to quickly find a changing target on the screen.

[0129] In an additional embodiment, the body-cognitive evaluation unit (340) can recommend customized training to improve balance by comprehensively analyzing physical and cognitive performance. For example, when a 75-year-old user performs a body and cognitive evaluation for the purpose of fall prevention, the physical activity data processing unit (310) can analyze lower limb muscle activity based on the user's electromyography signal. Additionally, the cognitive ability data processing unit (320) can evaluate the user's reaction speed to analyze neuromuscular responses related to maintaining balance. Based on the evaluation results, the body-cognitive evaluation unit (340) can recommend training to improve balance and reaction speed. For example, this may include balance maintenance training, visual-motor response training, or combined cognitive-motor training. Furthermore, the effectiveness of the training can be analyzed and the training intensity adjusted through re-evaluation after a certain period.

[0130] In an additional embodiment, the body-cognitive evaluation unit (340) can evaluate the physical and cognitive performance of a sports athlete and recommend an optimal training program. For example, if a soccer player performs a body-cognitive evaluation to improve reaction speed during a game, the physical activity data processing unit (310) can evaluate explosive power and muscle reaction speed by analyzing electromyogram signals of the lower limb muscles. Additionally, the cognitive ability data processing unit (320) can evaluate decision-making ability during a game by analyzing visual reaction speed and spatial cognitive ability. The body-cognitive evaluation unit (340) can recommend customized reaction training based on the analyzed data. For example, this may include training to improve reaction speed to rapidly changing visual signals, complex training simulating a game environment, or training requiring both strategic thinking and physical reaction simultaneously. The training intensity may also be adjusted by re-evaluating performance after the training is completed.

[0131] In another additional embodiment, the body-cognitive evaluation unit (340) can analyze the correlation between physical performance and cognitive performance by utilizing a machine learning model. The body-cognitive evaluation unit (340) inputs the user's physical activity index and cognitive ability index into an AI model and can predict a normal performance range based on learned patterns.

[0132] When a pattern is detected in which the user's physical activity index and cognitive ability index fall below a specific threshold, the body-cognitive evaluation unit (340) can apply an anomaly detection algorithm to analyze whether body-cognitive dissonance has occurred. For example, if the user's cognitive performance rapidly declines despite maintaining normal physical exercise performance, the body-cognitive evaluation unit (340) can warn of a possible neurological abnormality.

[0133] The body-cognitive evaluation unit (340) records detected abnormal patterns in an evaluation report and can guide the user to receive expert counseling if necessary. The body-cognitive evaluation unit (340) performs the function of detecting abnormal patterns early by learning long-term data and can continuously monitor the user's health change trends.

[0135] The control unit (350) can perform data processing and control processing between the physical activity data processing unit (310), the cognitive ability data processing unit (320), the correlation activity analysis unit (330), and the body-cognitive evaluation unit (340).

[0137] Figure 4 is a flowchart illustrating an AI-based physical activity and cognitive ability evaluation method performed in the AI-based physical activity and cognitive ability evaluation device of Figure 1.

[0138] In FIG. 4, the AI-based physical activity and cognitive ability evaluation device (130) performs an AI-based physical activity and cognitive ability evaluation procedure, and the AI-based physical activity and cognitive ability evaluation method includes a physical activity data processing step (S410), a cognitive ability data processing step (S430), a correlation activity analysis step (S450), and a body-cognitive evaluation step (S470).

[0140] In the physical activity data processing step (S410), the physical activity data processing unit (310) collects and analyzes the user's physical activity data. The physical activity data processing unit (310) can measure the user's electromyography data in real time using a surface electromyography (sEMG) sensor and evaluate physical performance by analyzing exercise intensity and exercise patterns. The physical activity data processing unit (310) can quantitatively evaluate the amplitude, Root Mean Square (RMS) value, and frequency changes of the electromyography signal and calculate a physical activity index by comparing it with a reference electromyography database. Additionally, the physical activity data processing unit (310) can apply a Dynamic Time Warping (DTW) algorithm to precisely analyze the user's exercise patterns and perform an evaluation of exercise performance.

[0141] In the cognitive ability data processing step (S430), the cognitive ability data processing unit (320) collects and analyzes the user's cognitive ability data. The cognitive ability data processing unit (320) can measure voice data in real time and derive voice clarity by analyzing voice features based on linear prediction coefficients (LPC). Additionally, the cognitive ability data processing unit (320) can extract formants of the voice signal and analyze voice frequency patterns by applying Mel Spectrogram transformation. The cognitive ability data processing unit (320) quantitatively evaluates the user's cognitive performance ability based on voice clarity and response speed, and can calculate a cognitive ability index based on this.

[0142] In the correlation activity analysis step (S450), the correlation activity analysis unit (330) integrately analyzes physical activity data and cognitive ability data. The correlation activity analysis unit (330) can analyze the correlation between the physical activity index and the cognitive ability index by utilizing a physical performance evaluation submodel and a cognitive performance evaluation submodel. To evaluate the relationship between physical performance and cognitive performance more precisely, the correlation activity analysis unit (330) applies a feature fusion technique and can generate a fusion embedding vector by projecting the physical and cognitive performance data onto a Common Latent Space. The correlation activity analysis unit (330) can model the interaction pattern between physical and cognitive performance using the fusion embedding vector and quantitatively evaluate the association between physical performance and cognitive performance. Additionally, the correlation activity analysis unit (330) can detect whether there is physical-cognitive dissonance by comparing the physical activity index and the cognitive ability index, and if an imbalance occurs between physical and cognitive performance, it can reflect this in the evaluation results.

[0143] In the body-cognitive evaluation stage (S470), the body-cognitive evaluation unit (340) can comprehensively evaluate the user's physical and cognitive performance capabilities. The body-cognitive evaluation unit (340) inputs the physical activity index and cognitive ability index into a reference index model to generate a user performance profile, and classifies the level of performance capabilities by intervals by comparing it with standard indices of the same age group and gender group. If the performance evaluation results fall below a specific threshold, the body-cognitive evaluation unit (340) can provide more effective training by adjusting the intensity and frequency of the user-customized exercise guide. Additionally, the body-cognitive evaluation unit (340) continuously monitors changes in the user's physical and cognitive performance capabilities and can recommend expert counseling if performance capabilities decline over the long term.

[0144] The body-cognitive evaluation unit (340) can generate a body-cognitive performance evaluation report by synthesizing body and cognitive performance evaluation data. The body-cognitive performance evaluation report can be provided so that the user can visually check changes in their body and cognitive performance and access customized exercise guides. The user can use the evaluation report to check changes in performance and, if necessary, adjust exercise intensity or add specific training.

[0146] Figure 5 is a flowchart illustrating the process of collecting data through sensors in an AI-based physical activity and cognitive ability evaluation device, and deriving a final evaluation result through filtering and analysis processes.

[0148] The sensor module (510) performs the role of collecting physical activity data and cognitive ability data. The surface electromyography (sEMG) sensor is used to evaluate physical performance by measuring electrical signals of muscles. The eye-tracking sensor can evaluate visual attention and reaction speed by detecting the user's eye movement patterns. The speech recognition sensor analyzes the user's speech patterns and speech clarity, and the motion recognition sensor measures the user's body movements to analyze motor skills and coordination.

[0149] The filter module (520) performs the function of refining raw data collected from the sensor module (510). The filter module (520) applies various filtering techniques to remove noise from the signal input from the sensor and to retain necessary signal components. A low-pass filter removes unnecessary high-frequency noise, while a band-pass filter preserves frequency components that are meaningful for evaluating physical activity and cognitive ability. In the case of physical activity data, signal normalization and smoothing processes may be applied to prevent signal distortion.

[0150] The data collection module (530) performs the role of integrating the refined data obtained through the filter module (520) and converting it into analyzable data. The data collection module (530) aligns physical activity data and cognitive ability data along a time axis and synchronizes data input from various sensors to generate a consistent dataset. The data collection module (530) performs a preprocessing process so that the collected data can be input into an AI-based analysis system and converts the physical activity data and cognitive ability data into a form that can be quantitatively evaluated.

[0151] The AI ​​analysis module (540) performs physical activity and cognitive ability evaluations based on data refined by the data collection module (530). The AI ​​analysis module (540) can learn and evaluate the user's physical activity patterns and cognitive performance patterns by utilizing a neural network-based deep learning model. The AI ​​analysis module (540) independently performs physical performance evaluations and cognitive performance evaluations, and then fuses the two data to analyze the correlation between physical and cognitive performance abilities.

[0152] The integrated control module (550) performs the role of evaluating performance based on the physical activity index and cognitive ability index derived from the AI ​​analysis module and generating user-customized feedback. The integrated control module (550) creates a user's performance profile by comparing the physical activity index and the cognitive ability index, and, if necessary, adjusts the intensity and frequency of the user-customized exercise guide. If an imbalance between physical and cognitive performance is detected, the integrated control module (550) reflects this in the physical-cognitive performance evaluation data and can automatically adjust the exercise intensity so that the user can perform optimal training.

[0153] The result output module (560) performs the role of providing the results of physical activity and cognitive ability evaluations to the user. The result output module (560) converts the physical activity index and cognitive ability index into an evaluation report format so that the user can check the status of their physical and cognitive performance. The evaluation results may be provided in the form of a visual report and a customized training guide, and the user can analyze changes in their performance through the evaluation report and, if necessary, adjust exercise intensity or add specific training.

[0154] As a result, FIG. 5 describes a method for evaluating physical activity and cognitive ability performed in a physical activity and cognitive ability evaluation device (130) step by step, and shows a series of processes including collecting sensor data, performing filtering, performing integrated control through AI analysis, and deriving a final evaluation result.

[0156] Figure 6 shows the structure of a surface electromyography (sEMG) sensor device for collecting physical activity data.

[0157] The sEMG sensor device of Fig. 6 is attached to the user's skin to detect electrical signals of the muscle and performs the function of transmitting them to an analysis device through a data transmission and reception module.

[0159] The sEMG sensor electrodes come into direct contact with the user's skin to detect electrical signals generated by muscles. The sEMG sensor electrodes consist of multiple electrodes arranged to precisely measure muscle activity in specific parts of the body, and can simultaneously analyze the activity of various muscle groups by adjusting the density of the electrodes.

[0160] The elastic sEMG sensor device is designed to adhere closely to the curves of the body and can flexibly deform according to the user's movements. The elastic structure of the elastic sEMG sensor device is designed to be attached to various parts of the body, allowing for the measurement of electromyography signals from various areas such as the wrists, arms, legs, and thighs. In addition, the elastic sEMG sensor device is designed to remain securely attached to the body even when worn for extended periods.

[0161] The data transmission and reception module and the battery serve to store and transmit electromyography data collected from the sEMG sensor electrodes. The data transmission and reception module is configured to wirelessly transmit data in real time and can be linked with an AI analysis system and a physical activity data processing device (310). The battery serves to supply power to the sensor device and can be configured to enable long-term use through an energy-efficient design.

[0162] In conclusion, the sEMG sensor device of Fig. 6 can collect physical activity data in real time and transmit it to an analysis device to be used for evaluating exercise intensity, muscle activation status, and physical performance. Through the sEMG sensor device, the user can analyze exercise patterns and receive expert feedback to perform more effective exercise and rehabilitation treatment.

[0164] FIG. 7 shows the configuration of a body-cognitive performance evaluation system (700) utilizing eye tracking, voice speech, and wirelessly connected sEMG sensors.

[0165] Referring to FIG. 7, a body-cognitive performance evaluation system (700) utilizing voice speech and a wirelessly connected sEMG sensor includes a body-cognitive performance evaluation system (700), an IR built-in camera (710) capable of eye tracking, a built-in microphone (720) for voice recognition, an eye guidance and voice speech content screen (730), and a wirelessly connected sEMG sensor (740).

[0167] The body-cognitive performance evaluation system (700) collects the user's gaze, voice utterance, and body electromyography data in real time and performs the function of integratively analyzing body and cognitive functions. The user can use a digital device such as a tablet to gaze at eye-tracking-based content or perform voice commands, and at the same time, measure muscle activity data through a wirelessly connected sEMG sensor.

[0168] The eye-tracking IR built-in camera (710) performs the function of detecting the user's eye movements and evaluating visual attention and reaction speed by analyzing the fixation time and movement patterns. The eye-tracking IR built-in camera (710) can quantitatively measure cognitive performance by analyzing data such as how long the user stares at a specific location on the screen and how quickly the user moves their eyes. In addition, the eye-tracking IR built-in camera (710) can perform the function of evaluating cognitive load by detecting changes in pupil size. The eye-tracking IR built-in camera (710) can detect the tendency of the pupils to dilate when the user processes complex visual information and evaluate the user's level of cognitive effort by analyzing these changes.

[0169] The built-in microphone (720) for voice recognition detects the user's voice and performs the function of analyzing speech rate, pronunciation accuracy, and acoustic characteristics by applying a voice recognition algorithm. The built-in microphone (720) for voice recognition can analyze cognitive functions by evaluating the user's language expression ability, voice response speed, and sentence construction ability. In addition, the built-in microphone (720) for voice recognition includes a background noise reduction function and an acoustic characteristic analysis function, enabling more accurate voice recognition.

[0170] The gaze guidance and voice utterance content screen (730) serves to provide a visual interface that allows the user to perform visual stimuli and voice commands simultaneously. The user can perform voice utterance according to given instructions while gazing at the gaze guidance and voice utterance content screen (730). The gaze guidance and voice utterance content screen (730) may also include a function to provide personalized feedback by reflecting the user's performance data in real time. For example, if the time the user gazes at a specific target is short or the voice response speed is slow, the gaze guidance and voice utterance content screen (730) may provide additional training content.

[0171] The wirelessly connected sEMG sensor (740) is attached to the user's body and performs the function of measuring muscle activity data in real time. The wirelessly connected sEMG sensor (740) detects electromyography (EMG) signals from specific parts of the user, such as the hands, arms, and legs, and serves to evaluate the state of muscle activity. In addition, the wirelessly connected sEMG sensor (740) supports the user's free movement through wireless connectivity and enables the measurement of body-cognitive performance in various environments. The wirelessly connected sEMG sensor (740) measures the muscle fatigue of the body while the user performs a specific movement and can analyze the relationship with cognitive performance based on the measured data.

[0172] The body-cognitive performance evaluation system (700) performs the function of quantitatively evaluating body-cognitive performance by integratively utilizing eye tracking, voice analysis, and electromyography signal analysis. The body-cognitive performance evaluation system (700) can be utilized in various fields, and major application examples are as follows.

[0173] First, it can be utilized as a system for monitoring cognitive decline in the elderly. By measuring the visual attention and reaction speed of older adults, it can detect cognitive decline at an early stage.

[0174] Second, it can be applied as a sports training evaluation system. By comprehensively analyzing an athlete's physical movements and cognitive responses, it can measure training effectiveness and support performance improvement.

[0175] Third, it can be utilized in rehabilitation therapy and the recovery process of patients with neurological injuries. By continuously monitoring the patient's physical and cognitive performance, it can contribute to analyzing the effectiveness of rehabilitation and establishing customized treatment plans.

[0176] Fourth, it can be applied as a musculoskeletal abnormality detection system. By analyzing specific muscle usage patterns, it can support the early diagnosis of musculoskeletal disorders and provide user-customized exercise guides.

[0177] The body-cognitive performance evaluation system (700) can perform the function of comprehensively evaluating changes in the user's physical and cognitive performance abilities and providing personalized feedback. The body-cognitive performance evaluation system (700) supports the improvement of the user's physical and cognitive functions and can be applied in various fields such as medical and sports training.

[0179] Although the present invention has been described above with reference to preferred embodiments, those skilled in the art will understand that various modifications and changes can be made to the invention without departing from the spirit and scope of the invention as described in the following claims. Explanation of the symbols

[0181] 100: AI-based Physical Activity and Cognitive Ability Assessment System 110: User terminal 130: AI-based physical activity and cognitive ability assessment device 150: Database 210: Processor 230: Memory 250: User I / O 270: Network I / O Section 290: Communication port section 310: Physical activity data processing unit 320: Cognitive ability data processing unit 330: Correlation Activity Analysis Department 340: Body-Cognitive Assessment Section 350: Control unit 510: Sensor module 520: Filter Module 530: Data Collection Module 540: AI Analysis Module 550: Integrated Control Module 560: Result output module 700: Physical-Cognitive Performance Assessment System 710: Eye-tracking built-in IR camera 720: Built-in microphone for voice recognition 730: Eye guidance and voice utterance content screen 740: Wirelessly connected sEMG sensor

Claims

Claim 1 An AI-based physical activity and cognitive ability evaluation device comprising: a physical activity data processing unit that generates physical activity data by collecting electromyography data through a surface electromyography (sEMG) sensor and deriving exercise intensity and exercise patterns; a cognitive ability data processing unit that generates cognitive ability data by collecting voice data through a voice sensor and deriving voice clarity and voice frequency patterns; a correlation activity analysis unit that generates physical-cognitive performance evaluation data analyzed based on respective function evaluations and correlation evaluations by inputting the physical activity data and the cognitive ability data into a multimodal body-cognitive performance evaluation model; and a body-cognitive evaluation unit that generates a body-cognitive performance evaluation report by calculating a physical activity index and a cognitive ability index through a quantitative evaluation of the body-cognitive performance evaluation data, wherein the body-cognitive performance evaluation index report is configured to allow access to a user-customized exercise guide through user interaction. Claim 2 The AI-based physical activity and cognitive ability evaluation device according to claim 1, wherein the physical activity data processing unit detects a frequency change of the electromyogram data and, if the frequency change of the electromyogram data is greater than or equal to a threshold, determines a time series analysis interval of the electromyogram data to extract the electromyogram data to be analyzed, determines the most similar reference electromyogram data by applying Dynamic Time Warping (DTW) to the electromyogram data to be analyzed from a reference electromyogram database to derive the exercise intensity and exercise pattern, and generates muscle activity information estimated through the exercise intensity and exercise pattern as the physical activity data. Claim 3 The AI-based physical activity and cognitive ability evaluation device according to claim 1, wherein the cognitive ability data processing unit extracts voice features based on linear prediction coefficients (LPC) for the voice data to derive the voice clarity and extracts voice formants for the voice data to derive the voice frequency pattern, and generates cognitive activity information based on energy changes by frequency band estimated through Mel Spectrogram transformation of the voice clarity and the voice frequency pattern as the cognitive ability data. Claim 4 The AI-based physical activity and cognitive ability evaluation device according to claim 1, wherein the correlation activity analysis unit generates physical function evaluation data by analyzing physical activity data through a physical performance evaluation submodel among the multimodal body-cognitive performance evaluation models, generates cognitive function evaluation data by analyzing cognitive ability data through a cognitive performance evaluation submodel among the multimodal body-cognitive performance evaluation models, and generates body-cognitive correlation function evaluation data by integrating and analyzing the physical activity data and the cognitive ability data through feature fusion between the physical performance evaluation submodel and the cognitive performance evaluation submodel. Claim 5 In claim 4, the correlation activity analysis unit is characterized by generating a fusion embedding vector by projecting the physical function evaluation data and the cognitive function evaluation data onto a common latent space formed between the physical performance evaluation submodel and the cognitive performance evaluation submodel, and generating body-cognitive correlation function evaluation data by modeling the interaction pattern between physical performance and cognitive performance through the fusion embedding vector. Claim 6 In claim 5, the correlation activity analysis unit detects whether muscle activity is reduced by evaluating the average power of the electromyogram data during the analysis of the physical activity data through the physical performance evaluation submodel, detects whether cognitive ability is reduced by evaluating the average speech rate of the voice data during the analysis of the cognitive ability data through the cognitive performance evaluation submodel, and detects whether there is body-cognitive dissonance during the integrated analysis of the physical activity data and the cognitive ability data through the fusion embedding vector, and reflects this in the body-cognitive correlation function evaluation data, characterized by an AI-based physical activity and cognitive ability evaluation device. Claim 7 The AI-based physical activity and cognitive ability evaluation device according to claim 1, wherein the body-cognitive evaluation unit inputs the physical activity index and cognitive ability index into a reference index model to generate a user performance profile, and classifies the user's physical and cognitive performance levels by segment based on the user performance profile to generate a user-customized exercise guide. Claim 8 In claim 7, the body-cognitive evaluation unit compares and analyzes the physical activity index and the cognitive ability index with standard indices of the same age group and gender group based on the reference index model, and if the user's physical and cognitive performance level is below a specific threshold, adjusts the intensity and frequency of the user-customized exercise guide, continuously monitors the trend of change in the user's physical and cognitive performance level, and if the user's physical and cognitive performance level is determined to be in a continuous downward trend, recommends expert counseling to the user, and the reference index model is dynamically subdivided based on the trend of change in the user's physical and cognitive performance level to increase the accuracy of comparative analysis, characterized by an AI-based physical activity and cognitive ability evaluation device.

Images