Behavior analysis device and method using acceleration data

Abstract

Disclosed are a behavior analysis device and method using acceleration data. The behavior analysis device according to one embodiment of the present invention comprises: a data collection unit for collecting acceleration data from a sensor attached to one side of the body of a subject; a data conversion unit for expanding the acceleration data into a plurality of pieces of derivative data; a feature extraction unit for extracting feature data from the plurality of pieces of derivative data; and an analysis unit for deriving the behavior of the subject on the basis of the feature data.

Description

Behavior analysis device and method using acceleration data

[0001] The present invention relates to a behavior analysis device and method using acceleration data.

[0002] Currently, behavior analysis technology using acceleration data is being actively researched in various application fields, driven by advancements in wearable devices and IoT (Internet of Things) technology. Accelerometer sensors embedded in devices such as smartphones, smartwatches, and fitness trackers provide the foundation for precisely measuring and analyzing daily movements, gait, and exercise intensity. In particular, combining this data with artificial intelligence technologies such as deep learning enables not only simple motion recognition but also advanced services such as behavior pattern analysis, abnormal behavior detection, and the provision of personalized feedback.

[0003] Recently, various methodologies have been proposed to enhance the accuracy of data processing and interpretation. For example, data fusion technology utilizing multiple sensors enables more precise analysis by supplementing the limited information available through single-sensor sources. Furthermore, with the advancement of cloud computing and edge computing technologies, research is progressing toward processing large-scale data in real time and rapidly providing personalized services. In addition, attempts are being made to apply behavioral analysis data to various healthcare fields, such as fall detection, chronic disease management, and rehabilitation monitoring.

[0004] However, these technologies present challenges in ensuring data accuracy and reliability. Acceleration data is susceptible to noise, and measurement results can vary depending on environmental factors (e.g., device location, individual user movement characteristics). Furthermore, while various signal characteristics can be identified, the high computational cost of calculating all features in real-time leads to a degradation in model generalization performance, necessitating improvements in energy efficiency for data processing and training.

[0005] Accordingly, we aim to provide a behavior analysis device and method using acceleration data that can operate stably even on low-power devices such as wearable devices by detecting changes in the subject's activity and state in real time, reducing unnecessary computations during feature selection and data preprocessing, and lightweighting the model to minimize battery consumption.

[0006] The technology forming the background of this invention is disclosed in Korean Registered Patent Publication No. 10-1560695.

[0007] The present invention aims to solve the problems of the aforementioned conventional technology by precisely analyzing the subject's behavior and activities based on acceleration data to monitor their health status and provide behavior recommendation information.

[0008] The present invention aims to solve the problems of the aforementioned conventional technology by collecting acceleration data, generating derived data, extracting statistical feature data based thereon, and applying it to a behavioral analysis model, thereby increasing the efficiency of data processing and analysis and ensuring reliability.

[0009] The present invention aims to solve the problems of the aforementioned conventional technology by monitoring the behavior or activities of a subject in real time through time-series changes and threshold-based analysis, and by immediately detecting abnormal behavior or specific conditions to provide notifications.

[0010] However, the technical problems that the embodiments of the present invention aim to solve are not limited to the technical problems described above, and other technical problems may exist.

[0011] As a technical means for achieving the above-mentioned technical task, a behavior analysis device using acceleration data according to one embodiment of the present invention may include a data collection unit that collects acceleration data from a sensor attached to one side of a subject's body, a data conversion unit that expands the acceleration data into a plurality of derived data, a feature extraction unit that extracts feature data from the plurality of derived data, and an analysis unit that derives the behavior of the subject based on the feature data.

[0012] In addition, the data collection unit can collect the acceleration data at preset time intervals and sampling rates.

[0013] In addition, the data collection unit can initialize the buffer area where the collected acceleration data is stored.

[0014] In addition, the data conversion unit can generate the plurality of derived data including signal vector magnitude and vertical and horizontal acceleration components for each axis data constituting the acceleration data.

[0015] In addition, the feature extraction unit can extract the feature data including a plurality of statistical feature data from the plurality of derived data.

[0016] In addition, the analysis unit can derive the behavior of the subject by inputting the above feature data into a pre-established behavior analysis model.

[0017] In addition, the analysis unit can derive the subject's activities by analyzing the time-series changes in the subject's behavior derived based on the behavior analysis model.

[0018] In addition, the analysis unit can derive recommended behavior information including at least one of the behaviors of the subject based on the activity of the subject.

[0019] In addition, a behavior analysis device according to one embodiment of the present invention may further include a notification unit that provides notification information based on a preset threshold value regarding the behavior of a subject.

[0020] In addition, the behavior analysis device according to one embodiment of the present invention may further include a data preprocessing unit for preprocessing the acceleration data.

[0021] Meanwhile, a method for analyzing a subject's behavior using acceleration data according to one embodiment of the present invention may include the steps of collecting the acceleration data from a sensor attached to one side of the subject's body, a data conversion step of expanding the acceleration data into a plurality of derived data, a step of extracting feature data from the plurality of derived data, and a step of deriving the subject's behavior based on the feature data.

[0022] The means for solving the problem described above are merely exemplary and should not be interpreted as intended to limit the present invention. In addition to the exemplary embodiments described above, additional embodiments may exist in the drawings and the detailed description of the invention.

[0023] According to the aforementioned means for solving the problem of the present invention, it is possible to precisely analyze the behavior and activities of subjects based on acceleration data to derive results of high reliability, thereby providing data that can be utilized in various application fields.

[0024] According to the solution to the problem of the present invention described above, there is an effect of detecting abnormal behavior or dangerous situations of a subject in real time and responding immediately through time-series data analysis and threshold-based notification functions.

[0025] According to the solution to the problem of the present invention described above, the efficiency of the analysis process can be increased through data preprocessing, derivative data generation, and feature extraction processes, and a device capable of operating stably in various environments can be provided.

[0026] However, the effects obtainable from this invention are not limited to those described above, and other effects may exist.

[0027] FIG. 1 is a schematic diagram of a behavior analysis system using acceleration data according to one embodiment of the present invention.

[0028] FIG. 2 is a schematic block diagram of a behavior analysis device using acceleration data in one embodiment of the present invention.

[0029] FIG. 3 is a diagram exemplarily showing acceleration data collected through a sensor according to one embodiment of the present invention.

[0030] FIG. 4 is a diagram illustrating an exemplary classification of a subject's behavior using acceleration data according to one embodiment of the present invention.

[0031] FIG. 5 is a diagram exemplarily showing the output of current status and notification information regarding a subject's behavior according to one embodiment of the present invention.

[0032] FIG. 6 is a flowchart of an operation for a behavior analysis method using acceleration data according to one embodiment of the present invention.

[0033] Embodiments of the present invention are described below with reference to the attached drawings to enable those skilled in the art to easily implement the invention. However, the present invention may be embodied in various different forms and is not limited to the embodiments described herein. Furthermore, in order to clearly explain the present invention in the drawings, parts unrelated to the explanation have been omitted, and similar parts throughout the specification are denoted by similar reference numerals.

[0034] Throughout this specification, when a part is described as being "connected" to another part, this includes not only cases where they are "directly connected," but also cases where they are "electrically connected" or "indirectly connected" with other elements interposed between them.

[0035] Throughout the entire specification, when a component is described as being located "on," "on top," "on top," "under," "on bottom," or "on bottom" of another component, this includes not only cases where the component is in contact with the other component but also cases where another component exists between the two components.

[0036] Throughout this specification, when a part is described as "comprising" a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components.

[0037] The present invention relates to a behavior analysis device (30) and a method using acceleration data.

[0038] FIG. 1 is a diagram showing the schematic configuration of a behavior analysis system (10) using acceleration data according to one embodiment of the present invention.

[0039] Referring to FIG. 1, a behavior analysis system (10) using acceleration data may include a behavior analysis device using acceleration data (hereinafter, a behavior analysis device (30)) and a user terminal (40).

[0040] Although FIG. 1 illustrates a behavior analysis device (30) connected to a single user terminal (40), it is not limited thereto and may include multiple user terminals (40) owned by medical professionals or administrators that receive information about the behavior of the subject provided by the behavior analysis device (30).

[0041] The user terminal (40) includes all personal computers and laptops, and may include all types of wired / wireless communication devices such as smartphones, smartpads, tablet PCs, PCS (Personal Communication System), GSM (Global System for Mobile communication), PDC (Personal Digital Cellular), PHS (Personal Handyphone System), PDA (Personal Digital Assistant), IMT (International Mobile Telecommunication)-2000, CDMA (Code Division Multiple Access)-2000, W-CDMA (W-Code Division Multiple Access), and Wibro (Wireless Broadband Internet) terminals.

[0042] The behavior analysis device (30) and the user terminal (40) can communicate with each other through a network (20). The network (20) refers to a connection structure that enables information exchange between each node, such as terminals and servers. Examples of such a network (20) include, but are not limited to, a 3GPP (3rd Generation Partnership Project) network, an LTE (Long Term Evolution) network, a 5G network, a WIMAX (World Interoperability for Microwave Access) network, the Internet, a LAN (Local Area Network), a Wireless LAN (Wireless Local Area Network), a WAN (Wide Area Network), a PAN (Personal Area Network), a Wi-Fi network, a Bluetooth network, a satellite broadcasting network, an analog broadcasting network, and a DMB (Digital Multimedia Broadcasting) network.

[0043] The behavior analysis model according to the present invention may be integrated into a behavior analysis device (30) or executed within an application. In one embodiment of the present invention, the behavior analysis model may be located within an application to process and analyze data on a user terminal (40). However, the behavior analysis model is not limited to a specific platform and may be designed to operate in various environments, including devices, applications, and web-based platforms.

[0044] FIG. 2 is a schematic block diagram of a behavior analysis device (30) using acceleration data according to one embodiment of the present invention.

[0045] Referring to FIG. 2, the behavior analysis device (30) may include a data collection unit (110) that collects acceleration data from a sensor attached to one side of the subject's body, a data conversion unit (120) that expands the acceleration data into a plurality of derived data, a feature extraction unit (130) that extracts feature data from the plurality of derived data, and an analysis unit (140) that derives the subject's behavior based on the feature data.

[0046] According to one embodiment of the present invention, the data collection unit (110) can collect acceleration data from a sensor attached to one side of the subject's body.

[0047] The sensor disclosed herein may refer to an accelerometer, but it goes without saying that additional sensors, such as a gyroscope or magnetometer, can be combined as needed to collect analysis data for deriving the subject's behavior.

[0048] FIG. 3 is a diagram exemplarily showing acceleration data collected through a sensor according to one embodiment of the present invention.

[0049] Referring to FIG. 3, the data collection unit (110) can detect changes in acceleration according to the movement of the subject using an accelerometer and collect acceleration data for each of the three axes (X, Y, Z). The acceleration data is collected according to a set sampling period after converting the analog signal into a digital value through an Analog-to-Digital Converter (ADC) built into the accelerometer. The sampling period can generally be set considering the resolution and power efficiency required for analyzing the subject's behavior. For example, a high period (100 Hz or higher) may be applied to detect fast movements.

[0050] In addition, the sensor is designed to be securely attached to the subject's body, and since the characteristics of the collected acceleration data vary depending on the attachment site (e.g., wrist, waist, ankle, etc.), the sensor may be attached to the center of the body, such as the ankle or waist, when analyzing the subject's gait pattern, and the acceleration sensor may be attached to the subject's wrist or arm when analyzing upper body movements. The position where the acceleration sensor is attached can be determined by considering the characteristics of the acceleration data collected from each part of the subject's body, as well as the amplitude and frequency of movement.

[0051] Additionally, the data collection unit (110) can collect acceleration data or electrocardiogram data from an acceleration sensor provided inside the electrocardiogram device.

[0052] Specifically, the data collection unit (110) can collect acceleration data and electrocardiogram data of a subject from an acceleration sensor integrated inside the electrocardiogram device via the network (20). By collecting acceleration data or electrocardiogram data of a subject from the acceleration sensor inside the electrocardiogram device, the data collection unit (110) can analyze the electrocardiogram effects according to the subject's behavior in the analysis unit (140) described later, or extract behavioral information based on the electrocardiogram data. For example, it may be possible to analyze the correlation between the subject's electrocardiogram changes and specific behaviors using acceleration data along with electrocardiogram data.

[0053] According to one embodiment of the present invention, the data collection unit (110) can collect acceleration data at a preset time interval and sampling rate.

[0054] Specifically, the data collection unit (110) can adjust the time interval and sampling rate for collecting acceleration data according to the subject's behavior and the purpose of analysis. For example, a high sampling rate of 100 Hz or higher can be set to detect fine movements of high frequency, and a sampling rate of 10 Hz to 50 Hz can be used when analyzing relatively slow movements or daily activities.

[0055] The time interval for collecting acceleration data in the data collection unit (110) disclosed herein may be 1 second and the sampling rate may be 25 Hz, but is not limited thereto.

[0056] Specifically, the data collection unit (110) collects acceleration data at 1-second intervals and can obtain acceleration data for each axis (X, Y, Z) at a sampling rate of 25 Hz. Accordingly, the acceleration data collected during 1 second may include a total of 25 samples for each axis.

[0057] Additionally, the data collection unit (110) uses a stable clock signal to collect acceleration data at regular time intervals, and the collected acceleration data is stored in an internal memory or buffer in real time. During the data collection process, fluctuations in the sampling rate are minimized to maintain data consistency, and a filtering technique may be applied to ensure the quality of acceleration data according to environmental factors (e.g., body movement, external impact).

[0058] Here, filtering techniques can remove unnecessary noise from acceleration data and enhance necessary signals. For example, noise such as external shocks or high-speed vibrations can be suppressed through a low-pass filter technique that extracts low-frequency signals by removing high-frequency components from acceleration data. Conversely, vibrations or behavioral patterns corresponding to the subject's actions can be emphasized through a high-pass filter technique that separates dynamic motion data by removing low-frequency components. However, it goes without saying that the techniques are not limited to those described above; various filtering methods such as band-pass filters, median filters, and Kalman filters can be utilized depending on the characteristics of the acceleration data (the subject's actions).

[0059] According to one embodiment of the present invention, the data collection unit (110) can initialize a buffer area where collected acceleration data is stored.

[0060] Specifically, the data collection unit (110) includes a buffer area for temporarily storing collected acceleration data, and the buffer may be initialized at regular time intervals or at the time of completion of the acceleration data processing step. The initialization process is performed to secure storage space for new data by removing unnecessary data accumulated in the buffer and to prevent data collisions or duplication.

[0061] More specifically, the data collection unit (110) can automatically initialize the buffer area according to a set period, and the initialization time can be set to a value less than or equal to a predetermined value to support real-time processing of acceleration data. For example, after the behavior of the subject is derived through the analysis of acceleration data, the buffer area can be initialized immediately to efficiently utilize memory resources.

[0062] According to one embodiment of the present invention, the data preprocessing unit can preprocess acceleration data.

[0063] Specifically, the data preprocessing unit can also perform preprocessing to detect and remove outliers. By removing abnormal acceleration data values ​​caused by sudden impact or abnormal movement using a median filter, the consistency and reliability of the acceleration data can be ensured.

[0064] In addition, the data preprocessing unit can adjust the range and scale of the acceleration data by performing preprocessing to standardize or normalize the acceleration data. For example, the comparability between data can be increased through standardization, which sets the mean value of the acceleration data to 0 and the variance to 1.

[0065] According to one embodiment of the present invention, the data conversion unit (120) can expand acceleration data into a plurality of derived data.

[0066] Specifically, the data conversion unit (120) can generate derived data for each of the X, Y, and Z axes of the acceleration data collected at preset time intervals and sampling rates. For example, the Signal Vector Magnitude (SVM) is calculated by combining the X, Y, and Z axis data, and by expressing the intensity of the subject's behavior as a single scalar value, the complex 3-axis data can be simplified while maintaining key features.

[0067] According to one embodiment of the present invention, the data conversion unit (120) can generate a plurality of derived data including signal vector magnitude and vertical and horizontal acceleration components for each axis data constituting acceleration data.

[0068] Here, the derived data may include, but is not limited to, Signal Vector Magnitude (SVM), Vertical Acceleration Component, and Horizontal Acceleration Component. Signal Vector Magnitude (SVM) is data representing the overall intensity of movement by combining X, Y, and Z axis data from the acceleration data, and the magnitude of the acceleration vector can be expressed as a single scalar value. The Vertical Acceleration Component is data representing the component projected from the acceleration data in the direction of the gravity vector, indicating how the subject's body is moving relative to gravity. The Horizontal Acceleration Component is data representing the acceleration component in a plane perpendicular to the gravity vector, signifying the directionality and intensity of the subject's movement.

[0069] Specifically, the data conversion unit (120) can generate derived data by expanding the acceleration data by calculating vertical and horizontal components for each axis in addition to the signal vector magnitude (SVM) based on acceleration data for each of the X, Y, and Z axes. The vertical acceleration component is calculated through projection in the direction of the gravity vector and can be used to analyze static behaviors such as changes in the subject's posture or tilt. On the other hand, the horizontal acceleration component represents movement within a plane perpendicular to the gravity vector and can be used to analyze the subject's dynamic behavior.

[0070] According to one embodiment of the present invention, the feature extraction unit (130) can extract feature data from a plurality of derived data.

[0071] Specifically, the feature extraction unit (130) can extract key feature data that reflects the change pattern or intensity of a specific action regarding the subject's behavior based on multiple derived data.

[0072] More specifically, the feature extraction unit (130) can extract key features by applying mathematical operations and signal processing techniques to multiple derived data. For example, in the feature extraction process, the mean value and standard deviation are calculated to indicate the central tendency and degree of dispersion of the acceleration data to analyze the magnitude and degree of change of the acceleration data, and the intensity of the acceleration data is quantified using RMS (Root Mean Square). The peak value and peak occurrence frequency detect specific patterns that repeat with the maximum value in the signal, and higher-order statistical features such as the signal's rate of change (Derivative), skewness, and kurtosis can be calculated to analyze the asymmetry of the acceleration data and the kurtosis of the data distribution.

[0073] In summary, the feature extraction unit (130) can extract key features from multiple derived data to reduce the complexity of acceleration data and to quantitatively express key information necessary for analyzing the behavior of the subject. The feature data may include, but is not limited to, an average value representing the central tendency of the acceleration data, a standard deviation representing the variability of the data, an RMS representing the intensity of movement by quantifying the energy of the signal, skewness representing the asymmetry of the data, kurtosis describing the kurtosis of the data distribution, and a signal change rate that detects rapid changes.

[0074] According to one embodiment of the present invention, the feature extraction unit (130) can extract feature data including a plurality of statistical feature data from a plurality of derived data.

[0075] Specifically, the feature extraction unit (130) can calculate the average value, standard deviation, maximum value, minimum value, median, variance, etc. between adjacent acceleration data to extract statistical feature data that quantitatively expresses the behavior of the subject.

[0076] In addition, NxM feature data can be extracted by combining M features (mean, standard deviation, maximum, minimum, median, variance, RMS, rate of change, crossover frequency, etc.) for each of the N derived data (signal vector magnitude, vertical acceleration component, horizontal acceleration component) generated for each axis (X-axis, Y-axis, Z-axis) of the acceleration data.

[0077] FIG. 4 is a diagram illustrating an exemplary classification of a subject's behavior using acceleration data according to one embodiment of the present invention.

[0078] Referring to FIG. 4, according to one embodiment of the present invention, the analysis unit (140) can derive the behavior of the subject based on feature data.

[0079] Specifically, the analysis unit (140) can analyze feature data to derive any one of the subject's behaviors, including "Heavy Activity," "Light Activity," "No Activity," "lying (back)," "lying (left)," "lying (right)," "lying (front)," and "Fall down."

[0080] According to one embodiment of the present invention, the analysis unit (140) can derive the behavior of the subject by inputting feature data into a pre-established behavior analysis model.

[0081] Specifically, the analysis unit (140) can derive the behavior of a subject based on a pre-established behavior analysis model that learns from training data labeled with feature data on the behavior of multiple subjects collected in advance, analyzes input feature data, and derives the behavior of the subject.

[0082] More specifically, the analysis unit (140) can classify the behavior of a subject by analyzing input feature data while the pre-established behavior analysis model is trained using a machine learning (ML) or deep learning-based algorithm. The behavior analysis model can calculate a probability value for each behavior class by comparing the input feature data with the training data and derive the behavior class with the highest probability value as the behavior of the subject.

[0083] In the learning phase of the behavior analysis model, the behavior analysis model can learn the characteristics of each behavior class by utilizing a training dataset in which feature data for various behaviors of the subject is labeled. For example, "Heavy Activity" is learned as feature data with high RMS and rate of change, and "lying (back)" is learned as feature data with low RMS and a constant signal pattern. Through this, the analysis unit (140) can derive the subject's behavior by analyzing the feature data extracted from the collected acceleration data.

[0084] In addition, the behavior analysis model according to one embodiment of the present invention may be, for example, a model based on a plurality of machine learning neural networks. The object recognition model may be designed to simulate the structure of the human brain on a computer and may include a plurality of network nodes that simulate neurons of a human neural network and have weights. The plurality of network nodes may each form a connection relationship to simulate the synaptic activity of neurons, in which neurons exchange signals through synapses. In addition, the general model may include, for example, a neural network model or a deep learning model developed from a neural network model. In a deep learning model, a plurality of network nodes may be located at different depths (or layers) and exchange data according to connection relationships. Examples of artificial intelligence models may include, but are not limited to, Deep Neural Networks (DNN), Recurrent Neural Networks (RNN), and Bidirectional Recurrent Deep Neural Networks (BRDNN).

[0085] In addition, a behavior analysis model according to one embodiment of the present invention may be implemented based on a machine learning algorithm. Machine learning is a field of artificial intelligence (AI) that learns patterns from data and performs prediction or classification tasks, characterized by enabling the system to improve itself by utilizing statistical approaches and algorithms. Machine learning algorithms may include various learning methodologies such as supervised learning, unsupervised learning, and reinforcement learning. Supervised learning is a method of training a model with labeled data and is mainly used for classification and regression, while unsupervised learning is used to identify data structures or clusters by analyzing unlabeled data. Reinforcement learning is a method of learning optimal behavior through rewards while interacting with an environment. The model of the present invention may be designed to recognize or analyze specific behavioral patterns by utilizing these machine learning technologies. Examples of machine learning models include Decision Tree, Random Forest, Support Vector Machine (SVM), Naive Bayes, k-Nearest Neighbors (k-NN), Linear Regression, and Logistic Regression, but the machine learning models applicable to the present invention are not limited to the above examples, and it goes without saying that other algorithms may be used depending on various data and analysis purposes.

[0086] According to one embodiment of the present invention, the analysis unit (140) can derive the subject's activity by analyzing the time-series change in the subject's behavior derived based on a behavior analysis model.

[0087] Specifically, the analysis unit (140) can derive activities including "running," "walking," "resting," and "standing" by accumulating records along a time axis based on the subject's behavior derived through a behavior analysis model and analyzing the activity transition patterns at regular time intervals. The behavior analysis model can derive the subject's activities by pre-learning training data labeled with time-series changes of behavior corresponding to the activities of multiple subjects. The aforementioned activities of the subject are not limited to running, walking, resting, and standing, but may include all activities that can be analyzed based on training data labeled with time-series changes of behavior corresponding to the activities of multiple subjects. For example, activities such as "climbing stairs," "descending stairs," "sitting," and "lifting objects" may be included.

[0088] Additionally, the analysis unit (140) can store time-series changes in the behavior of the subject by quantifying the derived behavior of the subject into a predetermined integer value. For example, the behavior of the subject, which analyzes acceleration data collected according to a preset time interval, can be quantified into an integer value to store time-series changes, such as “vigorous activity” being 6, “light activity” being 5, “no activity” being 4, “lying on one’s back” being 3, “lying on one’s left side” being 2, “lying on one’s right side” being 1, “lying face down” being 0, and “falling” being -1.

[0089] According to one embodiment of the present invention, the analysis unit (140) can derive recommended behavior information including at least one of the subject's behaviors based on the subject's activity.

[0090] Specifically, the analysis unit (140) can generate recommended behavior information suitable for the subject by analyzing the subject's current activity status and activities prior to the preset interval based on the subject's activities derived from the subject's time-series behavioral changes according to the preset interval. For example, the analysis unit (140) can recommend light activities such as "walking" if "No Activity" persists for a long time, or suggest "rest" if "Heavy Activity" persists excessively.

[0091] An analysis unit (140) according to one embodiment of the present invention can derive the behavior of the subject based on electrocardiogram data.

[0092] Specifically, the analysis unit (140) can derive the behavior of a subject through a behavior analysis model that has been pre-trained with training data that includes acceleration data and physiological information such as heart rate changes, heart rate intervals (RR intervals), and heart rate variability (HRV) regarding the behavior of multiple subjects.

[0093] For example, if high RMS values ​​appear in acceleration data and the heart rate spikes and HRV decreases in the electrocardiogram (ECG) data, the behavioral analysis model can identify the subject's behavior as "Heavy Activity" or "Fall Down." Conversely, if there is no change in acceleration data and the heart rate is stable with high HRV in the ECG data, the subject's behavior can be identified as "No Activity" or "lying."

[0094] According to one embodiment of the present invention, the notification unit may provide notification information based on a preset threshold value regarding the behavior of the subject.

[0095] Specifically, the notification unit monitors the behavior of the subject derived from the analysis unit (140) in real time and can provide notification information to a user terminal (40) or an administrator terminal when the subject's behavior exceeds a preset threshold or satisfies specific conditions. For example, if the subject remains in a "No Activity" state for longer than a set time, it can provide "long-term inactivity warning" notification information, or if "Heavy Activity" exceeds a threshold, it can generate "excessive activity warning" notification information, and if "Fall down" is detected, it can generate an immediate warning and send notification information to a pre-registered guardian or medical service.

[0096] FIG. 5 is a diagram exemplarily showing the output of current status and notification information regarding a subject's behavior according to one embodiment of the present invention.

[0097] Referring to FIG. 5, the notification unit is connected to a user terminal (40) or an administrator terminal via a network (20) and can provide information (text, graph, notification information) about the target's behavior through an interface built on the user terminal (40) or the administrator terminal.

[0098] Referring to FIG. 5(a), the notification unit can output notification information, such as "Fall detected," through the interface of the user terminal (40) to provide the subject's current behavioral state in a visual and intuitive manner. Additionally, the notification unit can add warning icons or graphic elements to highlight the notification information, thereby helping the subject or relevant parties immediately recognize the behavioral state and take necessary measures.

[0099] Referring to FIG. 5(b), the notification unit can provide information on changes in behavior regarding the subject's behavior in chronological order. For example, it can provide information on changes in behavior including "5 minutes ago: no movement", "3 minutes ago: light activity", "currently: fall detected", etc.

[0100] Below, based on the details described above, we will briefly examine the operation flow of the present invention.

[0101] FIG. 6 is a flowchart of an operation for a behavior analysis method using acceleration data according to one embodiment of the present invention.

[0102] The behavior analysis method illustrated in FIG. 6 can be performed by the behavior analysis device (30) described above. Therefore, even if the content is omitted below, the description of the behavior analysis device (30) can be applied equally to the description of the behavior analysis method.

[0103] In step S11, the data collection unit (110) can collect acceleration data from a sensor attached to one side of the subject's body.

[0104] Additionally, in step S11, the data collection unit (110) can collect the acceleration data at a preset time interval and sampling rate.

[0105] Additionally, in step S11, the data collection unit (110) can initialize a buffer area where the collected acceleration data is stored.

[0106] Next, in step S12, the data conversion unit (120) can expand the acceleration data into a plurality of derived data.

[0107] Additionally, in step S12, the data conversion unit (120) can generate the plurality of derived data including signal vector magnitude and vertical and horizontal acceleration components for each axis data constituting the acceleration data.

[0108] Next, in step S13, the feature extraction unit (130) can extract feature data from the plurality of derived data.

[0109] Additionally, in step S13, the feature extraction unit (130) can extract the feature data including a plurality of statistical feature data from the plurality of derived data.

[0110] Next, in step S14, the analysis unit (140) can derive the behavior of the subject based on the feature data.

[0111] Additionally, in step S14, the analysis unit (140) can input the feature data into a pre-built behavior analysis model to derive the behavior of the subject.

[0112] Additionally, in step S14, the analysis unit (140) can derive the activity of the subject by analyzing the time-series change in the subject's behavior derived based on the behavior analysis model.

[0113] Additionally, in step S14, the analysis unit (140) can derive recommended behavior information including at least one of the behaviors of the subject based on the activity of the subject.

[0114] In the description above, steps S11 through S14 may be further divided into additional steps or combined into fewer steps according to an embodiment of the present invention. Additionally, some steps may be omitted as necessary, and the order of the steps may be changed.

[0115] A method for analyzing behavior using acceleration data according to one embodiment of the present invention may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc., either individually or in combination. The program instructions recorded on the medium may be those specifically designed and configured for the present invention, or they may be those known and available to those skilled in the art of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes; optical recording media such as CD-ROMs and DVDs; magneto-optical media such as floptical disks; and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, and flash memory. Examples of program instructions include machine code, such as that generated by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc. The hardware devices described above may be configured to operate as one or more software modules to perform the operation of the present invention, and vice versa.

[0116] In addition, the aforementioned behavior analysis method can also be implemented in the form of a computer program or application executed by a computer that is stored on a recording medium.

[0117] The foregoing description of the present invention is for illustrative purposes only, and those skilled in the art will understand that other specific forms can be easily modified without altering the technical concept or essential features of the present invention. Therefore, the embodiments described above should be understood as illustrative in all respects and not restrictive. For example, each component described as a single unit may be implemented in a distributed manner, and components described as distributed may likewise be implemented in a combined form.

[0118] The scope of the present invention is defined by the claims set forth below rather than by the detailed description above, and all modifications or variations derived from the meaning and scope of the claims and the concept of equivalents thereof should be interpreted as being included within the scope of the present invention.

Claims

1. In a device for analyzing a subject's behavior using acceleration data, A data collection unit that collects the acceleration data from a sensor attached to one side of the subject's body; A data conversion unit that expands the above acceleration data into a plurality of derived data; A feature extraction unit for extracting feature data from the above plurality of derived data; and An analysis unit that derives the behavior of the subject based on the above feature data, A behavior analysis device including 2. In Paragraph 1, The above data collection unit is, A behavior analysis device that collects the acceleration data at preset time intervals and sampling rates.

3. In Paragraph 2, The above data collection unit is, A behavior analysis device that initializes a buffer area where the above-mentioned collected acceleration data is stored.

4. In Paragraph 3, The above data conversion unit is, A behavior analysis device that generates a plurality of derived data including signal vector magnitude and vertical and horizontal acceleration components for each axis data constituting the acceleration data.

5. In Paragraph 4, The above feature extraction unit is, A behavior analysis device that extracts feature data including multiple statistical feature data from multiple derived data.

6. In Paragraph 5, The above analysis unit is, A behavior analysis device that derives the behavior of the subject by inputting the above feature data into a pre-established behavior analysis model.

7. In Paragraph 6, The above analysis unit is, A behavior analysis device that derives the activity of the subject by analyzing time-series changes in the behavior of the subject derived based on the behavior analysis model.

8. In Paragraph 7, The above analysis unit is, A behavior analysis device that derives recommended behavior information including at least one of the behaviors of the subject based on the activity of the subject.

9. In Paragraph 8, The above behavior analysis device is, A behavior analysis device further comprising a notification unit that provides notification information based on a preset threshold value regarding the behavior of the subject.

10. In Paragraph 1, The above behavior analysis device is, A behavior analysis device further comprising a data preprocessing unit for preprocessing the above acceleration data.

11. In a method for analyzing a subject's behavior using acceleration data, A step of collecting acceleration data from a sensor attached to one side of the subject's body; A data conversion step for expanding the above acceleration data into a plurality of derived data; A step of extracting feature data from the above plurality of derived data; and A step of deriving the behavior of the subject based on the above feature data, A behavioral analysis method including 12. A computer-readable recording medium storing a program for executing the method of paragraph 11 on a computer.

Images