Multi-model-based ftp estimation method, device, and computer program

Abstract

Disclosed is a multi-model-based FTP estimation method for implementing the above-described task. The method comprises the steps of: acquiring user time-series data; selecting, as a user-customized FTP estimation model, one model from among a plurality of FTP estimation models on the basis of the user time-series data; and generating FTP information by using the user-customized FTP estimation model, wherein the user time-series data includes time-power data and heart rate data generated during riding by a user, and the plurality of FTP estimation models may be models for estimating the FTP information in different ways.

Description

Multi-model based FTP estimation method, device and computer program

[0001] The present invention relates to a method for estimating FTP (Functional Threshold Power) based on multiple models, and more specifically, to a method, device, and computer program for analyzing a user's time series data to select a suitable model among a plurality of FTP estimation models, and generating user-customized FTP information by utilizing the selected model.

[0002] FTP (Functional Threshold Power) is the exercise intensity at which lactic acid levels remain stable (Maximal Lactate Steady State, MLSS). It is the power that can be continuously exerted for 20 to 70 minutes, depending on training level. It has recently become widely used as a key indicator for managing cycling competitions and training. FTP provides essential data for assessing cycling performance and developing training plans, making it crucial for various training platforms and equipment to include FTP estimation capabilities. This allows for efficient analysis of a user's fitness level and physical condition, enabling the development of personalized training plans.

[0003] The most common method of measuring FTP currently in use is based on the FTP test protocol. The most widely known method is the 20-minute FTP test, which measures the average power measured by maximal effort over 20 minutes and multiplies it by a correction factor of 0.95 to estimate the power value at the exercise intensity at which lactate levels remain stable (MLSS, Maximal Lactate Steady State). Other methods include the RAMP test, which gradually increases power, and methods that measure maximum peak power. However, these methods can result in varying results depending on the user's physical condition and test protocol. Furthermore, they are burdensome and difficult to perform, especially for the average user.

[0004] Recently, a method for estimating submaximal FTP has emerged, leveraging the correlation between heart rate and power. This method offers the advantage of being able to estimate FTP based on data collected while the user is running at moderate or higher intensities, without the need for a direct FTP test. However, heart rate varies greatly between individuals and is easily influenced by environmental factors, limiting the accuracy of this method.

[0005] In particular, existing FTP estimation methods rely on fixed testing protocols, often failing to adequately reflect variations in a user's driving patterns, physical condition, training level, and testing environment. For example, short runs at high intensity may result in overestimated FTP, while long runs at low intensity may result in underestimated FTP. To address these issues, a data analysis method that comprehensively considers various factors, such as heart rate, peak power, driving time, and exercise intensity patterns, is needed.

[0006] Research continues to explore FTP estimation methods that analyze multiple peak powers or utilize the heart rate-power correlation. However, these methods suffer from significant individual variability and fail to provide consistent results for all users. Therefore, a new approach is needed that can more accurately reflect diverse user conditions and driving patterns.

[0007] The present invention has been conceived in response to the aforementioned background technology, and is intended to provide a method, device and computer program capable of estimating FTP information in a more accurate and personalized manner by precisely reflecting the user's various driving patterns, physical conditions, training levels, etc.

[0008] The problems to be solved by the present invention are not limited to the problems mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art from the description below.

[0009] In one embodiment of the present invention for solving the above-described problem, a multi-model based FTP estimation method is disclosed. The method includes the steps of acquiring user time series data, selecting one model from among a plurality of FTP estimation models as a user-customized FTP estimation model based on the user time series data, and generating FTP information using the user-customized FTP estimation model. The user time series data includes time-power data and heart rate data generated during the user's driving, and the plurality of FTP estimation models may be models that estimate the FTP information in different ways.

[0010] In an alternative embodiment, the plurality of FTP estimation models may include a first model that estimates FTP based on time series data corresponding to an FTP test protocol, a second model that estimates FTP based on maximum peak power data, and a third model that estimates FTP based on a correlation between heart rate data and power data.

[0011] In an alternative embodiment, the step of selecting one of the plurality of FTP estimation models includes the step of detecting a behavioral intention based on the user time series data by utilizing a classification model, the step of selecting one of the first model and the second model as the user-customized FTP estimation model when the behavioral intention is detected, and the step of selecting the third model as the user-customized FTP estimation model when the behavioral intention is not detected, wherein the behavioral intention may include information related to whether to perform an intentional maximum effort and an FTP test.

[0012] In an alternative embodiment, the classification model may be characterized in that it detects the behavioral intention by analyzing the user time series data with a neural network model pre-trained through a learning data set including time series data, driving data, and corresponding intention information of a plurality of users.

[0013] In an alternative embodiment, when the behavioral intention is detected, the step of selecting one of the first model and the second model as the user-customized FTP estimation model may include the step of selecting the first model as the user-customized FTP estimation model when intention information related to performing the FTP test is detected, and the step of selecting the second model as the user-customized FTP estimation model when intention information related to the intentional maximum effort is detected.

[0014] In an alternative embodiment, the first model may be characterized by calculating an average power corresponding to a specific driving section included in the time series data, identifying a test type according to the driving pattern, identifying a correction factor corresponding to the identified test type, and generating the FTP information based on the calculated average power and the identified correction factor.

[0015] In an alternative embodiment, the second model may be characterized by analyzing the user's weight and driving section data based on the maximum peak power data to derive relationship information between the peak power and weight, generating a power duration curve based on the derived relationship information, and generating the FTP information based on the generated power duration curve.

[0016] In an alternative embodiment, the third model may be characterized by analyzing a correlation between a user's heart rate and power to calculate an average power in a heart rate stable zone, and estimating FTP information based on the calculated average power and the user's maximum heart rate.

[0017] In an alternative embodiment, the output of each of the plurality of models has a priority in the order of the first model, the second model, and the third model, and the method may further include a step of performing an update on the FTP information when FTP information higher than the FTP information of a previous point in time is generated.

[0018] In an alternative embodiment, the method further includes a step of estimating maintenance time information based on the FTP information and body weight data, and a step of correcting the maintenance time information by applying an anaerobic weight based on user information to generate customized maintenance time information, wherein the customized maintenance time information may be characterized as information defining a time for which maximum power can be maintained in response to the user.

[0019] In another embodiment of the present invention, a device for performing a multi-model-based FTP estimation method is disclosed. The device includes a memory storing one or more instructions and a processor executing the one or more instructions stored in the memory, wherein the processor can perform the multi-model-based FTP estimation method by executing the one or more instructions.

[0020] In another embodiment of the present invention, a computer program stored on a computer-readable recording medium capable of performing a multi-model-based FTP estimation method is disclosed. The computer program is coupled to a computer, which is hardware, to perform the multi-model-based FTP estimation method.

[0021] Other specific details of the present invention are included in the detailed description and drawings.

[0022] The present invention provides a natural FTP estimate during routine cycling exercise, enabling a more universal, systematic power-based aerobic exercise management system without the need for a fixed testing protocol. This allows both professional athletes and casual users to develop effective training plans and manage performance based on FTP and time to effort (TTE).

[0023] Furthermore, the present invention overcomes the arduous testing protocols and data bias inherent in existing FTP estimation methods, which primarily target trained athletes. This allows FTP-based aerobic exercise management models, previously limited to elite athletes and enthusiastic cyclists, to be utilized by a wider range of users.

[0024] Additionally, the present invention estimates not only FTP information but also time-to-tenure (TTE) information, allowing the user to predict how long they can maintain their FTP. This can be utilized as a variable highly correlated with a user's resilience, contributing to more efficient performance management during long-distance driving by leveraging both performance and recovery variables.

[0025] The effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description below.

[0026] Various aspects are now described with reference to the drawings, wherein like reference numerals are used to refer to similar components generally. In the following examples, for purposes of explanation, numerous specific details are set forth to provide a comprehensive understanding of one or more aspects. However, it will be apparent that such aspects may be practiced without these specific details.

[0027] FIG. 1 is an exemplary diagram schematically illustrating a system for implementing a multi-model based FTP estimation method related to one embodiment of the present invention.

[0028] FIG. 2 is a hardware configuration diagram of a computing device that performs a multi-model-based FTP estimation method related to one embodiment of the present invention.

[0029] FIG. 3 illustrates a flowchart exemplifying a multi-model based FTP estimation method related to one embodiment of the present invention.

[0030] Figures 4 to 6 are various examples of user time series data related to one embodiment of the present invention.

[0031] FIG. 7 and FIG. 8 are exemplary flowcharts for explaining a process of selecting a user-customized FTP estimation model based on a behavioral intent detection result related to one embodiment of the present invention.

[0032] The advantages and features of the present invention, and the methods for achieving them, will become clearer with reference to the embodiments described in detail below together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below and may be implemented in various different forms. These embodiments are provided solely to ensure that the disclosure of the present invention is complete and to fully inform those skilled in the art of the scope of the present invention, and the present invention is defined solely by the scope of the claims.

[0033] The terminology used herein is for the purpose of describing embodiments only and is not intended to limit the present invention. In this specification, the singular also includes the plural unless specifically stated otherwise. As used herein, the terms "comprises" and / or "comprising" do not exclude the presence or addition of one or more other components in addition to the mentioned components. Like reference numerals refer to like components throughout the specification, and "and / or" includes each and any combination of one or more of the mentioned components. Although "first", "second", etc. are used to describe various components, these components are not limited by these terms. These terms are only used to distinguish one component from another. Therefore, it should be understood that a first component mentioned below may also be a second component within the technical spirit of the present invention.

[0034] Unless otherwise defined, all terms (including technical and scientific terms) used herein may be used in their common sense to those skilled in the art to which the present invention pertains. Furthermore, terms defined in commonly used dictionaries are not to be interpreted ideally or excessively unless explicitly and specifically defined otherwise.

[0035] The term "part" or "module" as used herein refers to a software or hardware component such as an FPGA or ASIC, and the "part" or "module" performs certain functions. However, the "part" or "module" is not limited to software or hardware. The "part" or "module" may be configured to reside on an addressable storage medium and may be configured to execute one or more processors. Thus, by way of example, the "part" or "module" includes components such as software components, object-oriented software components, class components, and task components, as well as processes, functions, attributes, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuitry, data, databases, data structures, tables, arrays, and variables. The functionality provided within the components and "parts" or "modules" may be combined into fewer components and "parts" or "modules" or further separated into additional components and "parts" or "modules."

[0036] In this specification, the term "computer" refers to any type of hardware device including at least one processor, and may also be understood to encompass software components operating on the hardware device, depending on the embodiment. For example, the term "computer" may be understood to encompass, but is not limited to, smartphones, tablet PCs, desktops, laptops, and all user clients and applications running on each device.

[0037] Those skilled in the art should further appreciate that the various illustrative logical blocks, configurations, modules, circuits, means, logics, and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, configurations, means, logics, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. Skilled artisans may implement the described functionality in varying ways for each particular application. However, such implementation decisions should not be interpreted as causing a departure from the scope of the present disclosure.

[0038] Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings.

[0039] Although each step described in this specification is described as being performed by a computer, the subject of each step is not limited thereto, and at least some of each step may be performed by different devices depending on the embodiment.

[0040]

[0041] Figure 1 is an example schematic diagram illustrating a system for implementing a multi-model based FTP estimation method.

[0042] As illustrated in FIG. 1, a system according to embodiments of the present invention may include a computing device (100), a user terminal (200), an external server (300), and a network (400). The components illustrated in FIG. 1 are exemplary, and additional components may exist or some of the components illustrated in FIG. 1 may be omitted. The computing device (100), the external server (300), and the user terminal (200) according to embodiments of the present invention may exchange data for a system according to embodiments of the present invention via the network (400).

[0043] A system according to embodiments of the present invention may include a computing device (100), a user terminal (200), an external server (300), and a network (400). The computing device (100), the user terminal (200), and the external server (300) according to embodiments of the present invention may mutually transmit and receive data for the system according to embodiments of the present invention through the network (400).

[0044] The network (400) according to embodiments of the present invention can use various wired communication systems such as a public switched telephone network (PSTN), xDSL (x Digital Subscriber Line), RADSL (Rate Adaptive DSL), MDSL (Multi Rate DSL), VDSL (Very High Speed ​​DSL), UADSL (Universal Asymmetric DSL), HDSL (High Bit Rate DSL), and a local area network (LAN).

[0045] Additionally, the network presented herein can use various wireless communication systems such as Code Division Multi Access (CDMA), Time Division Multi Access (TDMA), Frequency Division Multi Access (FDMA), Orthogonal Frequency Division Multi Access (OFDMA), Single Carrier-FDMA (SC-FDMA), and other systems.

[0046] The network (400) according to embodiments of the present invention can be configured regardless of the communication mode, such as wired or wireless, and can be configured as various communication networks, such as a personal area network (PAN) and a wide area network (WAN). In addition, the network may be the well-known World Wide Web (WWW), and may also utilize a wireless transmission technology used for short-distance communication, such as infrared (IrDA) or Bluetooth. The technologies described herein can be used not only in the networks mentioned above, but also in other networks.

[0047] According to an embodiment of the present invention, a computing device (100) (hereinafter referred to as 'computing device (100)') that performs a multi-model-based FTP estimation method can detect whether an intentional maximum effort is made based on user time series data, and select an optimal model among a plurality of FTP estimation models to generate user-customized FTP information.

[0048] In the present invention, FTP information can refer to the user's maximum exercise intensity (MLSS, Maximal Lactate Steady State) at which lactic acid levels remain stable, and can be the power an individual can sustain for 20 to 70 minutes. FTP information can be used to evaluate cycling performance or develop training plans, and can serve as a valuable indicator for assessing a user's fitness level and analyzing their ability to maintain sustained exercise intensity. Furthermore, FTP information can be utilized in various sports and fitness-related activities, such as setting training goals and predicting game performance.

[0049] In the present invention, the user time series data relates to time-power data, heart rate data, and exercise intensity changes that occur during the user's driving, and may include, for example, power output and heart rate changes that occur during a 20-minute FTP test, or data during everyday driving. For example, the user's driving refers to exercise activities recorded during outdoor cycling or using an indoor trainer, and includes power output, heart rate changes, driving speed, and distance recorded in real time according to the exercise intensity. Such time series data is collected in real time by an external device (e.g., a power meter, a heart rate monitor, a speedometer, a GPS device, etc.), and the data is transmitted to a computing device (100) to analyze the user's driving pattern, select an appropriate FTP estimation model, and generate user-customized FTP information.

[0050] Specifically, the computing device (100) can collect time-power data and heart rate data generated during the user's driving in real time, detect whether the user is making intentional maximum effort based on the data, and select an optimal model among multiple FTP estimation models based on the detection result to generate user-customized FTP information.

[0051] More specifically, the computing device (100) selects one of a plurality of FTP estimation models that generate FTP information through different methods as a user-customized FTP estimation model, and thereby generates optimal FTP information based on the user's time series data. In an embodiment, the plurality of FTP estimation models, in the present invention, may include a first model according to an FTP test protocol, a second model utilizing maximum peak power data, and a third model utilizing a correlation between heart rate and power.

[0052] The computing device (100) analyzes the user's driving pattern, time-power data, and heart rate data to select the optimal model among multiple FTP estimation models. The computing device (100) utilizes a classification model to detect whether the user's time-series data indicates intentional maximum effort or whether an FTP test has been performed, and selects the most appropriate FTP estimation model based on this.

[0053] Here, the classification model is a pre-trained neural network model that uses multiple user data and learned patterns to determine which model to use when a specific driving pattern is detected. For example, if an intentional FTP test is detected, the first model is selected, if maximum peak power is detected, the second model is selected, and for typical driving patterns, the third model is selected to generate user-tailored FTP estimation information.

[0054] That is, the computing device (100) can provide more accurate and personalized FTP information by selecting one of a plurality of FTP estimation models in a user-customized manner.

[0055] In an embodiment, the present invention implements a multi-model-based FTP estimation method that reflects driving data and individual characteristics of various users to address data bias issues. Conventional FTP estimation methods often result in results that are unsuitable for general users, as they are based on data from trained elite athletes or based on fixed testing protocols. However, the present invention provides a personalized FTP estimation method that reflects the user's actual driving data, heart rate data, and various variables such as weight, age, and gender.

[0056] Specifically, by dynamically selecting the appropriate model for each user through a classification model using multiple FTP estimation models, each user can estimate FTP in a personalized manner, moving beyond the conventional, data-based estimation method. For example, for general users, FTP can be estimated based on the correlation between heart rate and power in everyday running patterns, while for trained athletes, a method based on high-intensity test data can be applied.

[0057] This eliminates data bias tailored only to a specific user group (e.g., a trained user group), and allows general users to obtain accurate FTP information through a model optimized for each individual. In other words, by providing customized FTP estimation to users at different levels, a universal system applicable to all users without data bias can be implemented. In addition, the method provided by the computing device (100) of the present invention has the advantage of enabling FTP information to be naturally estimated through everyday driving data, and reducing dependence on existing fixed test protocols. A detailed description of the multi-model-based FTP estimation method of the present invention and the resulting effects will be described below with reference to FIG. 3.

[0058] Although only one computing device (100) is illustrated in FIG. 1, it will be apparent to those skilled in the art that more servers may also fall within the scope of the present invention and that the computing device (100) may include additional components. That is, the computing device (100) may be comprised of multiple computing devices. In other words, a set of multiple nodes may constitute the computing device (100).

[0059] According to one embodiment of the present invention, the computing device (100) may be a server providing a cloud computing service. More specifically, the computing device (100) may be a server providing a cloud computing service, a type of Internet-based computing that processes information using another computer connected to the Internet rather than the user's computer. The cloud computing service may be a service that stores data on the Internet and allows users to access it anytime and anywhere via an Internet connection without having to install necessary data or programs on their own computers. Furthermore, the data stored on the Internet can be easily shared and transmitted with simple operations and clicks. Furthermore, the cloud computing service may be a service that not only stores data on an Internet server, but also allows users to perform desired tasks using the functions of web-based application programs without having to install separate programs. Furthermore, the cloud computing service may be a service that allows multiple people to simultaneously share and work on documents. Furthermore, the cloud computing service may be implemented in at least one of the following forms: Infrastructure as a Service (IaaS), Platform as a Service (PaaS), Software as a Service (SaaS), a virtual machine-based cloud server, and a container-based cloud server. In other words, the computing device (100) of the present invention may be implemented in at least one of the aforementioned cloud computing services. The specific description of the cloud computing service described above is merely an example, and may include any platform that constructs the cloud computing environment of the present invention.

[0060] According to one embodiment of the present invention, the user terminal (200) is a device that can receive FTP estimation information and exercise performance analysis information from the user through information exchange with the computing device (100), and may refer to a mobile device carried by the user. For example, the user terminal (200) may be a device related to a user who wishes to obtain exercise performance and FTP information. In addition, the user terminal (200) may be a device that can collect driving data and heart rate data in real time by linking with an indoor trainer or bicycle, and may serve to provide exercise-related information to the user.

[0061] The user terminal (200) provides the user with FTP information and exercise performance analysis information acquired from the computing device (100), thereby assisting in establishing a training plan and evaluating driving performance. For example, the user terminal (200) can provide the user with information on peak power and FTP values ​​for each driving section, and can also be utilized as a device that provides personalized feedback based on training performance.

[0062] Additionally, the user terminal (200) can connect to a computing device (100) or an external server (300) to transmit exercise data in real time or receive comprehensive analysis information on exercise results. For example, data such as heart rate changes, time-power data, and exercise intensity can be transmitted to the external server (300) to receive results based on the user's exercise capacity.

[0063] The external server (300) collects users' driving and exercise data and provides comprehensive analysis. It can provide users with exercise performance evaluations, FTP estimations, and time-to-train (TTE) information. For example, the external server (300) compares and analyzes data from various users to calculate more sophisticated correction coefficients, which are then used to support exercise planning and performance evaluation.

[0064] Although the external server (300) and the computing device (100) are depicted as separate entities in FIG. 1, according to an embodiment of the present invention, the computing device (100) may be integrated with the external server (300) to perform FTP estimation, exercise performance analysis, and result provision functions in a single device. In this case, the user terminal (200) can access the external server (300) to analyze its exercise performance and establish a personalized exercise plan.

[0065] Additionally, when the computing device (100) and the external server (300) are separated, the computing device (100) can provide the user's exercise capacity information to the external server (300), and the external server (300) can generate analysis results tailored to the user's exercise performance based on this. In this way, the user terminal (200) can be linked to the user's actual exercise data and receive comprehensive feedback on the user's actual driving performance.

[0066] The external server (300) may be a digital device equipped with a processor, memory, and computing power, such as a laptop computer, notebook computer, desktop computer, web pad, or mobile phone. The external server (300) may be a web server that processes services. The types of servers described above are merely examples, and the present invention is not limited thereto.

[0067]

[0068] FIG. 2 is a hardware configuration diagram of a computing device that performs a multi-model-based FTP estimation method related to one embodiment of the present invention.

[0069] Referring to FIG. 2, a computing device (100) performing a multi-model-based FTP estimation method related to an embodiment of the present invention may include one or more processors (110), a memory (120) for loading a computer program (151) executed by the processor (110), a bus (130), a communication interface (140), and a storage (150) for storing the computer program (151). Here, only components related to the embodiment of the present invention are illustrated in FIG. 2. Therefore, a person skilled in the art to which the present invention pertains may understand that other general components may be further included in addition to the components illustrated in FIG. 2.

[0070] According to one embodiment of the present invention, the processor (110) can typically process the overall operation of the computing device (100). The processor (110) can process signals, data, information, etc. input or output through the components described above, or run an application program stored in the memory (120), thereby providing or processing appropriate information or functions to a user or a user terminal.

[0071] Additionally, the processor (110) may perform operations for at least one application or program for executing a method according to embodiments of the present invention, and the computing device (100) may have one or more processors.

[0072] According to one embodiment of the present invention, the processor (110) may be configured with one or more cores and may include a processor for data analysis and deep learning, such as a central processing unit (CPU), a general purpose graphics processing unit (GPGPU), and a tensor processing unit (TPU) of a computing device.

[0073] The processor (110) can read a computer program stored in the memory (120) and perform data processing for an artificial intelligence model according to an embodiment of the present invention. According to an embodiment of the present invention, the processor (110) can perform operations for neural network learning. The processor (110) can perform calculations for neural network learning, such as processing input data for learning in deep learning (DL), extracting features from input data, calculating errors, and updating the weights of the neural network using backpropagation.

[0074] In addition, the processor (110) can process network function learning using at least one of a CPU, a GPGPU, and a TPU. For example, the CPU and the GPGPU can jointly process network function learning and data classification using the network function. In addition, in one embodiment of the present invention, the processors of multiple computing devices can be used together to process network function learning and data classification using the network function. In addition, the computer program executed in the computing device according to one embodiment of the present invention can be a CPU, GPGPU, or TPU executable program.

[0075] In this specification, the term "network function" may be used interchangeably with "artificial neural network" or "neural network." In this specification, the network function may include one or more neural networks, in which case the output of the network function may be an ensemble of the outputs of one or more neural networks.

[0076] The processor (110) can read a computer program stored in the memory (120) and provide a classification model according to one embodiment of the present invention. According to one embodiment of the present invention, the processor (110) can perform calculations to train the classification model.

[0077] According to one embodiment of the present invention, the processor (110) can typically process the overall operation of the computing device (100). The processor (110) can process signals, data, information, etc. input or output through the components described above, or run an application program stored in the memory (120), thereby providing or processing appropriate information or functions to a user or a user terminal.

[0078] Additionally, the processor (110) may perform operations for at least one application or program for executing a method according to embodiments of the present invention, and the computing device (100) may have one or more processors.

[0079] In various embodiments, the processor (110) may further include a Random Access Memory (RAM, not shown) and a Read-Only Memory (ROM, not shown) that temporarily and / or permanently store signals (or data) processed within the processor (110). In addition, the processor (110) may be implemented in the form of a system on chip (SoC) that includes at least one of a graphics processing unit, RAM, and ROM.

[0080] The memory (120) stores various data, commands, and / or information. The memory (120) can load a computer program (151) from the storage (150) to execute methods / operations according to various embodiments of the present invention. When the computer program (151) is loaded into the memory (120), the processor (110) can perform the method / operation by executing one or more instructions constituting the computer program (151). The memory (120) may be implemented as a volatile memory such as RAM, but the technical scope of the present invention is not limited thereto.

[0081] The bus (130) provides a communication function between components of the computing device (100). The bus (130) may be implemented as various types of buses, such as an address bus, a data bus, and a control bus.

[0082] The communication interface (140) supports wired and wireless Internet communication of the computing device (100). Furthermore, the communication interface (140) may support various communication methods other than Internet communication. To this end, the communication interface (140) may be configured to include a communication module well known in the technical field of the present invention. In some embodiments, the communication interface (140) may be omitted.

[0083] Storage (150) can non-temporarily store a computer program (151). When performing a process for performing a multi-model-based FTP estimation method through a computing device (100), storage (150) can store various information necessary to provide a process for performing the multi-model-based FTP estimation method.

[0084] Storage (150) may be configured to include non-volatile memory such as ROM (Read Only Memory), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), flash memory, a hard disk, a removable disk, or any type of computer-readable recording medium well known in the art to which the present invention pertains.

[0085] The computer program (151) may include one or more instructions that, when loaded into the memory (120), cause the processor (110) to perform a method / operation according to various embodiments of the present invention. That is, the processor (110) may perform the method / operation according to various embodiments of the present invention by executing the one or more instructions.

[0086] In one embodiment, the computer program (151) may include one or more instructions for performing a multi-model based FTP estimation method, including the steps of obtaining user time series data, selecting one of a plurality of FTP estimation models as a user-customized FTP estimation model based on the user time series data, and generating FTP information using the user-customized FTP estimation model.

[0087] The steps of a method or algorithm described in connection with an embodiment of the present invention may be implemented directly in hardware, implemented as a software module executed by hardware, or implemented by a combination thereof. The software module may reside in a random access memory (RAM), a read only memory (ROM), an erasable programmable ROM (EPROM), an electrically erasable programmable ROM (EEPROM), a flash memory, a hard disk, a removable disk, a CD-ROM, or any other form of computer-readable recording medium well known in the art to which the present invention pertains.

[0088] The components of the present invention may be implemented as a program (or application) and stored on a medium to be executed in conjunction with a computer, which is hardware. The components of the present invention may be implemented as software programming or software elements, and similarly, the embodiments may be implemented in a programming or scripting language such as C, C++, Java, assembler, etc., including various algorithms implemented as a combination of data structures, processes, routines, or other programming components. Functional aspects may be implemented as algorithms executed on one or more processors. Hereinafter, a multi-model-based FTP estimation method performed by a computing device (100) will be described in detail with reference to FIGS. 3 to 8.

[0089]

[0090] Figure 3 illustrates a flowchart exemplifying a multi-model-based FTP estimation method according to one embodiment of the present invention. The steps illustrated in Figure 3 may be rearranged as needed, and at least one step may be omitted or added. In other words, the steps illustrated in Figure 3 are merely one embodiment of the present invention, and the scope of the present invention is not limited thereto.

[0091] According to one embodiment of the present invention, a multi-model based FTP estimation method may include a step (S100) of acquiring user time series data.

[0092] In one embodiment, user time-series data may include time-series change data acquired during the user's driving, which may serve as a basis for determining the user's exercise intensity or performance at specific points or intervals during exercise. In one embodiment, the user's exercise patterns, fatigue, or exercise performance may be assessed based on the acquired time-series data.

[0093] User time-series data may include time-power data and heart rate data. However, this data is not limited to these data, and may also include data such as distance traveled, altitude, location, temperature, air resistance (CdA), speed, and pedal rotation rate (cadence). This data can be used to evaluate the user's actual riding performance and condition, and can be used for various exercise performance analyses and personalized FTP estimation.

[0094] According to one embodiment, obtaining user time series data may be by receiving or loading data stored in memory (120). Obtaining user time series data may be by receiving or loading multiple learning data from another storage medium, another computing device, or a separate processing module within the same computing device based on a wired / wireless communication means.

[0095] According to one embodiment, the process of acquiring user time-series data is performed by having external devices collect data in real time and transmit the data to a computing device (100). Here, the external devices may include exercise performance measurement devices such as a power meter, a heart rate monitor, a speedometer, a GPS device, etc., and these external devices measure time-power data, heart rate, speed, distance, pedal rotation speed (cadence), etc. generated during the user's riding in real time. In an embodiment, the power meter measures the power (output) generated when pedaling a bicycle and collects power data exerted by the user during riding in real time. This data indicates the number of watts of force per second and can be used as an important indicator for evaluating the user's exercise intensity. The heart rate monitor measures the user's heart rate in real time and is used to evaluate how much energy the body expends during exercise and the body's response to exercise intensity. This heart rate data is acquired through a heart rate monitor worn on the wrist or in the form of a chest strap. In addition, the speedometer and GPS devices measure the user's riding distance and speed to record exercise performance, and collect location information and route data in real time during riding. The data acquired in this manner is transmitted to a computing device (100) via a wired / wireless communication means, and the computing device (100) analyzes the user's exercise intensity, fatigue, and performance based on the received data, and uses it to select an optimal FTP estimation model.

[0096] According to one embodiment of the present invention, a multi-model based FTP estimation method may include a step (S200) of selecting one model among a plurality of FTP estimation models as a user-customized FTP estimation model based on user time series data.

[0097] In some embodiments, multiple FTP estimation models may be used to generate more precise FTP information by analyzing a user's running patterns, heart rate, and power data from various perspectives. For example, utilizing only a single FTP estimation (or calculation) method may result in inaccurate FTP information that is not suitable for some users or may not reflect various exercise situations.

[0098] For example, FTP estimation methods using a single model are only suitable for specific test protocols or driving situations, making it difficult to accurately estimate FTP based on general driving data. Furthermore, using only heart rate-based estimation methods can reduce the reliability of FTP information, as it varies depending on environmental changes, individual heart rate response patterns, and the reliability of previously stored information such as resting heart rate and maximum heart rate.

[0099] Accordingly, the present invention utilizes multiple FTP estimation models that estimate a user's FTP in different ways. Each of these models processes the user's running pattern, power output, and heart rate data differently, thereby deriving more precise FTP information.

[0100] In other words, multiple FTP estimation models can be characterized as models that analyze user time-series data in different ways and estimate FTP information in different ways. Since each model utilizes different data and calculation methods, the most appropriate model can be selected based on the user's actual driving data and circumstances, providing accurate FTP information.

[0101] In a specific embodiment, the plurality of FTP estimation models may include a first model that estimates FTP based on time series data corresponding to an FTP test protocol, a second model that estimates FTP based on maximum peak power data, and a third model that estimates FTP based on a correlation between heart rate data and power data.

[0102] According to one embodiment of the present invention, the first model may be characterized by calculating an average power corresponding to a specific driving section included in time series data, identifying a test type according to a driving pattern, identifying a correction coefficient corresponding to the identified test type, and generating FTP information based on the calculated average power and the identified correction coefficient.

[0103] In an embodiment, the first model may be a Tested FTP model that estimates a power data time series using an FTP calculation formula when a user performs a test using a traditional FTP protocol.

[0104] Specifically, the first model can generate average power based on time-power data generated during the user's driving. First, the first model calculates the average power for a specific driving section based on the time-series data input when the user performs an FTP test. Then, based on the driving pattern, the model identifies the test type, such as a 20-minute FTP test or a RAMP test, and applies a correction factor based on the identified test type. For example, for a 20-minute FTP test, a correction factor of 0.95 can be applied to calculate the final FTP value. This process compensates for instantaneous power fluctuations that occur during the test and estimates an FTP value based on an exercise intensity standard at which lactate in the body is stably maintained when the user exercises.

[0105] For example, a 20-minute FTP test might have a correction factor of 0.95 based on data from a user riding at their maximum effort for 20 minutes, while a RAMP test might have a correction factor of 0.71 based on a gradually increasing power output.

[0106] Additionally, in an embodiment, the first model can apply a personalized correction factor based on personal information such as the user's fitness level, training data, age, gender, peak power statistics of other users, and training status. For example, the correction factor can be fine-tuned based on the training intensity or fatigue of a specific user to provide a more personalized FTP value. For example, if a user's anaerobic peak power is estimated to be higher than that of other users with similar FTP, a shorter TTE can be applied instead of the general correction factor, taking into account the user's anaerobic energy capacity, to provide a higher FTP estimate. Conversely, if a user's aerobic peak power is estimated to be higher than that of other users with similar FTP, a correction factor can be used to apply a longer TTE, to provide a slightly lower FTP estimate.

[0107] For another example, the correction factor can be differentiated based on physical differences according to the user's age or gender. For example, for elderly users, since their anaerobic energy reserves are relatively low, the correction factor can be lowered and the TTE can be adjusted to provide a lower FTP value. On the other hand, for young users, since their anaerobic energy reserves are high, the TTE can be shortened and a higher FTP value can be estimated. Considering the difference in physical fitness by gender, it is also possible to apply correction factors that increase the TTE and decrease the FTP appropriately for women, considering that women have lower anaerobic capacity and higher aerobic capacity compared to men.

[0108] These personalized corrections contribute to deriving FTP information that is much more tailored to the user's experience than the conventional one-size-fits-all FTP method.

[0109] In other words, the first model is characterized by analyzing power data generated during driving sections based on various FTP test protocols and combining it with correction factors to generate final FTP information. This first model can be advantageous in deriving accurate FTP values ​​in a standardized testing environment.

[0110] According to one embodiment of the present invention, the second model may be characterized by analyzing the user's weight and driving section data based on maximum peak power data to derive relationship information between the peak power and weight, generating a power duration curve based on the derived relationship information, and generating FTP information based on the generated power duration curve.

[0111] In the embodiment, the second model may be a Weighted FTP model, which means a statistical model that applies correction statistics for the maximum peak power and body weight recorded by the user.

[0112] Specifically, the second model analyzes the maximum power generated by the user during a specific period of time using Maximal Peak Power (MMP) data. Maximal Peak Power refers to the maximum power a user can exert during a specific period of time and is utilized as an important factor in evaluating the user's driving performance. In an embodiment, Maximal Peak Power refers to the average power of the maximum effort a user can exert during a specific period of time. For example, if a subject consistently exerts 400 watts through maximal effort for two minutes, the user's two-minute peak power may be recorded as 400 watts. The criteria for determining Maximal Effort can be measured during the time period during which the user maintains the maximum amount of power they can exert.

[0113] The second model generates a power-duration curve based on the correlation between maximum peak power and body weight. The second model can detect that there is no data in which the power drops below a set threshold compared to the existing FTP by applying a moving average corresponding to a set time (e.g., 30 seconds) to the time series data of a specific section based on the maximum peak power. In one embodiment, the second model can detect that there is no data in which the power drops below 0.8% of the existing FTP by applying a 30-second moving average to the time series data of the relevant section. If the time period in which there is no data in which the power drops below 0.8% of the existing FTP continues for a certain period of time (e.g., 4 minutes), it is primarily considered an intentional action. Thereafter, the user's body weight and driving section data are analyzed to derive information on the relationship between peak power and body weight. If the w / kg of the derived specific time period is higher than the w / kg of the same time period in the past, a power-duration curve can be generated based on previously held statistical information.

[0114] A power-duration curve visualizes a user's power output over various time intervals, allowing for the prediction of power decline over riding time. For example, while peak power may be very high for short periods of time, a power-duration curve reflects a pattern where power gradually decreases over longer periods. A power-duration curve allows a user to predict the power they can produce during each riding time interval, thereby providing accurate FTP information.

[0115] For example, the second model generates a power-duration curve based on the time period and body weight data in which the user recorded maximum peak power, thereby predicting the power that the user can exert over various time intervals and calculating FTP. For example, based on the generated power duration curve, an appropriate time for the subject is identified between 20 and 80 minutes, and an estimated power is calculated based on this, which can then be used to generate FTP information. Maximum peak power occurs during short, high-intensity running sessions, and this can be used to estimate the FTP value that an individual can sustain during the TTE time.

[0116] Additionally, the second model can use the power-duration curve to adjust FTP values ​​based on body weight changes. For example, if a user's weight increases relative to peak power during a specific time period, the FTP can be calculated by considering the decrease in aerobic and aerobic power. If the user's weight decreases, a higher FTP can be calculated based on the assumption that aerobic power has improved. In this way, the user's weight relative to the same peak power can be used as a parameter to adjust the slope of the power duration curve by adjusting the aerobic to anaerobic ratio.

[0117] For example, because users in their 20s can exert higher peak power over shorter periods of time than users in their 50s, their power-duration curves will reflect higher peak power output over shorter periods of time (less than 10 minutes). In contrast, older users, even with the same peak power / body mass, will exhibit lower output over shorter periods of time (less than 10 minutes). This subtle adjustment, based on age-specific statistics, allows for more realistic FTP estimates. This approach applies equally to both genders, with women projecting lower peak power estimates over shorter periods of time than men.

[0118] That is, the second model is characterized by generating a power-duration curve based on the correlation between maximum peak power data and body weight, thereby providing FTP information that accurately reflects the user's power change over time.

[0119] According to one embodiment of the present invention, the third model may be characterized by analyzing the correlation between the user's heart rate and power to calculate an average power in a heart rate stable zone, and estimating FTP information based on the calculated average power and the user's maximum heart rate.

[0120] In an embodiment, the third model may be a submaximal FTP model that identifies time-series data intervals where the user runs at a stable heart rate based on a linear relationship between heart rate and power, and estimates FTP using the average power of those intervals. By analyzing power data measured within a stable heart rate range, normalized FTP information is derived based on the power generated by the user between 70% and 90% of their maximum heart rate.

[0121] In one embodiment, the third model, based on Alan Couzens' Submaximal FTP estimation model, analyzes the linear relationship between heart rate (HR) and power to identify sections where a stable HR is achieved, and estimates FTP information based on the average power data for those sections. The existing Submaximal FTP model maintains a stable HR during activities above 70% of the maximum HR, calculates the average power value within that section, and estimates FTP as approximately 87% of the maximum HR.

[0122] However, the existing submaximal model had a problem in that the FTP value was over- or under-measured due to the measurement of resting heart rate and maximum heart rate, which have large individual differences depending on activity tendencies, unstable wearing of the heart rate monitor, and environmental factors such as temperature or the user's condition, which caused delay or incorrect measurement of the heart rate value. To complement this, the third model of the present invention defines the resting heart rate as 55% of the maximum heart rate when it is unclear, and defines the maximum heart rate as the maximum heart rate directly set by the user as a top priority, but defines 180 as the reference value when there is no user-defined maximum heart rate. In the driving data, a method is adopted in which the stable heart rate value is not used for FTP estimation if it is less than 70% of the maximum heart rate.

[0123] In one embodiment, the third model can search for a section in which the heart rate is 70% or higher of the maximum heart rate and the power does not fall below 0.8% of the user's FTP by performing a moving average process corresponding to a set time (30 seconds or 1 minute) for the user's heart rate and power, and in each moving averaged data, there is no data. The third model can analyze the correlation between the power and the heart rate in the identified section to calculate the maximum power matching the user's maximum heart rate, and multiply the calculated maximum power by a coefficient between 0.86 and 0.88 to generate an FTP correction.

[0124] In this example, power data is normalized to a 1-minute moving average to compensate for temporary fluctuations and remove device errors and environmental factors before deriving an FTP value. This preprocessing process ensures more reliable data, and enables accurate FTP estimation by utilizing peak power and corresponding heart rate data over a specific period.

[0125] For example, power data generated based on heart rate stability between 70% and 90% of maximum heart rate can be analyzed to estimate FTP based on the average power within that range. Applying a 1-minute moving average of power data minimizes data instability due to heart rate fluctuations, enabling accurate FTP estimation.

[0126] Additionally, to compensate for measurement errors that may arise from the instability of wearing a heart rate monitor, normalized power data is used to calculate FTP values ​​only for data intervals where moderate-intensity efforts exceed 70% of maximum heart rate. This prevents overestimation issues that arise in existing submaximal FTP models.

[0127] Therefore, the third model analyzes the correlation between heart rate and power, and provides personalized FTP information based on the average power in the heart rate zone maintained at a stable moderate intensity or higher, enabling realistic and reliable FTP estimation for users regardless of the success or failure of high-intensity tests and FTP tests.

[0128] As described above, the computing device (100) of the present invention can utilize multiple FTP estimation models to provide customized FTP information to a user. For example, when a user performs a high-intensity FTP test, the first model is selected, and when the user records maximum peak power, the second model is selected. Conversely, in everyday driving situations where FTP must be estimated through the correlation between heart rate data and power without performing a maximum-intensity test, the third model is selected.

[0129] As a specific example, Fig. 4 illustrates a high-intensity FTP test situation. The X-axis represents the driving time, the Y-axis represents the power output (watts) and the heart rate (bpm), and different colors distinguish the high-intensity and low-intensity sections. Fig. 4 visually shows the fluctuations in power data and heart rate data that occurred during the drive. For example, in the high-intensity section, a pattern can be confirmed in which the power output increases rapidly and the heart rate also rises accordingly. In such sections, the first model is applied to estimate the user's FTP information.

[0130] Figure 5 also illustrates intermittent interval running, which may be a suitable case for the third model. The X-axis shows distance traveled (km), and the Y-axis shows power output (watts) and heart rate (bpm). Figure 5 visualizes a case where a user exhibits irregular running patterns and the correlation between heart rate and power output is not consistent. In such cases, since a fixed testing protocol is not applicable, the third model can be selected to analyze the relationship between heart rate and power to derive FTP information.

[0131] Thus, Models 1 and 2 are utilized in situations where a clear test protocol has been performed or where high-intensity peak power data has been recorded. For example, if a user performs a 20-minute FTP test or records maximum peak power, FTP information is derived using Models 1 and 2. Conversely, Model 3 analyzes data generated during everyday driving in real time and dynamically calculates FTP values ​​based on driving patterns and heart rate changes. This allows for reliable FTP values ​​based on various data collected during driving, even when the user does not perform a clear test.

[0132] The computing device (100) statistically analyzes the user's accumulated time-power time series data to determine whether there has actually been an intentional, maximum-effort peak power or FTP test, and then dynamically selects the most appropriate FTP estimation model based on the success or failure of the attempt. Through this process, accurate FTP information tailored to the user's individual circumstances can be provided.

[0133] That is, the computing device (100) can determine whether the user is actually exerting effort by selecting an FTP estimation model suitable for each situation based on the user's driving data and heart rate data, and can generate optimized FTP information accordingly, so that not only professional athletes but also general users can easily obtain suitable FTP information, and there is an advantage of providing a reliable FTP value through everyday driving data without a high-intensity test.

[0134] This allows us to provide accurate and customized FTP estimates to a wide range of users, minimizing data bias without relying on existing, fixed testing protocols.

[0135] Additionally, in an embodiment, the computing device (100) can utilize a classification model to analyze the user's driving pattern and heart rate data to detect whether there was intentional maximum effort. Referring to FIG. 6, an area where the heart rate and power output suddenly increased in a specific section during the user's driving can be visualized. FIG. 6 shows the distance traveled (km) on the X-axis, the power output (watt) and the heart rate (bpm) on the Y-axis, and shows the sudden increase in heart rate and power output in a high-intensity section. For example, in a section where the heart rate reached 169 bpm and 80% of the maximum heart rate, the classification model can detect that the user intentionally continued high-intensity effort.

[0136] If intentional effort is detected, the classification model selects either the first or second model to estimate FTP information based on whether the user performed a high-intensity FTP test or recorded maximum peak power. Conversely, irregular sections in Figure 6, where the relationship between heart rate and power remains relatively unstable, may be detected as unintentional driving patterns. In this case, the third model is selected to dynamically estimate FTP values ​​based on heart rate and power data.

[0137] As a result, the computing device (100) selects a model suitable for each driving situation through a classification model that distinguishes whether or not intentional effort is made, thereby providing optimal FTP information, thereby enabling the user to estimate an accurate FTP value through everyday driving data without relying on a high-intensity test protocol.

[0138] In a specific embodiment, referring to FIG. 7, the step of selecting one model from among a plurality of FTP estimation models may include a step of detecting behavioral intent based on user time series data by utilizing a classification model (S210), a step of selecting one of the first and second models as the user-customized FTP estimation model when the behavioral intent is detected (S220), and a step of selecting the third model as the user-customized FTP estimation model when the behavioral intent is not detected (S230).

[0139] Here, behavioral intention may include information regarding whether to perform intentional maximal effort and FTP tests.

[0140] In one embodiment, the classification model may be characterized by detecting the behavioral intention by analyzing the user time series data with a neural network model pre-trained through a learning data set including time series data, driving data, and corresponding intent information of a plurality of users.

[0141] In this embodiment, the classification model is a machine learning-based neural network model that can be pre-trained using a training data set containing time-series data, driving data, and intent information from multiple users. This classification model analyzes user time-series data and heart rate data to detect whether the user is performing an intentional maximum effort or FTP test. During the training process, the classification model develops the ability to detect behavioral intent based on the driving patterns and data of various users.

[0142] For example, when performing an intentional FTP test, the time-series data may detect a steady, stepwise increase in power output of 20 watts per minute, or a sustained, high power output for approximately 20 minutes. If maximum peak power is recorded, a rapid increase in power and corresponding heart rate changes over a period of four minutes or more are observed. These patterns are repeatedly analyzed during the model training process and used as criteria for the classification model to detect behavioral intent.

[0143] Additionally, in this embodiment, the classification model analyzes the correlation between driving patterns and heart rate patterns based on a pre-trained neural network, thereby determining whether the user has intentionally exerted maximum effort. For example, if a pattern of rapid increases in power output and concomitant increases in heart rate are observed in the user's time-series data during a specific period, an FTP test performance is detected. Conversely, if the changes in power output are inconsistent and the heart rate changes are minimal, the classification model may recognize this as routine driving and select the third model.

[0144] As a result, the classification model can provide accurate FTP information for each situation by selecting the optimal model for user-tailored FTP estimation.

[0145] More specifically, the computing device (100) analyzes user time series data and heart rate data in step S210 to detect the user's behavioral intention. The classification model analyzes the correlation between driving patterns and heart rate patterns based on a pre-trained neural network, thereby determining whether the user exerted intentional maximum effort or performed an FTP test.

[0146] For example, if a user's time series data shows a pattern of sustained high power output and high heart rate for a specific period of approximately 20 minutes, an FTP test is detected. Conversely, if power output varies inconsistently and heart rate changes are minimal, this is considered routine running, and the third model is selected.

[0147] In step S220, when an action intention is detected, the computing device (100) can select one of the first model and the second model as a user-customized FTP estimation model.

[0148] More specifically, referring to FIG. 8, when an action intention is detected, the step (S220) of selecting one of the first model and the second model as the user-customized FTP estimation model may include a step (S221) of selecting the first model as the user-customized FTP estimation model when intention information related to performing an FTP test is detected, and a step (S222) of selecting the second model as the user-customized FTP estimation model when intention information related to intentional maximum effort is detected.

[0149] To elaborate, if intent information related to performing an FTP test is detected, the first model is selected. At this time, the first model estimates FTP information based on data from the user performing a standard FTP test, applying a correction factor corresponding to the specific test protocol. For example, for a 20-minute FTP test, a correction factor of 0.95 can be applied to calculate the FTP value.

[0150] Meanwhile, if intentional information related to intentional maximal effort is detected, a second model is selected. This model generates a power-duration curve based on peak power data and body weight, and uses this to estimate FTP information. This method relies on short-term peak power data and takes into account the user's various physical conditions to provide a more appropriate FTP value.

[0151] Meanwhile, if no behavioral intent is detected in step S230, the computing device (100) may select the third model as a customized FTP estimation model. The third model analyzes the correlation between heart rate and power and estimates FTP information based on the user's daily driving data. This model calculates FTP values ​​based on power data generated in the 70-90% range of the maximum heart rate, thereby enabling accurate FTP information to be obtained even in daily driving environments.

[0152] As described above, the computing device (100) utilizes a classification model to comprehensively analyze the user's real-time driving patterns and heart rate data to dynamically select an optimal FTP estimation model. This allows for optimal FTP information tailored to each situation, enabling personalized FTP estimation. If a user's intentional behavior is detected, an appropriate model can be selected to derive accurate FTP information, reducing reliance on fixed testing protocols. Consequently, the method offers the advantage of being suitable for both professional athletes and general users, providing a flexible FTP estimation method suitable for users of all skill levels.

[0153] According to one embodiment of the present invention, a multi-model-based FTP estimation method may include a step (S300) of generating FTP information using a user-customized FTP estimation model. A computing device (100) processes the user's time series data and heart rate data based on the FTP estimation model (i.e., the user-customized FTP estimation model) selected using a classification model to produce optimal FTP information. The computing device (100) provides personalized FTP information that reflects each user's driving pattern and physical condition, thereby producing and providing accurate and reliable FTP values ​​not only in everyday driving but also in intentional test situations.

[0154] In one embodiment, the outputs of each of the multiple models may be prioritized in the order of Model 1, Model 2, and Model 3. This prioritization is intended to provide users with the most reliable results based on the accuracy and application context of each model.

[0155] This configuration is necessary to more accurately reflect high-intensity data generated when a user intentionally exerts maximum effort or during a test situation. Model 1 is preferred because it provides the most reliable FTP estimation based on a deliberately performed FTP test protocol. Model 2 reflects maximal effort recorded over a short period of time based on the user's maximum peak power, effectively estimating FTP even with shorter data periods compared to test situations. Conversely, Model 3 utilizes the correlation between heart rate and power observed during everyday driving to provide reliable FTP information in normal situations.

[0156] The reason for prioritizing may be that each model is optimized based on data collected in different situations. For example, if a user performs a 20-minute FTP test, Model 1 may provide the most accurate results, but if the user records their maximum peak power for a short period or on the field, Model 2 may be more appropriate. Model 3 is advantageous for routine training or situations requiring heart rate-based estimation.

[0157] The advantage of this prioritization is that it provides users with optimal FTP information tailored to their specific situation. This allows users to automatically obtain appropriate FTP information during routine training and to receive accurate results after performing intentional FTP tests. Furthermore, even when users only exert maximal effort for a short period of time, lasting four minutes or more, the second model can quickly analyze the results.

[0158] As a result, the computing device (100) can provide optimal FTP estimation results in various driving situations by utilizing the advantages and characteristics of each model according to priority, which provides the effect of implementing a personalized FTP estimation system suitable for not only professional athletes but also general users.

[0159] In addition, the computing device (100) can dynamically update FTP information based on data collected in real time, and can contribute to establishing an efficient exercise plan by providing personalized feedback considering the user's training intensity and fatigue level.

[0160] In one embodiment, a multi-model-based FTP estimation method may include a step of updating FTP information only when FTP information higher than that of a previous point in time is generated. This configuration ensures that updates are performed only when performance improvements are clearly demonstrated, thereby preventing FTP from being overestimated due to incorrect data.

[0161] More specifically, the computing device (100) analyzes time-series data collected in real time during the user's driving, and updates the FTP information when the user exerts intentional effort or improves training performance. For example, if the user records a new peak power or if long-term training performance accumulates and exceeds the existing FTP, the previous FTP value is updated with the new FTP information. Conversely, if performance deteriorates due to fatigue during training or if there is a temporary data error, the existing FTP information is maintained.

[0162] This update method is designed to accurately reflect a user's physical fitness changes while not overly reflecting performance declines due to temporary deterioration in condition. This allows for stable performance management and provides a system that promotes sustained performance improvement over the long term.

[0163] Additionally, in an embodiment, a multi-model-based FTP estimation method may further include a step of estimating time-to-exhaustion (TTE) information based on FTP information and body weight data, and a step of correcting the time-to-exhaustion information by applying an anaerobic weight based on user information to generate customized time-to-exhaustion information. Here, the customized time-to-exhaustion information may be characterized as information defining a time for which a user can maintain maximum power.

[0164] More specifically, the computing device (100) can calculate Time to Exhaustion (TTE) information based on the user's FTP information and body weight data. The time to exhaustion (TTE) is defined as the time a user can maintain FTP. While this is typically close to one hour for athletes, it can vary significantly for the general public or those with less training. Therefore, accurately estimating FTP and TTE, the time required to maintain FTP, is crucial. In an embodiment of the present invention, in addition to existing FTP information, customized TTE information is provided to enhance training performance.

[0165] In the process of estimating TTE, the computing device (100) may apply a formula for calculating TTE based on the user's body weight and FTP value. For example, TTE may be calculated in the form of TTE = (W / kg for one hour) × 10 + 5. This calculation provides the user with information on how long he or she can continue exercising at a certain intensity.

[0166] Additionally, anaerobic weighting based on user information such as age and gender is applied to TTE, resulting in accurate retention times tailored to individual physical characteristics. For example, if age is set at 40, a linear calculation can be applied so that a 25-year-old user's TTE is 10% earlier, and a 55-year-old's is 10% longer. Depending on gender, weighting can be applied so that women's TTE is approximately 10% higher. These weightings provide customized TTE information that reflects physical differences by age and gender.

[0167] As a specific example, if a 30-year-old male weighing 65 kg has an FTP of 260 watts, the TTE can be calculated using the following formula:

[0168] TTE = ((260 / 65) × 10 + 5) × 0.9333

[0169] This yields a TTE of approximately 42 minutes, meaning that this user can maintain his FTP for approximately 42 minutes.

[0170] In conclusion, the computing device (100) calculates personalized TTE information based on the user's weight, age, gender, and FTP value, which can play an important role in establishing a training plan and managing performance.

[0171] In summary, the computing device (100) can comprehensively analyze the user's weight, age, gender, FTP value, and previous driving data to provide customized and optimized TTE (Time to Exhaustion) information to each user. This allows the user to establish a training plan tailored to their physical characteristics and efficiently manage performance based on accurate information on how long a given power can be maintained. Furthermore, the computing device (100) collects and analyzes data in real time, dynamically updating the TTE according to changes in the user's physical condition and training intensity, thereby helping the user set and achieve customized training goals.

[0172]

[0173] The steps of a method or algorithm described in connection with an embodiment of the present invention may be implemented directly in hardware, implemented as a software module executed by hardware, or implemented by a combination thereof. The software module may reside in a random access memory (RAM), a read only memory (ROM), an erasable programmable ROM (EPROM), an electrically erasable programmable ROM (EEPROM), a flash memory, a hard disk, a removable disk, a CD-ROM, or any other form of computer-readable recording medium well known in the art to which the present invention pertains.

[0174] The components of the present invention may be implemented as programs (or applications) and stored on a medium to be executed in conjunction with a computer, which is hardware. The components of the present invention may be implemented as software programs or software elements. Similarly, the embodiments may be implemented in a programming or scripting language such as C, C++, Java, or an assembler, including various algorithms implemented as a combination of data structures, processes, routines, or other programming components. Functional aspects may be implemented as algorithms that are executed on one or more processors.

[0175] Those skilled in the art will appreciate that the various illustrative logical blocks, modules, processors, means, circuits, and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, various forms of programs or design code (referred to herein, for convenience, as “software”), or a combination of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. Those skilled in the art may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.

[0176] The various embodiments presented herein can be implemented as a method, apparatus, or article of manufacture using standard programming and / or engineering techniques. The term "article of manufacture" includes a computer program, carrier, or media accessible from any computer-readable device. For example, computer-readable media include, but are not limited to, magnetic storage devices (e.g., hard disks, floppy disks, magnetic strips, etc.), optical disks (e.g., CDs, DVDs, etc.), smart cards, and flash memory devices (e.g., EEPROMs, cards, sticks, key drives, etc.). Furthermore, the various storage media presented herein include one or more devices and / or other machine-readable media for storing information. The term "machine-readable medium" includes, but is not limited to, wireless channels and various other media capable of storing, retaining, and / or carrying instructions and / or data.

[0177] It should be understood that the specific order or hierarchy of steps in the presented processes is merely an example of exemplary approaches. It should be understood that the specific order or hierarchy of steps in the processes may be rearranged within the scope of the present invention based on design priorities. The appended method claims provide elements of various steps in a sample order, but are not intended to be limited to the specific order or hierarchy presented.

[0178] The description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other embodiments without departing from the scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments disclosed herein, but is to be construed in the broadest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method performed on one or more processors of a computing device, Step of acquiring user time series data; A step of selecting one model among multiple FTP estimation models as a user-customized FTP estimation model based on the above user time series data; and A step of generating FTP information by utilizing the above user-customized FTP estimation model; including; The above user time series data includes time-power data and heart rate data generated during the user's driving. The above multiple FTP estimation models are models that estimate the FTP information in different ways. A multi-model based FTP estimation method.

2. In paragraph 1, The above multiple FTP estimation models are, A first model for estimating FTP based on time series data corresponding to the FTP test protocol; A second model that estimates FTP based on maximum peak power data; and A third model that estimates FTP based on the correlation between heart rate data and power data; including; A multi-model based FTP estimation method.

3. In paragraph 2, The step of selecting one model among the above multiple FTP estimation models is: A step of detecting behavioral intention based on the user time series data by utilizing a classification model; When the above-mentioned behavioral intention is detected, a step of selecting one of the first model and the second model as the user-customized FTP estimation model; If the above-mentioned behavioral intention is not detected, a step of selecting the third model as the user-customized FTP estimation model is included; The above behavioral intention includes information regarding whether to perform intentional maximum effort and FTP tests. A multi-model based FTP estimation method.

4. In paragraph 3, The above classification model is, A pre-trained neural network model using a training data set containing time series data, driving data, and corresponding intent information of multiple users. Characterized in that the above-mentioned user time series data and heart rate data are analyzed to detect the above-mentioned behavioral intention. A multi-model based FTP estimation method.

5. In paragraph 3, When the above behavioral intention is detected, the step of selecting one of the first model and the second model as the user-customized FTP estimation model is as follows: When detecting intention information related to the performance of the above FTP test, a step of selecting the first model as the user-customized FTP estimation model; and Including a step of selecting a second model as the user-customized FTP estimation model when detecting intention information related to the above intentional maximum effort; A multi-model based FTP estimation method.

6. In paragraph 2, The above first model is, A method characterized in that it calculates an average power corresponding to a specific driving section included in the above time series data, identifies a test type according to a driving pattern, identifies a correction coefficient corresponding to the identified test type, and generates the FTP information based on the calculated average power and the identified correction coefficient. A multi-model based FTP estimation method.

7. In paragraph 2, The above second model is, Based on the maximum peak power data, the user's weight and driving section data are analyzed to derive relationship information between the peak power and weight, a power duration curve is generated based on the derived relationship information, and the FTP information is generated based on the generated power duration curve. A multi-model based FTP estimation method.

8. In paragraph 2, The third model above is, A method characterized in that the correlation between the user's heart rate and power is analyzed to calculate the average power in the heart rate stable zone, and FTP information is estimated based on the calculated average power and the user's maximum heart rate. A multi-model based FTP estimation method.

9. In paragraph 2, The output of each of the above multiple models is: The priority is in the order of the first model, the second model, and the third model. The above method, If FTP information higher than the FTP information of the previous point in time is generated, a step of performing an update on the FTP information is further included; A multi-model based FTP estimation method.

10. In paragraph 1, The above method, A step of estimating maintenance time information based on the above FTP information and weight data; and A step of generating customized retention time information by correcting the retention time information by applying an anaerobic weight based on user information; further comprising; The above custom maintenance time information is characterized by being information defining the time for which maximum power can be maintained in response to the user. A multi-model based FTP estimation method.

11. Memory that stores one or more instructions; and A processor comprising: a processor that executes one or more instructions stored in the memory; The processor executes one or more of the instructions, A device for performing the method of claim 1.

12. A computer program stored on a computer-readable recording medium that is combined with a computer as hardware and can perform the method of claim 1.

Images