Apparatus, method and computer program product for providing training guidance

Abstract

An apparatus, method, and computer program product for providing training guidance. The method comprises: obtaining (220) training data representing a power output of a plurality of athletic performance of the user (100), the training data includes first training data representing a power output of a first athletic performance for a first duration and second training data representing a power output of a second athletic performance for a second duration different from the first duration; fitting (222) the training data into an attenuated performance power curve representing a power output as a function of a duration of continuous athletic performance; determining (224), based on the fitting, a set of training power output regions conforming to the performance power curve by decaying according to duration; generating (226) a training plan for physical exercise based on the set of training power output regions; and causing (228) a user interface to output the training plan to the user.

Description

Technical Field

[0001] Various embodiments are directed to apparatus, methods, and computer program products for providing training guidance. Background Art

[0002] Sports performance testing is traditionally based on laboratory testing of athletes. However, laboratory testing is time consuming and expensive, and the results may not translate well to real-world performance outside the laboratory. Laboratory testing can be used to find some user parameters, such as maximum oxygen uptake for determining the user's aerobic threshold and anaerobic threshold. These thresholds are then used to determine personalized heart rate zones that guide the user's training. Alternatively, other training intensities and corresponding zones, such as speed, can be used. Exercise can be based on exercising with a certain training intensity profile for a certain duration.

[0003] A mathematical model describing running performance was recently studied in Mullig et al. (2018), A minimal power model for human running performance, PLoS ONE 13 (11): e0206645. Compared with data collected in a laboratory environment, the model can utilize real-world data, such as world record times. Further research on this model was conducted in Emig & Peltonen, Human running performance from real-world big data, Nat Commun 11, 4936 (2020). However, the practical application of this model is currently very limited and needs further development. Summary of the invention

[0004] According to an aspect, the subject matter of the independent claims is provided. The dependent claims define some embodiments.

[0005] One or more examples of implementations are set forth in more detail in the drawings and the description of the embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0006] Some embodiments will now be described with reference to the accompanying drawings, in which

[0007] Figure 1 An embodiment of an apparatus for providing training guidance is illustrated;

[0008] Figure 2A and Figure 2B An embodiment of a method for providing training guidance is illustrated;

[0009] Figure 3 and Figure 4An example embodiment of a representation power curve is illustrated;

[0010] Figure 5 and Figure 6 illustrating embodiments related to providing training guidance;

[0011] Figure 7 and Figure 8 Graphs illustrating embodiments related to fitting performance power curves;

[0012] Fig. 9 and Fig.10 An embodiment related to acquisition of time series data is illustrated;

[0013] Fig.11 illustrates an embodiment for using and updating a performance power curve; and

[0014] Fig.12 An embodiment of an apparatus for providing training guidance is illustrated as a block diagram. DETAILED DESCRIPTION

[0015] The following embodiments are merely examples. Although the specification may refer to "an" embodiment in several places, this does not necessarily mean that each such reference refers to the same embodiment, or that the feature applies only to a single embodiment. Individual features of different embodiments may also be combined to provide other embodiments. In addition, the words "comprise" and "include" should be understood as not limiting the described embodiments to consist only of those features that have been mentioned, and such embodiments may also include features / structures that are not specifically mentioned.

[0016] In the description of the embodiments and in the claims, reference numerals are used to illustrate the embodiments with reference to the figures and are not intended to limit them to these examples alone.

[0017] Embodiments and features disclosed in the following description that do not belong to the scope of the independent claims, if any, should be construed as examples to assist in understanding various embodiments of the present invention.

[0018] Figure 1 The diagram illustrates an embodiment of an apparatus for providing training guidance to a user 100. The user 100 is shown performing a physical exercise or athletic performance. The physical exercise may be running (such as Figure 1 Power output may refer to the power generated or output by a user during athletic performance.

[0019] Figure 1The apparatus may be or may include a wearable device 102, a portable electronic device 104, a wearable sensor device 106, and / or a server computer 108. The apparatus may be a single physical apparatus or a distributed apparatus (system). For example, the wearable device 102 may operate as a single physical apparatus. A distributed apparatus or system may include multiple communicatively coupled physical devices, such as the wearable device 102, the wearable sensor device 106, the portable electronic device 104, and / or the server computer 108.

[0020] The device may be configured to perform Figure 2A The computer-implemented method shown in . Figure 2A , the method includes: acquiring 220 training data representing power output of multiple sports performances of a user 100, the training data including first training data representing power output of a first sports performance of a first duration and second training data representing power output of a second sports performance of a second duration different from the first duration; fitting 222 the training data into a decaying performance power curve, the performance power curve representing power output as a function of duration of continuous sports performance; based on the fitting, determining 224 a set of training power output areas that conform to the performance power curve by decaying according to duration; generating 226 a training plan for physical exercise based on the set of training power output areas; and causing a user interface to output the training plan to a user.

[0021] The fitting and calculation of the resulting apparent power curve may be performed as described in Mulligan et al mentioned in the background. In the following, some embodiments of the calculation of the apparent power curve are described.

[0022] Thus, a performance power curve may define the shape of a training power output zone.Performance power curves and training power output zones may be defined based on average power as a function of duration of athletic performance.A workout may refer to a major work session of a physical exercise or competition.

[0023] Figure 2BAnother aspect of a method for improved training guidance is illustrated. The method includes: acquiring 200 time series data representing power output of a user's athletic performance; extracting 202 a plurality of segments from the time series data, wherein the plurality of segments include a first segment having a first duration and a second segment having a second duration, the second duration being different from the first duration; determining 204 segment power values ​​based on the extracted time series data of the plurality of segments; forming 206 a set of scattered points describing how the segment power values ​​vary according to the respective durations of the extracted plurality of segments, and fitting 208 a performance power curve to the set of scattered points; determining 210 a set of training zones based on the performance power curve; generating 212 training instructions based on the set of training zones; and causing 214 a user interface to output the training instructions.

[0024] The solution described herein provides personalized training guidance to users. The performance power curve and the training zones derived therefrom characterize the user's athletic performance. By generating training instructions based on a set of training zones, users can receive personalized feedback and guidance on their training. The performance power curve is intended to present how well the user is able to maintain a certain average power output for a certain duration. Figure 3 As shown in the performance power curve, when the duration of the sports performance is short, the user is able to maintain a greater average power output. It is well known that people can only maintain a higher constant rate in a short duration, and can maintain a slower rate in a long duration. However, this feature has not been introduced into the training area in a personalized manner as performed by the present invention. The training area and the training power output area refer to the same area in this specification. The training area refers to the area defined by the upper and lower limits defined according to the power output of the user. As described herein, the power output can be quantified as, for example, heart rate or rate or speed (velocity). The time series data represents the power output of multiple sports performances performed by the user, each of which has a specific duration (t). Sports performance can be understood as a single physical exercise, a game or a competition, or a part of a physical exercise. The power output of the sports performance can be directly measured by, for example, a power sensor commonly used with a (fixed) bicycle. In one embodiment, the time series data is power data measured by a power sensor. Power output can also be derived from other types of time series data. It has been found that there is a linear relationship between power, speed and heart rate. These linear relationships can be used to convert, for example, speed data and / or heart rate data into power data, and vice versa. Furthermore, the device may derive speed data, for example from position data measured by a position sensor, and / or the device may derive heart rate data, for example from electrocardiogram (ECG) data measured by an ECG sensor.

[0025] The time series data may be time series data of past athletic performances acquired from one or more memories, and / or time series data acquired in real time by the device. The device may include or be coupled to one or more sensors configured to measure the time series data, and the device may be configured to acquire the time series data from the one or more sensors. The acquisition may also include receiving the time series data from the one or more sensors or one or more memories via a wired or wireless connection.

[0026] Time series data is an example of training data.

[0027] The training data may be a collection of samples, where each sample has a power value representing a power output and a duration value representing a duration of a corresponding athletic performance. Figure 3 Such a group of samples (hereinafter referred to as scatter points) is illustrated. The power value can be the average rate of counting within the corresponding duration. In the simplest form, the device can obtain corresponding information as user input from a user interface or from a memory to generate the group of samples. Such information required for each sample can include, for example, the distance of the motion performance and the duration required to cross the distance. Then, the average rate can be calculated by the ratio of distance and duration to obtain the power value. In order to correctly fit, at least two samples of the motion performance with different durations are required. Preferably, more than two samples are obtained under a fairly wide duration span to improve the accuracy of the fit.

[0028] From another perspective, the training data may include time series data. Time series data may be understood as a collection of measurement data measured during one or more athletic performances. The training data may include measurement data of at least one athletic performance stored in at least one memory, and at least one processor is configured to calculate the power output of the athletic performance based on the measurement data and use the power output as training data.

[0029] In one embodiment, the device is configured to extract 202 multiple segments from the time series data. The multiple segments include a first segment having a first duration and a second segment having a second duration, and the second duration is different from the first duration. The segment can be a part of a physical exercise, a game or a physical exercise. Preferably, the segments are taken from different exercises or games. In principle, at least two different durations may be required to fit the performance power curve. Alternatively or additionally, the duration of the segment can be checked via the distance traveled during the segment. For example, when the time series data includes position data, the device can determine the distance traveled by the user based on the position data. The device can select the segment to be extracted based on the assumption that different distances are usually traveled in different durations. Therefore, the first segment can have a first distance, and the second segment can have a second distance, and the second distance is different from the first distance.

[0030] The performance power curve and the associated training area are specially designed for running, and are used to guide users in running exercises. Therefore, the training data can include or consist of training data from past running exercises or running competitions. However, the same principle can be directly applied to other sports, such as race walking, which is very similar to running. Considering the metric of the average training power output target determined based on the training area (the training area selected for exercise), the same principle can also be applied to other sports, such as riding a bicycle, swimming or cross-country skiing. For example, rate is a training power output target metric suitable for running, but is not suitable for riding a bicycle or skiing. When riding a bicycle, pedaling power may be a suitable metric. In any case, the training data is preferably selected only from the sports type (such as running) of exactly the same type. It can be achieved that different sports types are mixed into the same performance power curve, but it may easily lead to an inaccurate estimate of the ability of the user.

[0031] In one embodiment, training data represents the maximum power output of the user in the corresponding sports performance (such as exercise or competition). The example of maximum power output includes the best performance (recording average power output) of the user in the sports performance. Therefore, training data can include or consist of the user's record time and corresponding distance (such as 5km record, 10km record, half marathon record and marathon record) for various games / exercises of different distances. These examples produce a performance power curve based on the maximum physiological ability of the user. Another alternative is to use the training data representing the maximum level of the user's desired exercise, representing the maximum power output of comfort. The user can manually input this information to the device, or the device can harvest this information from the past exercise of the user stored in the memory of the device. In this case, the performance power curve is adapted to the level that the user wishes to train. In order to select suitable training data, the user can indicate whether the performance power curve should be based on the maximum physiological ability (record) of the user or based on the comfortable maximum performance of the user. The user can personally select or approve the training data to be included in the fitting.

[0032] The device may apply some filters to decide whether particular information qualifies as training data. For example, a maximum performance that qualifies as a performance power curve must be no older than a specified time period, e.g., only maximum performances achieved within a year qualify. In the event that a sufficient amount of recent training data is not available, the device may notify the user of this and request that the user provide more training data (e.g., exercise for a certain distance). Alternatively, the device may notify the user that the fit may be inaccurate because the training data is old.

[0033] As described herein, even a portion of a past athletic performance may be considered eligible as training data. For example, if a running workout includes a work period in which the user achieves the best maximum performance over a certain distance and during a work period that forms a subset of the work periods for the running workout, the distance and duration of the work period may be counted as training data. In such an embodiment, the device may identify a continuous time interval of athletic performance associated with a power output greater than an inclusion threshold (maximum performance), and select the duration and power output of the continuous time interval for fitting, and exclude the power output of at least one other time interval of the first athletic performance from the fitting. The inclusion threshold may be set so that it represents the user's best performance in a particular workout (exercise competition / workout at a particular distance).

[0034] The apparatus is configured to determine 204 a segment power value based on the extracted time series data of the plurality of segments. In one embodiment, the apparatus is configured to calculate 216 a segment average value based on the extracted time series data of the plurality of segments to determine the segment power value. Thus, the segment power value may be a segment average power value. The apparatus may calculate the average value for each segment based on the time series (measurement) data of the corresponding segment. For example, it has been observed that the average value is sufficient to characterize the performance of the user and is computationally efficient compared to some other statistical parameters.

[0035] In one embodiment, the time series data includes speed data and / or heart rate data, and the apparatus is configured to convert 218 the speed data and / or heart rate data into power output data, and determine the segment power value based on the power output data. In the context of the described embodiments, rate and speed data are technically equivalent because the absolute direction of the user's movement is not of interest. Therefore, rate and speed are used interchangeably herein. When the apparatus is configured to calculate segment averages, the averages may be calculated before or after conversion to power output data.

[0036] When the time series data includes heart rate data, the device may perform a heart rate-power conversion on the heart rate data to obtain power output data. The heart rate-power conversion may be a transformation of the form P=A(HR)+b, where P represents power, HR represents heart rate, A represents a linear mapping coefficient, and b represents a translation parameter. The translation parameter b may be zero.

[0037] Heart rate is linearly proportional to power, at least under certain conditions. In contrast to anaerobic exercise, at least when the user is performing aerobic exercise, the user's heart rate is linearly proportional to the user's power output. The relationship between heart rate and power may also be distorted by the phenomenon of heart rate drift. Heart rate drift is observed when heart rate increases despite constant power output. Rapid changes in exercise intensity may result in large heart rate drifts. Decreases in blood volume, increases in core temperature, and neuromuscular fatigue may result in gradual heart rate drift. The device may be configured to detect heart rate drift based on time series data, or to detect whether the user is performing anaerobic exercise. If neither heart rate drift is detected nor the user is performing anaerobic exercise, the device may perform a heart rate-to-power conversion on the heart rate data to obtain power output data.

[0038] When the time series data includes speed data, the device may perform speed-to-power conversion on the speed data to obtain power output data. The speed-to-power conversion may be a transformation of the form P=Cv+d, where P represents power, v represents speed, C represents linear mapping coefficients, and d represents a translation parameter. The translation parameter d may be zero.

[0039] In some sports such as running, speed is proportional to power, but in other sports such as cycling this is not the case. The device may be configured to detect a sport of athletic performance performed by a user based on the time series data. Alternatively or additionally, the device may receive a sport as input from the user via a user interface. The device may compare the detected sport with a list of sports in which speed is proportional to power, for example, stored in a memory of the device. If the sport is included in the list, the device may perform a speed-to-power conversion on the speed data to obtain power output data.

[0040] The device may utilize a multimodal conversion (such as Polar's running power) to obtain power output data. The device may calculate running power based on speed and gradient time series data obtained from a global positioning system (GPS) and a barometric pressure sensor to generate power output data. The device may determine segment power values ​​based on the power output data.

[0041] The apparatus is configured to form 206 a set of scattered points describing how the fragment power values ​​vary according to the respective durations of the extracted plurality of fragments. For example, the apparatus may obtain the durations of the extracted plurality of fragments from timestamps or metadata of the time series data. The apparatus is further configured to fit 208 a representation power curve to the set of scattered points.

[0042] Figure 3 An example of a set of scatter points and a representative power curve 300 is illustrated. Figure 3 In , the scattered points are plotted in a coordinate system where the vertical axis represents the average power and the horizontal axis represents the duration. Figure 3 In the example, the segment power value is the segment average power value. The performance power curve 300 is fitted to the set of scattered points using, for example, the least square method. Because it is scientifically known that the performance power curve is a decay curve, in particular an exponential decay curve (e -x ), so a basic curve for fitting can be provided. For example, the basic curve can be based on the function e -x , or based on a more complex function that better approximates the general shape of the average performance power curve. It should be appreciated that although the general shapes of the performance power curves of different users are similar, each user has a unique performance power curve. For example, the average power that each user is able to output over a long duration varies depending on the endurance characteristics of the user. In addition, the degree of attenuation may vary from user to user.

[0043] As mentioned above, under some conditions, speed or rate may be proportional to power. The following paragraphs describe a mathematical model in terms of rate, also referred to as a general performance model. The model represents exercise intensity with a relative power scale that varies between 0 corresponding to basal metabolic rate and 1 when maximum aerobic power (MAP) is produced. Alternatively or additionally, the model can be represented in terms of power or heart rate by means of a conversion similar to the conversion described above, for example.

[0044] The relationship between the fastest performance time T and the distance d traveled within the time T is represented by the two equations (1) and (2) shown below. c When d <= d c When , equation (2) applies, where d c By d c =t c v c The crossover distance given, where t c and v c is the crossing time and rate. The crossing rate corresponds to the rate or speed at maximum oxygen consumption (VO2max), while the crossing distance is the distance traveled in time t c The speed v c The distance traveled. Parameter v c This corresponds to the rate at which VO2max can be achieved, and is therefore also referred to as the maximum aerobic rate (MAS). MAS is an example of maximum aerobic power (MAP), which is the highest power a user can output. It is estimated that t c It may be 5 to 7 minutes. In one embodiment, t c For 6 minutes.

[0045]

[0046]

[0047] Parameter γ s and γ l and endurance parameter E s and E l Related. Long distance endurance parameter E l According to the relation E l =exp(0.1 / γ l ) is given, and the short-distance endurance parameter E s According to the relation E s =exp(0.1 / γ s ) are given. The subscripts l and s refer to the long-distance performance and short-distance performance, respectively. -1 is the real branch of the Lambert W function defined as the multivalued inverse of the function w→w exp(w).

[0048] E l Also known as the endurance index, it is the ratio of the duration that the athlete can maintain 90% of MAS to the duration that the athlete can maintain MAS. For example, if the athlete can maintain 90% of MAS for 18 minutes and can maintain MAS for 6 minutes, the athlete's endurance index is 18 / 6=3.

[0049] The above model can be used to obtain a threshold power curve. In one embodiment, the performance power curve is a threshold power curve. The performance power curve can be a power law or a broken power law. Although the performance power curve is described herein in the form of a power and duration function, power and / or duration can be replaced by other parameters whose relationship with power is known or can be estimated. For example, power can be replaced by heart rate or rate or speed, and / or duration can be replaced by distance. The device can use speed data and the formula d=tv to determine the relationship between duration and distance, where d is the distance, t is the duration, and v is the speed. In addition, the variables power and duration (and / or their substitutions) can be switched so that, for example, the duration of multiple segments extracted by the scatter point description varies according to the segment power value.

[0050] Training zones were previously based on the idea of ​​a "threshold power" that describes the power that a muscle can provide for an infinite period of time. This time-invariant threshold power leads to static training zones that are also time-invariant. However, this is contrary to observations, as no power can be sustained indefinitely. Analysis of world records in running shows that the maximum power a runner can sustain decreases rapidly within the first 12 minutes and then decreases at a slower rate indefinitely. A performance power curve personalized to a user in the manner described above describes this characteristic of the user, and the performance power curve is used to generate training zones that differ according to duration, wherein the different properties mean that the training zones are defined differently for physical exercise of different durations.

[0051] The apparatus is configured to determine 210, 224 a set of training areas based on the performance power curve. Determining a set of training areas based on the performance power curve may refer to determining the set of training areas based on the curve itself (e.g., the shape and / or value of the curve), and / or based on one or more parameters characterizing the curve. Figure 4 The diagram illustrates a performance power curve 400 in an average power versus duration coordinate system. Each training zone 402 to 406 has a decay shape that matches the decay shape of the performance power curve. Each training zone may define a (average) power range per duration t for a physical exercise or a portion of a physical exercise. A portion may refer to a major portion of a workout, e.g., it may exclude a warm-up cool-down period if a warm-up cool-down period is set in the workout. Figure 4 In FIG. 4 , three training regions 402, 404, and 406 are determined based on the performance power curve. The training region 402 includes power values ​​above the performance power curve, the training region 404 includes power values ​​equal to, above, and below the performance power curve, and the training region 406 includes power values ​​below the performance power curve. Figure 4 As shown in , the training regions can be non-overlapping.

[0052] Figure 4 The diagram illustrates a performance power curve on a training area 404 having at least one training area 402 above it. In this case, the performance power curve may not be suitable for levels below the user's physiological maximum performance (the maximum comfort level mentioned above). In another embodiment in which maximum performance is used to generate the performance power curve, the maximum power output area 402 can be arranged around the performance power curve. In order to know the training area to which the performance power curve belongs, the device can receive information as user input about whether the performance power curve is based on physiological maximum performance (maximum power output area 402 around the performance power curve) or on a lower performance level (another training area around the performance power curve).

[0053] In one embodiment, the apparatus is configured to determine a training area 402, 404, 406 in a set of training areas by multiplying the performance power curve 400 by a constant multiplier. In this way, information about the relationship between power and duration contained in the performance power curve can be efficiently transferred to the set of training areas, and the training areas can closely follow the shape of the performance power curve. The apparatus can generate the boundaries of the training areas by multiplying the performance power curve 400 by a constant multiplier. In one embodiment, each training area 402, 404, 406 in the set of training areas is defined by multiplying the performance power curve 400 by a constant multiplier. The constant multiplier used for each training area can be different, that is, it is region-specific.

[0054] As an example of some of the above embodiments, three training zones may be defined as follows. The training zones may be somewhat similar to traditional training zones in terms of how each zone characterizes the level of performance (power output) exhibited by the user. The first training zone (the "recovery" training zone 406) may be defined by the performance power curve multiplied by 0.95 as an upper bound. The second training zone (the "production" training zone) may be defined by the performance power curve multiplied by 0.95 as a lower bound, and the performance power curve multiplied by 1.05 as an upper bound. The third training zone (the "maximum" training zone) may be defined by the performance power curve multiplied by 1.05 as a lower bound. The training zones are not necessarily completely bounded; the first training zone may have no lower bound, and / or the third or final training zone may have no upper bound.

[0055] In one embodiment, the constant multiplier is between 0 and 1. As an example, three training zones may be defined in order of increasing power as follows: A first training zone ("recovery" training zone) may be defined with the performance power curve multiplied by 0.5 as a lower bound, and the performance power curve multiplied by 0.7 as an upper bound. A second training zone ("production" training zone) may be defined with the performance power curve multiplied by 0.7 as a lower bound, and the performance power curve multiplied by 0.9 as an upper bound. A third training zone ("max" training zone) may be defined with the performance power curve multiplied by 0.9 as a lower bound.

[0056] In one embodiment, the training regions are fully bounded (in terms of power). Considering the above example again, the third training region may have an upper bound of the performance power curve multiplied by 1.0 or even 1.1, resulting in three fully bounded training regions.

[0057] It should also be noted that the number of training regions in a set of training regions is not limited to three, but may be any suitable number, such as 2, 3, 4 or 5, for example.

[0058] The apparatus is configured to generate 212 training instructions based on a set of training zones. The training instructions may guide the user in one or more physical exercises to achieve a desired training effect. The training instructions may be based on selecting a training zone according to the desired training effect of the exercise, and then selecting the duration and / or average power of the exercise within the training zone. The apparatus is configured to cause 214 a user interface to output the training instructions to the user 100.

[0059] Figure 5 An embodiment is illustrated, including: obtaining 500, 502 input parameters and training goals of a physical exercise (training process); based on the training goals, selecting 504 a training area for the exercise from a set of training areas; determining 506 a power output target based on the input parameters and the selected training area; and causing 508 a user interface to output the power output target of the exercise. In one embodiment, the input parameter is a duration parameter or a distance parameter. For example, the training goal can be a qualitative goal that describes the training effect or training benefit sought from the exercise, such as "recovery", "development" or "competition" (equivalent to peak performance) that describes the purpose of the training process or a part of the training process. The device selects a training area mapped to the training goal from a set of training areas. The mapping of the training goal and the training area can be stored in the memory of the device, for example, the goal "recovery" can correspond to the recovery training area, the "development" goal can correspond to the production training area, and the "competition" goal can correspond to the maximum training area. The device can be configured to use the mapping to select the training area for the training process. The device selects the power output target based on the input parameters and the selected training area. Figure 4The diagram illustrates an example where the input parameter is a duration parameter t1, the training goal is "race", and the selected training zone is the maximum training zone 402 mapped to the training goal "race". The device can determine power values ​​P1 and P2 corresponding to the upper and lower bounds of the training zone at duration t1. The device can determine the power output target as a range, such as [P1, P2], or as a power value, such as the average of P1 and P2. For example, the power output target can be defined based on a rate range or average rate value mapped to the input duration t1. In other words, the user can input that he / she wants to exercise at maximum effort for a specific duration, and the device then uses the maximum effort training zone to find an average rate or rate range specific to that duration to train for maximum effort. As Figure 4 As shown in , for shorter exercise durations, the rate may be greater, while for longer exercise durations, the rate may be reduced. In other words, for different exercise durations, the maximum training effort is achieved at different rates or rate ranges. Instead of using duration as an input parameter, the user can enter the distance he / she wishes to travel during training. Similarly, the device can calculate an average rate or rate range mapped to the selected training area and the input distance parameter. For example, the device causes the user interface to output the power output target to the user, for example as a value or range of values ​​in watts. Since watts are abstract to the average user, rate or heart rate can be used instead.

[0060] The following table shows examples of training area descriptions, where the Figure 4 A finer granularity in the .training zone is provided as a measure of the training effect. In addition, the user can provide the distance or time of the exercise as an input parameter and then evaluate different average times and rates and the corresponding training effect.

[0061]

[0062] %MAP refers to the percentage of maximum aerobic power. The above table is based on a target duration of 30 minutes, where the user achieves maximum effort (training effect 100) in a 30-minute workout at an average rate of 12.1 km / h (4 minutes 58 seconds per kilometer). Now, if the user selects, for example, training effect 90, the device can set a target rate of 10.9 km / h as the power output target for the workout and guide the user to achieve or maintain this rate during the workout. The same user with the same performance power curve enters a 60-minute target duration as an input parameter, and can then obtain the power output target from the table below.

[0063]

[0064] The following table illustrates another example of a power curve for the same user and the same performance as in the above table, where the user now enters a desired distance (3 km in this example). In response to this input parameter, the device can provide a power output target from the following table according to the desired training effect.

[0065]

[0066]

[0067] It should be appreciated that the above table is intended only to illustrate how to convert a performance power curve and corresponding training zones into a power output target for a workout. Actual implementations and user interfaces may vary.

[0068] In the above embodiments, the power output target can represent the average power output of the exercise. For example, this is directly applicable to steady pace exercise, where the user is instructed to maintain a selected rate representing the power output target. In the case where the user exceeds the selected rate, the device can instruct the user to reduce the rate, and in the case where the user's rate is below the selected rate, the device can instruct the user to increase the rate. In other exercises that are not designed to be performed at a steady pace or steady power output, the performance power curve may still be available. Let us consider an interval exercise that includes several working periods and rest periods between consecutive working periods. If a certain training area is selected based on the desired training effect, the device can directly use the training area in the steady pace exercise, but for the purpose of the interval exercise, a certain level can be scaled down. Therefore, the selected exercise type can be another input to the device. Let us remind that in interval exercise, the user is usually unable to show maximum performance in any working period. Therefore, it may be feasible to scale down the power output target and use a lower power output target for the working period of the interval exercise.

[0069] The device may obtain input parameters and / or training goals from the user via the user interface as user input. Alternatively or additionally, the device may read input parameters and / or training goals from a memory of the device, such as from a training plan pre-designed for the user, and / or the device may receive input parameters and / or training goals from other devices, such as a server computer, via a wireless connection.

[0070] Figure 6The diagram illustrates an embodiment related to a power output target. In one embodiment, the device is configured to convert 600 the power output target to a heart rate target and / or a speed target, and cause 604 a user interface to output the heart rate target and / or the speed target as the power output target. Power may be difficult to understand, especially for an inexperienced user, and converting to heart rate or speed may allow the user to better understand the target. In addition, power data may be more difficult to obtain than speed or heart rate data, thus facilitating comparison of the target with the speed and / or heart rate data obtained. The speed-to-power conversion and / or heart rate-to-power conversion or other conversions described above may be used to convert the power output target to a heart rate target and / or a speed target.

[0071] In one embodiment, the device is configured to convert the power output target to a heart rate target, perform 602 drift correction on the heart rate target, and cause the user interface to output the corrected heart rate target as the power output target. As previously described, heart rate drift corresponds to an increase in heart rate, despite constant power output. When the power output target is output as a heart rate target, if the user is experiencing heart rate drift, the heart rate target may be incorrect, or even unattainable. Therefore, drift correction can be used to correct the heart rate target to adjust (usually increase) the heart rate target to a heart rate target that more accurately represents the desired power output target. Drift correction can be based on a heart rate demand, which represents the actual heart rate required for the user's effort. When the relationship between heart rate and heart rate demand is known, the user can be instructed to reach a suitable heart rate even in the event of heart rate drift. A model describing the relationship between heart rate and heart rate demand is described in Stirling, James Robert, et al., A Model of Heart Rate Kinetics in Response to Exercise. Journal of Nonlinear Mathematical Physics, October 2008, Volume 15, Supplement 3, pages 426-436.

[0072] The device may also be configured to obtain the user's power data, speed data and / or heart rate data. The acquisition can be performed similarly to the acquisition of time series data. The device may be configured to cause the user interface to output the acquired power data, speed data and / or heart rate data. The device may be configured to compare the power data with the power output target, compare the speed data with the speed target, and / or compare the heart rate data with the heart rate target. If the power / speed / heart rate data does not meet or reach the corresponding target, the device may be configured to cause the user interface to output an indication to the user. For example, meeting the target in this article may refer to data falling within a target range. For example, the indication may be a graphical indication displayed on a display screen or other display unit, an audio indication output by an audio output device such as a speaker, or a tactile indication output by a tactile output device configured to provide a tactile indication to the user.

[0073] The performance power curve may characterize both the aerobic performance and the anaerobic performance of the user. In order to obtain an accurate estimate of the aerobic performance and the anaerobic performance of the user, the time series data preferably includes data from both types of athletic performance. In one embodiment, the first duration is less than a crossover time representing the time for maintaining a maximum aerobic speed (MAS), and the second duration is greater than the crossover time. As previously described, the crossover time t c It can be 5 to 7 minutes, or in one embodiment, t c For 6 minutes.

[0074] exist Figure 7 and Figure 8 In the embodiment shown in , the apparatus is configured to divide 700 a set of scattered points into a first set of scattered points 800 and a second set of scattered points 802, wherein the duration of the first set of scattered points is less than or equal to a cut-off time, and the duration of the second set of scattered points is greater than or equal to the cut-off time; perform 702 fitting on the first set of scattered points to obtain a first segment 804 of the performance power curve, and perform 704 fitting on the second set of scattered points to obtain a second segment 806 of the performance power curve, and form 706 a performance power curve based on the first segment of the performance power curve and the second segment of the performance power curve. It is observed that the performance power curve can be best fitted with 2 segments. Therefore, the performance power curve can be a (2-segment) piecewise function. In one embodiment, the cut-off time is the cross-border time (t c ). When the power curve is at t c The best fit may be achieved when the device is a piecewise function with a breakpoint 810 at t. The device may perform the fit so that the breakpoint is fixed at the crossover time. Such a fit separates the aerobic performance and anaerobic performance into the first and second segments of the performance power curve, respectively, and allows for a more accurate characterization of the user's performance, since individuals may have distinct aerobic performance characteristics and anaerobic performance characteristics. The duration is less than or equal to tc The performance may be primarily anaerobic or at least represent anaerobic performance, and the duration is greater than or equal to t c of athletic performance may be primarily aerobic or at least represent aerobic performance.

[0075] exist Figure 8 In , the duration is expressed on a logarithmic scale. This facilitates advantages with respect to the fit and form of the performance power curve. In one embodiment, the performance power curve is a piecewise linear function. Fitting a linear function is computationally simple and may be more accurate than fitting a more complex function. The performance profile curve may be defined by parameters such as the respective slopes E, F of segments 804 and 806 and breakpoints 810. The breakpoints may correspond to durations t c .

[0076] Fig. 9 and Fig.10 The diagram illustrates an embodiment related to the acquisition of time series data. In one embodiment, the device is configured to obtain 900 training plan data 1000 associated with time series data 1002; based on the training plan data, identify 902 one or more time periods of the time series data having a training intensity greater than a threshold; and extract 904 multiple fragments from the time series data of the identified one or more time periods. The training plan data may include training programs and / or training process plan data. The training program may include a long-term plan spanning more than a few days, weeks, or months. For example, a training program that follows a 3-week cycle may include the following entries: 1) production week (2 key exercises), 2) production week (1 key exercise), and 3) recovery week. Key exercises may refer to high-intensity exercises that are expected to produce time series data suitable for the purposes described herein. The device may be configured to identify one or more time periods as one or more key exercises based on the training plan data. Training intensity may be defined as the number of training processes per week, energy consumption or calories burned, total power or average power, total duration, or other types of data that can be measured by sensors further described in this document. Training intensity may refer to planned or measured training intensity. The inclusion threshold may be a value of one or more of the above types of data. The device may or may not acquire the time series data for the entire one or more time periods; the device may be configured to acquire (only) a portion of the time series data for the identified one or more time periods.

[0077] The training session planning data may include information about training intervals and other time periods within a single training session. Fig.10The training plan data of the diagram illustrates intervals 1010, 1012, 1014 and 1016. The time series data segments corresponding to these intervals are 1020, 1022, 1024, 1026, respectively. Among these training intervals, 1010 and 1014 can be low-intensity training intervals, while 1012 and 1016 can be high-intensity intervals. Information about the intensity of the training intervals can be encoded in the training plan data. In one embodiment, one or more time periods include one or more training intervals 1012, 1016, and the device is configured to extract one or more of the multiple segments as one or more training intervals. Therefore, the device can extract the time series data segments 1022 and 1026 corresponding to the training intervals 1012, 1016.

[0078] In one embodiment, the apparatus is configured to acquire time series data representing power output of athletic performance of one or more users. The one or more users may include or exclude user 100. When the one or more users do not include user 100, the apparatus may provide training instructions to user 100 without acquiring time series data from the athletic performance of user 100. Thus, the user does not necessarily need to perform any athletic performance for data acquisition, but still receives training instructions from the apparatus.

[0079] Fig.11The diagram illustrates an embodiment related to utilizing and updating a performance power curve during a training process. The device may be configured to select 1110 a training program, for example, in response to a user input. The user may indicate that they wish to train, for example, for a marathon, and provide the device with a user input, for example, selecting a corresponding training program from a selection of training programs displayed on a user interface. For example, the device may be configured to determine 1112, 1114 training goals and input parameters, for example, based on the training program, by selecting a training goal and duration and / or distance of a prearranged training process. The device may be configured to start 1116 a training process using the training goals and input parameters specified above. During the training process, the device may enable 1118 a user interface to output a training guide based on a set of training areas to the user. The device may be configured to obtain 1120 sensor data or additional time series data during the training process. The device may determine 1122 whether the training process is a key exercise, and if so, the device may be configured to extract the training data of the training process and trigger the update of the performance power curve. The device may extract one or more segments from the sensor data or additional time series data obtained during the training process and use them as training data. Alternatively, the device may take the time and duration of one or more working periods of the training process (excluding warm-up and cool-down, if any) or the entire training process for calculation of the average power output of the training process. The device may be configured to update 1124 the performance profile curve and the set training area based on the training data. The update may be performed as in steps 204, 206, 208 and / or 210 of the method of FIG. 2, but using segments extracted from previously acquired time series data and sensor data acquired during the training process.

[0080] Typically, the device may acquire new training data after performing the fitting in 222 or 208; detect from the new training data that the user's maximum power output has increased since the fitting was performed (e.g., the user set a new record time for traversing a certain distance); perform the fitting again 208 or 222 using the new training data, and determine 210 or 224 a new set of training power output zones based on a new performance power curve generated by the newly performed fitting, generate 212 or 226 a new training plan for a new physical exercise based on the new set of training power output zones, and cause the user interface to output 214 or 228 the new training plan. In this way, the training process may be repeated after acquiring more training data and detecting a substantial change in the training data that crosses a certain trigger threshold. Figure 2A or Figure 2B process.

[0081] One or more steps of the above embodiments may be combined with Figure 2A or Figure 2B The method and / or its embodiments are combined to provide a self-updating training guidance solution.

[0082] Fig.12 The embodiment of a portion of an apparatus or system as a block diagram is illustrated. For example, the apparatus may be a wearable device 102 , a portable electronic device 104 , a wearable sensor device 106 , or a server computer 108 . Fig.12 The apparatus may include a device configured to perform Figure 2A or Figure 2B The method or processing system of any embodiment thereof. The processing system may include at least one processor 10 and at least one memory 20. The apparatus may also include a user interface 40, which includes a display or other display unit, an input device such as one or more buttons and / or a touch-sensitive surface, and an audio output device such as a speaker. In some embodiments, the user interface 40 includes a tactile output device configured to provide a tactile indication to the user.

[0083] The processor 10 may include a measurement signal processing circuit 14, which is configured to process time series data by controlling at least some steps of the process of Figure 2, such as steps 200, 202, 204, 206 and 208, or any embodiment thereof. The measurement signal processing circuit 14 may be configured to obtain time series data representing the power output of the athletic performance from the sensor 30 and / or from the sensor 32 connected to the processor 10. The processor 10 may also include a training controller (circuit) 16, which is configured to control the training-related steps of the process of Figure 2, such as generating 212 and outputting 214 training instructions. The measurement signal processing circuit 14 or the training controller 16 may be configured to determine 210 a training zone depending on the implementation.

[0084] The apparatus may include communication circuitry 52 connected to the processor 10. The communication circuitry 52 may include a communication circuit 52 suitable for supporting, for example Hardware and software for communication protocols (such as the Bluetooth Smart specification). It should be appreciated that other communication protocols are equivalent solutions as long as they are suitable for a personal area network (PAN) or suitable for the measurement scenarios described in this document. The communication circuit may include a radio modem and appropriate radio circuits for establishing a communication connection between the server computer and the wearable device. Appropriate radio protocols may include protocols based on IEEE 802.11 or cellular communication protocols. The processor 10 may use the communication circuit 52 to send and receive frames or data according to the supported wireless communication protocol. The frame may carry payload data including time series data measured by the sensor 30. In some embodiments, the processor 10 may use the communication circuit 52 to send time series data, segment power values, a set of scattered points, a performance power curve, a set of training areas and / or training instructions to other devices, for example, to a server computer or a wearable device.

[0085] In the case of a distributed device or system, a first device such as a server computer may be configured to perform some steps of the method of FIG. 2, such as steps 200, 202, 204, 206, and 210. For example, a second device such as a wearable device may be configured to perform steps 212 and 214. The first device may send the output of one or more steps, such as the output of step 210 (a set of training regions), to the second device to facilitate subsequent processing of the output by the second device. Now referring to Fig.11 In another example, the server computer may perform steps 1110 , 1112 , 1114 , 1122 , and 1124 , and the wearable device may perform steps 1116 , 1118 , and 1120 .

[0086] The memory 20 may store definitions for executing Figure 2A or Figure 2B The computer program product 22 of the method or computer program instructions of any embodiment thereof. The memory may also store a measurement database 24, which includes time series data of the measurements, and / or data associated with and / or derived from the time series data, such as segment power values, a set of scatter points, a performance power curve, a set of training areas and / or training instructions. The memory may also store a user profile 26, which stores personal characteristics of the user, such as age, weight, etc.

[0087] The time series data may include data measured from the user by one or more sensors, such as sensors 30 and 32. Sensors 30 and / or 32 may be configured to measure time series data from the user. The device may include sensors 30 and / or 32. The sensors may include one or more of: one or more heart activity sensors, one or more motion sensors, one or more position sensors, one or more air pressure sensors, one or more swimming sensors, one or more power sensors, one or more bicycle sensors, and / or one or more temperature sensors.

[0088] The cardiac activity sensor may be configured to determine cardiac activity, such as, for example, heart rate, heart beat interval (HBI), and / or heart rate variability (HRV). The cardiac activity sensor includes, but is not limited to, a cardiovascular sensor (such as an ECG sensor), an optical cardiac activity sensor (such as a photoplethysmography (PPG) sensor), or a bioimpedance plethysmography. The optical cardiac activity sensor may detect the cardiac activity of the user through optical heart rate measurement, which may include sending a light beam to the user's skin and measuring the light that bounces off and / or is emitted from the user's skin. The light beam may change when passing through the user's veins, and the change may be detected by the optical heart rate activity sensor. An embodiment of an ECG sensor is a wearable sensor device 106, such as, for example Figure 1 The PPG sensor may be integrated into a wearable device 102, such as a heart rate monitor belt as shown in the diagram. Figure 1 In addition, in addition to these types of cardiac activity sensors, other types of biosignal measurement sensors can also be embedded in the cardiac activity sensor. These types include, but are not limited to, the following types: laser Doppler-based blood flow sensors, magnetic blood flow sensors, electromechanical thin film (EMFi) pulse sensors, polarization blood flow sensors. In one embodiment, the cardiac activity sensor can generate raw sensor data of cardiac activity, and / or it can process the sensor data into cardiac activity information, such as, for example, heart rate data.

[0089] The motion sensor may be configured to measure movement of the motion sensor caused by the user by moving the user's hand, chest, head, or other body part to which the motion sensor is attached. The motion sensor may use other motion data, such as the user's position data, to determine the user's movement. In an example embodiment, the motion sensor includes at least one of: an accelerometer, a magnetometer, and a gyroscope. The motion sensor may also include sensor fusion software for combining accelerometer data and gyroscope data to provide a physical quantity, such as acceleration data, velocity data, or limb trajectory data, in a reference coordinate system having an orientation defined by a predetermined gyroscope orientation.

[0090] Position sensors can utilize the Global Navigation Satellite System (GNSS) or other satellite-based or radio-based systems to locate the user and measure various parameters related to the user's movement (speed, distance, location, route, altitude, slope). Indoor location can be detected by indoor location tracking methods, such as mapping techniques that include measuring the earth's magnetic field or radio frequency signals.

[0091] A barometric pressure sensor or altimeter can measure altitude or grade by measuring the atmospheric pressure and matching the measured atmospheric pressure to the altitude.

[0092] Swimming sensors can measure parameters specific to swimming, such as, for example, number of strokes or distance.

[0093] Power sensors can measure power in various sports. Examples of power sensors include bicycle power sensors that can be mounted on a user's bicycle, and running power sensors such as pedometers.

[0094] Bicycle sensors may be sensors attached to various parts of a (stationary) bicycle to measure, for example, rate, cadence or power (output).

[0095] The time series data measured by the sensor or determined by the device based on the time series data may include: heart rate zones, heart rate samples, heart rate variability samples, heart beat interval samples, fat consumption rate, calorie consumption rate, calorie consumption, activity zones, activity samples, speed and / or pace samples, power samples, cadence samples, altitude samples, slope samples, temperature samples, position samples, distance traveled, time elapsed, pedal index, left-right balance, running index, training load, galvanic skin response samples, fluid balance, skin temperature samples, heading samples and / or bicycle angle.

[0096] The term "circuitry" as used in this application refers to all of the following: (a) pure hardware circuit implementations, such as implementations using only analog circuits and / or digital circuits, (b) combinations of circuits and software (and / or firmware), such as (if applicable): (i) a combination of processors or (ii) various portions of processors / software, including digital signal processors, software and memory that work together to enable the device to perform various functions, and (c) circuits that require software or firmware to operate, such as a microprocessor or a portion of a microprocessor, even if the software or firmware is not physically present. This definition of "circuitry" applies to all uses of the term in this application. As another example, the term "circuitry" as used in this application would also cover implementations of only a processor (or multiple processors) or a portion of a processor and its (or their) accompanying software and / or firmware. For example and if applicable to the particular element, the term "circuitry" would also cover a baseband integrated circuit or application processor integrated circuit for a mobile phone, or a similar integrated circuit in a server, cellular network device or other network device.

[0097] In one embodiment, regarding Figure 2A or Figure 2B At least some of the described processes may be performed by a system including corresponding means for performing at least some of the described processes. Some example means for performing the processes may include at least one of the following: a detector, a processor (including a dual-core processor and a multi-core processor), a digital signal processor, a controller, a receiver, a transmitter, an encoder, a decoder, a memory, a RAM, a ROM, software, firmware, a display, a user interface, a display circuit, a user interface circuit, a user interface software, a display software, a circuit, and a circuit system. In one embodiment, at least one processor, a memory, and a computer program code form a processing means or include one or more computer program code portions for performing the processing according to Figure 2A or Figure 2B Any embodiment or operation thereof performs one or more operations.

[0098] The techniques and methods described herein can be implemented by various means. For example, these techniques can be implemented by hardware (one or more devices), firmware (one or more devices), software (one or more modules), or a combination thereof. For hardware implementations, the apparatus of the embodiment can be implemented in one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), processors, controllers, microcontrollers, microprocessors, other electronic units designed to perform the functions described herein, or a combination thereof. For firmware or software, the implementation can be performed by a module (e.g., a process, a function, etc.) of at least one chipset that performs the functions described herein. The software code can be stored in a memory unit and executed by a processor. The memory unit can be implemented within a processor or outside a processor. In the latter case, as known in the art, it can be communicatively coupled to the processor by various means. In addition, as will be appreciated by those skilled in the art, the components of the system described herein can be rearranged and / or supplemented by additional components to facilitate the implementation of various aspects of its description, etc., and they are not limited to the precise configurations stated in the given figures.

[0099] The described embodiments may also be implemented in the form of computer processing defined by a computer program or part thereof. Figure 2A or Figure 2B Embodiments of the described methods can be performed by executing at least a portion of a computer program including corresponding instructions. The computer program can be in source code form, object code form, or some intermediate form, and it can be stored in a carrier of some type, which can be any entity or device capable of carrying the program. For example, the computer program can be stored on a computer or processor readable computer program distribution medium 50. For example, the computer program medium can be, for example, but not limited to, a recording medium, a computer memory, a read-only memory, an electrical carrier signal, a telecommunication signal, and a software distribution package. The computer program medium can be a non-transient medium. The coding of the software for executing the illustrated and described embodiments is fully within the scope of ordinary technicians in the art.

[0100] Although the present invention has been described with reference to one or more embodiments according to the accompanying drawings, it is obvious that the invention is not limited thereto but may be modified in various ways within the scope of the appended claims. All words and expressions should be interpreted broadly and they are intended to illustrate rather than limit the embodiments. It will be obvious to those skilled in the art that with the advancement of technology, the inventive concept may be implemented in various ways.

Claims

1. A device comprising: at least one processor; as well as At least one memory storing computer program code, the computer program code configured to cause the at least one processor to perform operations comprising: Acquiring (220) training data representing power output of a plurality of athletic performances of a user (100), the training data comprising first training data representing power output of a first athletic performance of a first duration and second training data representing power output of a second athletic performance of a second duration different from the first duration; fitting (222) the training data to a decaying performance power curve, the performance power curve representing power output as a function of duration of a continuous motor performance; Based on the fitting, determining (224) a set of training power output regions that conform to the performance power curve by decaying according to duration; generating (226) a training plan for a physical exercise based on the set of training power output zones; as well as Cause (228) a user interface to output the training plan to the user.

2. An apparatus according to any preceding claim, wherein the training data represents the maximum power output of the user in a corresponding sporting performance.

3. The apparatus of claim 2, wherein the maximum power output is defined based on a recorded average power output that the user has performed.

4. The apparatus according to claim 2 or 3, wherein the apparatus is configured to: acquiring new training data after performing the fitting; detecting from the new training data that the user's maximum power output has increased since performing the fitting; and by performing the fitting again using the new training data and determining a new set of training power output zones based on a new performance power curve resulting from the newly performed fitting, and A new training plan for a new workout is generated based on the new set of training power output zones, and the user interface is caused to output the new training plan.

5. An apparatus according to any preceding claim, wherein the training data comprises measurement data of at least a first athletic performance stored in the at least one memory, and wherein the at least one processor is configured to calculate the power output of the first athletic performance based on the measurement data.

6. The apparatus of claim 5, wherein the at least one processor is configured to obtain the training data by at least performing the following operations: identifying (902) consecutive time intervals of a first athletic performance associated with a power output greater than an inclusion threshold; and The duration and power output of the consecutive time intervals are selected for the fitting, and the power output of at least one other time interval of the first athletic performance is excluded from the fitting.

7. An apparatus according to any preceding claim, wherein the at least one processor is configured to: obtaining (500, 502) input parameters representing a duration of the physical exercise and a training goal of the physical exercise; selecting (504) a training power output zone from the set of training power output zones based on the training intensity goal; determining (506) a power output target for the workout based on the input parameters and the selected training zone; as well as The user interface is caused (508) to output the power output target.

8. The apparatus of claim 7, wherein the power output target comprises a heart rate target and / or a speed target.

9. The apparatus according to claim 7 or 8, wherein the training goal describes a training effect sought for the user from the physical exercise, wherein selectable training goals include at least "recovery" and "maximum performance".

10. An apparatus according to any preceding claim 7-9, wherein the input parameter is a target duration or a target distance of the physical exercise.

11. A computer-implemented method comprising: Acquiring (220) training data representing power output of a plurality of athletic performances of a user (100), the training data comprising first training data representing power output of a first athletic performance of a first duration and second training data representing power output of a second athletic performance of a second duration different from the first duration; fitting (222) the training data to a decaying performance power curve, the performance power curve representing power output as a function of duration of a continuous motor performance; Based on the fitting, determining (224) a set of training power output regions that conform to the performance power curve by decaying according to duration; generating (226) a training plan for a physical exercise based on the set of training power output zones; as well as Cause (228) a user interface to output the training plan to the user.

12. The method of claim 11, wherein the training data represents the user's maximum power output in a corresponding athletic performance.

13. The method of claim 12, wherein the maximum power output is defined based on a recorded average power output that the user has performed.

14. The method according to claim 12 or 13, further comprising: acquiring new training data after performing the fitting; detecting from the new training data that the user's maximum power output has increased since performing the fitting; as well as by performing the fitting again using the new training data and determining a new set of training power output zones based on a new performance power curve resulting from the newly performed fitting, and A new training plan for a new workout is generated based on the new set of training power output zones, and the user interface is caused to output the new training plan.

15. A method according to any preceding claim 11-14, wherein the training data comprises measurement data of at least a first athletic performance stored in at least one memory, and wherein at least one processor is configured to calculate the power output of the first athletic performance based on the measurement data.

16. The method of claim 15, wherein the training data is obtained by at least performing the following operations: identifying (902) consecutive time intervals of a first athletic performance associated with a power output greater than an inclusion threshold; and The duration and power output of the consecutive time intervals are selected for the fitting, and the power output of at least one other time interval of the first athletic performance is excluded from the fitting.

17. The method according to any of the preceding claims 11-16, further comprising: obtaining (500, 502) input parameters representing a duration of the physical exercise and a training goal of the physical exercise; selecting (504) a training power output zone from the set of training power output zones based on the training intensity goal; determining (506) a power output target for the workout based on the input parameters and the selected training zone; as well as The user interface is caused (508) to output the power output target.

18. The method of claim 17, wherein the power output target comprises a heart rate target and / or a speed target.

19. The method according to claim 17 or 18, wherein the training goal describes the training effect sought for the user from the physical exercise, wherein selectable training goals include at least "recovery" and "maximum performance".

20. The method according to any of the preceding claims 17-19, wherein the input parameter is a target duration or a target distance of the physical exercise.

21. A computer program product, the computer program product being readable by a computer and comprising program instructions which, when executed by the computer, cause the computer to perform a computer process comprising: Acquiring (220) training data representing power output of a plurality of athletic performances of a user (100), the training data comprising first training data representing power output of a first athletic performance of a first duration and second training data representing power output of a second athletic performance of a second duration different from the first duration; fitting (222) the training data to a decaying performance power curve, the performance power curve representing power output as a function of duration of a continuous motor performance; Based on the fitting, determining (224) a set of training power output regions that conform to the performance power curve by decaying according to duration; generating (226) a training plan for a physical exercise based on the set of training power output zones; as well as Cause (228) a user interface to output the training plan to the user.

Images