JP2008530678A - Electronic device and method for selecting content items - Google Patents

Abstract

The method of selecting content items comprises selecting (6) a subset of content items from the collection of content items in response to feedback collected during the pre-physical activity of the human (3). The subset is smaller than the collection of content items, and the feedback has information about pre-physical activity. The electronic device has an electronic circuit that operates to perform the method of the present invention. The software of the present invention allows a programmable device to perform the method of the present invention.

Description

  The present invention relates to an electronic device for selecting content items, a music player, or a device for transmitting music to, for example, a music memory or a music player.

  The invention further relates to a method for selecting a content item.

  The invention also relates to software that allows a programmable device to perform a method of selecting a content item.

An electronic device for reproducing human perceptual signals is described in WO 2004/072767 pamphlet. WO 2004/072767 describes a sound reproduction device that adapts the tempo of an acoustic signal so that the tempo of the adapted acoustic signal falls within a predetermined acceptable deviation from the selected tempo. ing. The selected tempo can be based on the human heartbeat, the human pace, and / or the user's specific tempo. The piece of sound is selected from the sound source either manually by the user or automatically by the sound playback device itself. The disadvantages of known sound reproduction devices are selected, in particular when the user has to send a piece of sound to the music (player) memory before physical activity and the memory is relatively small in size. A piece of sound is often not appropriate for the physical activity the user is performing.
International Publication No. 2004/072767 Pamphlet

  A first object of the present invention is to provide an electronic device of the kind described in the opening paragraph, which provides a good selection of content items for playback during physical activity. It is.

  A second object of the present invention is to provide a method of the kind described in the opening paragraph, which provides a good selection of content items for playback during physical activity. .

  A first object in accordance with the present invention comprises an electronic circuit that operates to select a subset of content items from a collection of content items in response to feedback collected during human physical activity. The subset is smaller than the collection of feedback and content items with information about pre-physical activity. By using feedback with information about the human pre-physical activity in the content collection step, the content item playback is realized in that the content item that provides the most effective human performance can be selected. The feature of the content item that the playback of the content item results in the most effective human performance can be part of or derived from the feedback. This requires a significant amount of effort even if the user can manually select the appropriate content item. The present invention is particularly advantageous when the user needs to transfer a subset of the content items to the music (player) memory prior to physical activity.

  The content item can have, for example, acoustic, audio, light and / or haptic feedback. The electronic device transfers a subset of the content items to, for example, a memory or portable playback device, the portable playback device itself (eg, a mobile phone) or a fixed playback device (which can be integrated into a fitness device). It can be a PC for. The collection can be a personal collection or a public collection (eg, a content provider collection). The electronic device can be used, for example, in a health club, home or outdoors. The subset of content items can be stored locally or remotely (eg, at a content provider or service provider).

  The feedback can have a human physical state during pre-physical activity. This is particularly useful when a human must achieve the desired pace, but the content item is synchronized with the human physical condition (eg, heart rate). The physical condition can have, for example, a human heart rate. The physical state may be, for example, the average of the human physical state during multiple exercises, the average of the human physical state during each exercise, or the human physical state during each of a plurality of time periods. It can be average.

  Feedback can have a human pace during pre-physical activity. This is particularly useful when a human must achieve a desired physical condition (eg, heart rate), but the content item is synchronized with the human pace. The pace can have, for example, the number of revolutions per minute for riding a bicycle or the number of steps per minute for jogging. The pace can be, for example, an average of human paces during multiple exercises, an average of human paces during each exercise, or an average of human paces during each of multiple time periods.

  The feedback can have a pace associated with the human physical condition during the pre-physical activity. This is particularly useful when the tempo of the sound output signal depends on both the human pace and physical condition. Human fitness, and therefore the physical state of a person at a particular pace, is dynamic, which usually depends on the amount of recent physical activity of the person.

  The feedback may have information identifying which one of the subset of content items was skipped during a particular type of physical exercise. A particular content item may appear appropriate for a particular type of physical exercise based on the physical exercise tempo or the user's general preferences, It turns out that it is not appropriate for the type and, in fact, reduces human performance. This particular content item should not be selected for this particular type of exercise, and probably not even transferred to the music (player) memory.

  The feedback can comprise information identifying one or more effects of the subset of content items on the human physical condition. Some content items can have an unexpected beneficial effect on the human physical condition, ie performance. Those content items need to be suitably selected.

  The feedback can have a specific type of length of physical exercise. The specific type of exercise may need to be postponed because the goal for this exercise has not yet been achieved, or humans may be arbitrarily postponed to enjoy the specific type of exercise Is possible. The exact length of the exercise is the content for each exercise if the user must transfer a subset of the content items to the music (player) memory prior to physical activity and the memory is relatively small in capacity. This is important information for selecting the appropriate amount of items.

  A second object of the invention comprises the method wherein the method selects a subset of content items from a collection of content items in dependence on feedback collected during human pre-physical activity, Less than collecting feedback and content items with information about pre-physical activity.

  The above and other features of the electronic device and method of the present invention will be described in detail below with reference to the drawings.

  In the figures, the same reference numbers are used consistently for the same elements.

  Using music is attractive for sports exercises, particularly for sports that require personal, repetitive and perseverance. Listening to music before and during exercise is thought to increase the output output by dissociation, motivation, and improved annular conditions, and provides improved control and skills to understand motor control It is done. Therefore, music can be an excellent guide for keeping up with exercise, especially when exercise is monotonous, repetitive and boring. Runners running while listening to music can be seen every day on the street.

  Many recreational exercisers and patience athletes / sportsmen use personal training programs for their health and sports performance. Such a training program basically defines the sequence of exercises that need to be performed and the duration and intensity of each individual exercise. Such a training program adapts to the entire plan and is adapted over time to work towards either peak performance or injury recovery.

  The method of the present invention is preferably used in a personal music system for sports exercises that realizes integration of sports training programs, music selection and music playback. The method needs to make sport exercises more enjoyable and help direct the user to the intended purpose as a set in the training program. In short, before it works, the system assists the user with pre-selected favorite music suitable for exercise in the training program. When successful, the song is selected for playback and the music playback characteristics, particularly the music tempo, are adapted to indicate or guide the current user's ability to be manifested in heart rate and frequency of movement Is done. After successful performance and playback data are collected for future music selection improvements through visual inspection, training program adaptation and learning.

Sport exercises that require a lot of personal endurance are motivated to last even when their movements result in feelings of monotony, boredom, distress, exhaustion and fatigue without being immediately rewarded. It needs to be attached. In the field of sports and exercise psychology, exercisers employ a variety of related / internal and dissociated / external strategies to address the physical strength of sports exercise and the physical load for the required endurance It is assumed that With a low physical burden, the exerciser can optionally divert attention (ie 'dissociate') from internal sensations derived from the muscles, joints, cardiovascular and pulmonary systems. The dissociation strategy used can be image formation, positive mood, thinking about work, career or social relationships, ability to work effectively, gaining independence and confidence, or solving problems like mental arithmetic It is. Clearly, at about 70% of the maximum reserve heart rate (MHRR), exercisers are forced to begin switching their attention to return to the internal sensation of physical discomfort from external cues. At levels of physical load dominated by extreme weariness or fatigue sensations (eg, 90% VO ₂ MAX), exercisers can no longer “dissociate” them from the sensation of distress being provided. Can not. Instead, exercisers need to 'fight' directly with pain to continue the exercise, requiring a high level of motivation and tolerance for effort.

  Consumer products have already been brought about for a happy marriage between solo endurance sports and music reproduction. The partnership between electronic products company Philips and sportswear maker Nicke has created a combined device called MP3Run (PSA260), which is a portable music player, capability / training display and logging. Act as a system. The player is a 256 MB flash memory device for playing about 4-8 hours of MP3 or WMA encoded music and has FM radio. Music can be updated by downloading. The player is equipped with a pedometer that wirelessly transmits (using Bluetooth) measure and distance data to a device for display and on-demand acoustic feedback for performance logging. The performance data can be updated to the personal computer for inspection. However, music playback or updates are not tied to its performance.

  A training program defines a series of exercises that are scheduled in a single training. The exercise rules can include the following (see FIG. 1):

Types of exercises that refer to activities during sports include things like running, cycling, rowing, aerobics and fitness.
The duration of the exercise specifies the time period during which the exercise can occur at a predetermined intensity level.
Exercise intensity specifies how much level of exercise activity needs to be performed. For example, the aerobic heart rate region for improved patience and optimal cardiovascular training is 70-80% of MHRR.
Exercise severity indicates how hard or how slowly the exercise is performed in terms of duration and intensity.

  For a particular exercise type, the parameters are irrelevant or cannot be predicted. For example, heart rate during warm-up or stretch exercise is less relevant and probably stable. It is not understood in advance how long it takes to recover from activity before the heart rate stabilizes.

  Training programs can also have a single exercise, such as running for an hour, for those who prefer to run slowly in the park instead of doing a training program. It is necessary to keep in mind. Typically, the training program, for example, strives for peak performance in sporting events, increases patience, increases strength and power, loses weight, becomes healthier, stays healthy Or part of the overall personal coaching plan to achieve well-defined long-term objectives such as rehabilitation or physical rehabilitation. The plan, along with the training program, can be created by sports physiologists, coaches, trainers, sports and health consultancy firms, or by people themselves and supported by today's computer technology.

  In the embodiment of the present invention shown in FIG. 2, the use of favorite music is performed in four stages to assist the user in proceeding toward the intended purpose as a set in the training program. The four stages are an installation stage (not shown), a preparation stage 1, an exercise stage 2, and a feedback stage 4. The installation phase, preparation phase 1 and feedback phase 4 are offline. In other words, the tasks included in these stages are performed at home using a personal computer before or after exercise. In contrast, exercise phase 2 is online while the user is exercising.

Installation Phase The first use of the system requires an installation phase (not shown) to make the system personal. By interacting, the system allows relevant personal data (eg name, gender, age), exercise and physiological data (eg weight, body fat percentage, resting heart rate, maximum or peak heart rate, walking frequency, stride length). ) And sports experience (ie, beginner, intermediate, advanced). Furthermore, exercise definitions can be changed or added along with the likes and dislikes of music. Multiple users can use the system, but they all have personal profiles. Users can always return to this installation phase to reset their data.

Preparatory Phase Preparatory Phase 1 involves creating an offline training program and pre-selecting music. Existing training programs can be used for use, adapted for use, or created from scratch. Subsequently, the favorite music is selected from a large collection of music that adapts the training program. The pre-selection of music before the actual training begins ensures a simple sport and music experience. Furthermore, it addresses storage limitations in today's portable flash-based players. Different training exercises have completely different characteristics in movement, pace, strength and duration. For example, warming up differs from intermediate training exercises because it does not take into account the intensity of exercise or the user's ability. Instead, warming up requires the exerciser's full concentration. Those different exercise features require different selections of favorite music. Both music and training programs are downloaded to the portable player for take. The following requirements are designed for the pre-selection strategy of the system's music.
Personal music preferences regarding music, music genres, artist likes and dislikes, and preferences related to exercises in training programs are considered.
An amount of music that is proportional to the duration of the overall training program and of the individual's exercise is selected.
Music is selected based on the strength of the exercise. In particular, the strength of exercise is converted into a tempo characteristic of music.
User performance indicators are used to learn which songs need to be included or excluded in future pre-selections.

  Preparatory phase 1 has a step 6 of selecting a subset of content items from the collection of content items depending on the feedback collected during the pre-physical activity of human 3, see FIG. 2 for this. However, the subset is smaller than the collection of fiordbacks and content items with information about pre-physical activity. In the embodiment shown in FIG. 2, the feedback proceeds to feedback stage 4. In an alternative embodiment, step 6 can be performed as part of an exercise phase performed at the server. A content playback device (eg, known) can be connected to the server during physical activity. In an alternative embodiment, an off-line phase is not necessary.

The music is defined as a finite vector s = (ν ₁ ,..., Ν _K ) of _K attribute values so as to formalize the music pre-selection process. These attributes reflect catalog data tags or music specific features such as unique identification, title information, artist name, genre, duration, music tempo and time signature. These tag values can be easily supplied by online services from third parties or music signal processing algorithms. All valid values for each kth attribute are collected in the attribute domain _denoted by _Dk .

A finite set of m songs M = (s ₁ ,..., s _m ) represents the user's music collection. The pre-selection of music is accomplished by generating a set of reproductions defined as a (disordered) set of n songs p = (p ₁ ,..., P _K ), where i = 1,. . . , N, p _i εM. P _ik is used to define the value of the k-th attribute of the i-th song in the playback set. For convenience of notation, the size n of the reproduction set is assumed to be known. In practice, the size of the playback set is made variable.

  For each individual exercise, the set of constraints C is defined by either the user or the system on which the playback set can be generated. In addition, constraints that can be applied to the training program as a whole are defined, for example, to ensure that different songs are selected for exercise and that all music fits in the portable player. All playback sets are combined into a pre-selected collection of music used in the exercise stage. A constraint is a restriction expressed in the declaration section of which music needs to be in the playback set and which music should not be. The constraints can reflect restrictions on music attributes suitable for exercise in exercise or the user's musical preferences. A set of replays is said to satisfy the constraint if it meets the constraints specified by the constraint. If not, it is said to be against the constraints.

If a user does not like music with a specific attribute (eg, artist or genre) when performing a specific exercise, she will let the system know about this via the user interface. Selection of types can be avoided. The system is shown here by way of example with constraints being eliminated. Formally, the constraint to be eliminated is given by four tuples (p, n, k, V), where p is the set of regenerations,
n is the size of p,
k is the number of attributes, where 1 ≦ k ≦ K,
V is for all 1 ≦ i ≦ n

を適用できる必要があることを示す属性値の集合

A set of attribute values that indicate that must be applicable

である。

It is.

On the other hand, if the user prefers music with specific attributes, the system will exemplify the counting constraints that limit and require the number of specific attribute values to appear in the playback set. Formally, the counting constraint is given by six tuples (p, n, k, V, a, b), where p is the set of regenerations,
n is the size of p,
k is the number of attributes, where 1 ≦ k ≦ K,
V is a set of attribute values, where

であり、
ａ及びｂはそれぞれ、ａ≦｜｛１≦ｉ≦ｎ｜ｐ_ｉｋ∈Ｖ｝｜≦ｂを適用できる必要があることを示す、下限及び上限である。

And
a and b are a lower limit and an upper limit, respectively, indicating that a ≦ | {1 ≦ i ≦ n | p _ik εV} | ≦ b needs to be applicable.

If the exercise has a limited duration, a set of songs with a similar overall duration needs to be selected, for which a duration constraint can be used. Formally, the duration constraint is given by five tuples (p, n, k, a, b), where
p is the set of replays,
n is the size of p,
k is the number of attributes, where 1 ≦ k ≦ K,
a and b are respectively

を適用できる必要があることを示す、下限及び上限である。

Is a lower limit and an upper limit that indicate that it must be applicable.

Due to the strength and severity of the exercise, the exercise limits the generation of a playback set in the range of musical tempos that need to be included. If it is assumed that the frequency of movement is synchronized with the music (or vice versa), only music with a specific range of music tempo is needed, for which range constraints are valid. Appropriate limits on the music tempo associated with the exercise need to be learned for each particular user. Formally, the range constraints are given by five tuples (p, n, k, v, w), where
p is the set of replays,
n is the size of p,
k is the number of attributes, where 1 ≦ k ≦ K,
v and w are a lower limit and an upper limit, respectively, indicating that for all 1 ≦ i ≦ n, v, wεD _k needs to be able to apply v ≦ p _ik ≦ w. Similarly, a song can be generated for a playback set as if the value for the kth attribute is derived from a statistical distribution (eg, a normal distribution).

It should be noted that the constraints that limit the order of songs are not modeled because the playback set is considered a chaotic set.
1) Implementation of local search: Satisfying a set of constraints is a difficult combination problem for NPs. Thus, it is unlikely that there will be a polynomial algorithm that can compute a replay set that fits any given set of constraints. Therefore, instead of striving to obtain an accurate solution, an approximate solution is computed by using a local search. In this local search method, each playback set constraint is transformed into a normalized partially sensible first order penalty function. Each penalty function is defined to be 0 if the constraint is met, to be greater than 0 if the constraint is not met, and to increase with the amount of violation of the constraint. As an example, the penalty function for the duration constraint is defined as:

ここで、

here,

及びδ＝ｍａｘ（ａ−ｎ・ｍｉｎＤ_ｋ，ｎ・ｍａｘＤ_ｋ−ｂ）である。全てのｋ番目の属性値の和（例えば、全持続時間）が範囲［ａ，ｂ］内にある場合、ペナルティは０である。そうでない場合、そのペナルティは、その和と最近接の下限又は上限との間の正規化された差分である。

And δ = max (a−n · minD _k , n · maxD _k −b). If the sum of all kth attribute values (eg, total duration) is in the range [a, b], the penalty is zero. Otherwise, the penalty is the normalized difference between the sum and the nearest lower or upper bound.

  The total penalty is defined as a weighted convex combination of all included penalty functions. The weight indicates how severe it is when the constraint is not met. Some constraints are very difficult and others indicate a desire that can be partially violated. Choosing songs with very few favorite artists is more difficult than choosing music that can be adapted for exercise.

  The local search considers all solutions and replay sets so as to find a replay set that minimizes the total penalty. That goes from solution to solution by applying a small random change at each iteration. The types of changes that are given include (i) adding music, (ii) deleting music, and (iii) replacing music with new music. Changing the order of the songs is less relevant when constraints on the sequence are not addressed here. If the newly created music selection is more appropriate than the original selection, the new selection is allowed and can enter the next iteration of the local search. If the new solution is worse, the solution can be accepted with a certain probability, which decreases with the degree of deterioration, but also decreases in the course of the algorithm. In other words, the search process is a 'cooling down' known as simulated annealing to avoid becoming trapped at a local minimum. Simulated annealing requires a predetermined cooling schedule for temperature control, and a standard linear cooling schedule is used.

  A minor modification to the above method does not convert all constraints into a penalty function, but handles a small number of easy constraints with preprocessing (ie, constraint propagation principles). For example, if a user wants all songs to be rock songs, this can be easily achieved by using only rock music from the music collection and excluding all other songs. .

Exercise stage 2
Exercise stage 2 is an online stage in which pre-selected music is played when the user is exercising, see FIG. 2 for this. Exercise in a training program affects which music is played. The system (here, a portable player) monitors and records user performance data such as heart rate using a wireless connection with a heart rate sensor and running speed using a wireless connection with a pedometer. Depending on the motivational mode required, music playback is adapted to current user performance or current training goals. In addition to standard linear playback, the following music playback advance mode is realized.

  Constant Tempo: The music is played at a constant tempo to motivate the exerciser to synchronize with the exerciser's pace by music for improved patience. Its purpose is to help the exercise to distribute her effort evenly over a distance and to reduce recovery loss with a slightly higher strength effort.

  Pace adaptation: The music tempo is continuously adapted to the exerciser's current stride frequency to support running at varying stride frequencies without having to focus on tune.

  Pace impact: The tempo of a music increases / decreases the speed so that the exerciser is motivated to increase / decrease the speed in the same manner by synchronizing the pace of the music and the exerciser. In this case, the purpose is to reach a performance level as defined in the training program.

  Different modes of playback relate to different ways of exercising that can vary in type of movement (eg, running, cycling, interval, jogging), pace, strength and duration. It is assumed that user performance is measured when exercising. The user's performance can be expressed as the running pace at the number of steps using the pedometer, the cycling pace at the number of rotations using the pedometer, or the heart rate in beats per minute using the heart meter. For running, the number of steps can be taken as a unit of performance. For cycling, the number of cyclings can be taken as a unit of performance. For heartbeats, one heartbeat can be taken as a unit of performance. Different units of performance can be defined for different sports. Music playback characteristics (eg, tempo, loudness, sound source location, equalization) are then adapted based on the current user performance and the currently selected playback mode. For example, the tempo of music playback can be controlled by music conversion technology (eg, time expansion) as needed. The playback mode is expanded across songs, which means that a new song is selected based on the playback mode and the current user performance.

  Obviously, the user can control in playback and songs and exercises can be skipped or extended using interactive controls in the player. Finally, the duration and strength of the current exercise is monitored to signal the user to proceed to the next exercise. The selection of music is based on the current user performance, the current exercise goal, the music playback mode and the history of the music that has already been played.

  Different modes of music playback require different controls and music selection strategies. Standard linear reproduction does not require further sophistication. The system includes a portable music player, a heart rate monitor (typically having a chest belt with a wireless transmitter and a receiver connected to the player), and a stride frequency monitor (pedometer connected to the player). Are used). The chest belt used transmits a pulse for each beat. The system uses a 5-point moving average of the incoming pulse interval, such as the current heartbeat. For stride times, a 5-point moving average is also used.

  A) Constant tempo: Music playback at a constant tempo changes the tempo of a predetermined musical piece by time expansion if the original tempo of the musical piece does not match the desired tempo. In short, time extension is a shortening or extension of acoustic data without a pitch change according to a scaling factor. Obviously, this needs to happen faster than real-time for online decompression and playback, and standard time domain techniques based on synchronous overlap and add (SOLA) are used. The scaling factor represents the ratio between the desired tempo of the music and the original tempo. Time stretching is not done without difficulty for extreme scale factors because music that is changed too widely (eg, over 125%) may sound inappropriate. Therefore, the (next) song for playback is defined as the song whose original tempo is closest to the desired tempo and has not been played recently. If necessary, songs by one artist are not played continuously.

  B) Pace adaptation: Exercisers are synchronized by music that is synchronized with their exerciser's movements. Pace adaptation synchronizes music with the user's stride count. The number of strides or pace is the number of landings of the user's foot on the ground in one minute, and is expressed in steps per minute (spm). The tempo of music is expressed in beats per minute (bpm), in which one music unit in 4/4 time has 4 beats. Music needs to be played faster or slower by using time stretching to adapt the music tempo to the number of strides. A typical maximum value for the scaling factor involved in time stretching is -15% to + 25%. For songs with a tempo of 100 bpm, this results in an extension range [85,125].

  Playing music at 75 bpm tempo when the stride count is 150 spm is also considered fit because the foot landing moment still matches the beat of the music. Note that it is assumed that the music tempo and the user's stride are synchronized. In general, multiplication or integer division of the number of strides is considered adaptation to the music tempo.

  For example, if the current music tempo is 100 bpm (meaning the music extension range [85, 125]) and the stride count is 210 spm, the system changes the music playback tempo to 105 bpm. When the number of strides is 130 spm, it is not possible to obtain an effective fit within the extension range of the music. In this case, the music playback tempo is changed to the expansion boundary closest to the number of strides by the time expansion of the music. In that case, a new musical piece is selected under the condition that the current musical piece is reproduced sufficiently long (for example, to shorten a repeated musical piece that changes over time for 30 seconds). New music is required to have an expansion range with the current number of strides and the current playback tempo to allow changes without sudden tempo changes.

The playback tempo changes smoothly. When changes in tempo are made instantaneously, a temporary chick-up in music playback is obvious. On the other hand, when the change takes a long time, the system response is too slow to properly fit the user's stride count. The parameters specifying the system response time in pace adapted mode is T _m, the T _m is to identify the maximum change has to take time in the reproduction tempo is defined by extension range of the music. The linear function is used to calculate the actual time so as to change from a predetermined playback tempo to a new playback tempo. This function, represented by t _m , is given using two points: t _m (0) = 0 and t _m (Δ _max ) = T _m , where Δ _max is the maximum change in playback tempo. Represents. Although a primary relationship is assumed, as a future extension, the square root relationship can be used to propagate small changes relatively quickly rather than large changes.

Considering the above example with the assumption T _m = 1000 msec, the change from 100 bpm to ₁₀₅ bpm takes t _m (Δ _{100 → 105} ) = P | 100−105 | / | 80−125 | = 125 msec.

C) Pace impact: The method of the system uses the following four steps to motivate the exerciser to reach and maintain a heart rate within a specific heart rate range.
1) Adapt the user's stride count and music playback tempo to ensure the relationship between exercisers.
2) Determine the heart rate target, and then the stride frequency and music playback tempo sub-targets.
3) Propagate the change in the music playback tempo to the playback tempo required by the sub-target determined from the current playback tempo.
4) Wait for the heartbeat to stabilize.

  The above steps are repeated in a continuous loop until the exercise goal is achieved.

  By adapting the music playback tempo to the number of strides in stage 1, it is assumed that the change in playback tempo in stage 3 has a synchronization effect on the number of strides of the user. In other words, the exerciser automatically maintains the exerciser's movement in response to subsequent changes in the music playback tempo. The adaptation of the playback tempo and the number of strides is performed by using the pace adaptation mode described above. As soon as the stride count is met, the system proceeds to the next stage.

  In the second stage, the user's current heart rate is compared to the current exercise heart rate target and the percentage difference is determined. The required number of strides is predicted by applying the difference rate of the heart rate to the current number of strides. From sports physiology, it is known that heart rate varies linearly with the intensity of exercise. In other words, a 10% increase in heart rate means a 10% increase in stride frequency. The desired stride count is then converted to the desired music playback tempo, which can be a multiplication or an integer division of the desired stride count.

  When the current song cannot be further extended to accommodate changes in the music playback tempo, as in the pace adaptation mode, the system will return to the nearest neighbor value (ie, the maximum or minimum value). Music expansion boundary) Change the playback tempo. Subsequently, the system selects a new song with an extension range that has reached the desired and desired music playback tempo.

  The system stores a history of heart rate measurements. From that history, the system can predict whether the heart rate (and hence the desired playback tempo) will improve over the next 30 seconds. This prediction gives useful information about which song is best to be selected next if a change in song is needed for playback. The system effectively minimizes the number of song changes required.

Changes in the music playback tempo are not performed immediately because sudden changes in the playback tempo prevent the user from following the change, or because of disagreement or fatigue. Instead, the system considers propagation time. As in the pace adaptation mode, it is used to calculate the actual time so that the linear function changes from a predetermined playback tempo to a new playback tempo. This function, represented by t _i , is given using two points: t _i (0) = 0 and t _i (Δ _max ) = T _i , where Δ _max is the maximum change in the playback tempo. Represents.

  Finally, when the change is propagated, the system waits for the user's heart rate to stabilize before making additional changes. From experience when testing, the user's cardiovascular system requires some time to adapt to additional effort. Rather than specifying a fixed waiting time after each change, the system seeks a tendency to reduce the difference in successive heart rates with a small differential value. In addition, the music is played at a higher volume, lower frequencies are emphasized and / or as if the music is in front of the listener so as to prompt the user to concentrate and give better results. Can be played back. A song with a tempo corresponding to a slightly higher level of performance than the current user's performance can be selected. Furthermore, the music is played more gently, the lower frequencies are attenuated, and / or as if the music is after the listener so as to prompt the user to produce good results without much concentration. It can be played back. A new song can be selected that has a tempo corresponding to a slightly lower level of performance than the current user's performance.

Feedback stage 4
Feedback stage 4 is connected to offline performance storage and feedback. The performance data, the training program that is actually executed, and the list of songs that are actually played are stored for user inspection and system learning purposes. Performance data can be displayed graphically for testing, self-assessment, motivation, and reporting on fit to current training methods. The system obtains data to learn which music is best suited for which type of exercise.

Finding music that fits a training program The most important feature that determines the fitness of a song in the database of music used in a particular exercise is the tempo of the song. Other features include, for example, percussion characteristics (the amount of beats in music), time signature, keynote, musical instruments used in recording, and the like. Because the amount of storage space is often limited, consider user preferences to 'do not play the same song twice during one training program' or 'do not play songs by the same artist continuously' There needs to be a way to ensure that the right music is in the portable device so that the right music is always found. This is accomplished by generating tempo probability graphs for each exercise and then combining those graphs against the overall tempo probability graph for the entire training program.

  In an embodiment of the method of the present invention, a pace that fits the exercise heart rate range is determined to find music that fits the training program. Unfortunately, this relationship is not an easy relationship. The heart rate that results from running at a particular pace is particularly factors such as endurance (how long the run is already running), fatigue, runner fitness, weather conditions and time of day. Therefore, it varies from user to user and over time. However, a specific average tempo can be assumed and is obtained as in the following table.

エクササイズについての平均のテンポを決定するための他の重要な特徴は、トレーニングプログラムにおけるエクササイズの位置である。例えば、エクササイズを開始したときに１１０の心拍数に達する必要があるペースは９０であることが可能である一方、３０分後には、８０のペースが、この心拍数に達するには十分である。それ故、ユーザの疲労曲線は性別及び年齢に基づいて仮定されることが可能であり、その疲労曲線において、特定の心拍数レベルに達する必要があるペースに関する時間の影響が記録されている。それらの粗い表及び疲労曲線は、性別及び年齢群によるユーザの分類に基づいている。この情報は、ユーザにより容易に入力されることが可能であり、システムの較正について多くの時間を必要としない。

Another important feature for determining the average tempo for exercise is the position of the exercise in the training program. For example, the pace that needs to reach 110 heart rates when exercising can be 90, while after 30 minutes, 80 paces are sufficient to reach this heart rate. Thus, a user's fatigue curve can be assumed based on gender and age, in which the time effect on the pace at which a particular heart rate level needs to be recorded is recorded. Their rough tables and fatigue curves are based on the classification of users by gender and age group. This information can be easily entered by the user and does not require much time for system calibration.

  Given a predetermined value for each pace of exercise in the training program, a Gaussian distribution such as the following can be assumed from their average pace and the possibility of a specific tempo in exercise with a typical heart rate region: Can be evaluated.

この簡単なガウス関数における平均値μはエクササイズについて評価された平均値であり、ここでは、標準偏差として、σは得られるグラフの幅を表す。標準偏差は、前のエクササイズの評価されたテンポ（このテンポは、現時点のエクササイズが開始されるテンポである可能性があるため）及び次のエクササイズのテンポ（同様の理由による）を考慮して決定されることが可能である。上記の全ては、ユーザのクラス（年齢及び性別の群）のパフォーマンスに関する統計的情報を考慮する。また、トレーニングプログラムからの情報を用いることができる。しかしながら、エクササイズの間、ユーザのパフォーマンスに関するデータが記録され、このデータは、エクササイズの後のフィードバック段階で収集される。このデータは、その分布がより個人的にされるように用いられることができる。

The average value μ in this simple Gaussian function is the average value evaluated for the exercise, and here, as a standard deviation, σ represents the width of the obtained graph. Standard deviation is determined taking into account the estimated tempo of the previous exercise (since this tempo may be the tempo at which the current exercise begins) and the tempo of the next exercise (for similar reasons) Can be done. All of the above considers statistical information regarding the performance of the user's class (group of age and gender). Information from the training program can also be used. However, during the exercise, data about the user's performance is recorded and this data is collected in a feedback phase after the exercise. This data can be used to make the distribution more personal.

  First, an average deviation is determined for a particular type of exercise, and this average deviation is used with statistical information that is already valid to determine a good standard deviation σ for this class of exercise. Also, the average tempo μ in the function is determined fairly accurately when considering user performance. An effective way to perform such a combination is to generate an actual curve from the user's performance and average this curve with a previously generated curve. The same process can be applied to the user's fatigue curve. In this way, tempo prediction becomes increasingly appropriate with each new exercise completed by the user.

  When generating a more or less appropriate distribution for each individual exercise, all those curves are combined into one overall distribution for the entire training program. The relative weight of the exercise can be taken into account. This weight is determined by examining the duration of the exercise over the length of the overall training program. One problem that arises is that the duration of some exercises may be expressed in kilometers or target heart rate instead of time. For distance duration exercises, the distance can be multiplied by the average running speed, taking into account the user's fatigue curve. For the target heart rate, the user's fatigue polarity can also be taken into account.

  When the overall distribution is evaluated for the entire training program, this distribution can be mapped to the total playing time available in the music (player) memory. Preferably, the total playback time depends on the amount of space in memory and the average compression level for individual songs. This graph adaptation is done in one of the constraints that a song needs to satisfy in preparation stage 1.

User Interface In an embodiment of the system, the system allows the user to stop, skip, pause or resume exercises in the training program, current playback of songs in the playlist, and music playback mode. The music player itself is fully trained by playing only pre-selected music for each exercise and taking into account the time required to achieve a pre-defined performance level or the duration of the exercise Kept against the program. User control of music playback is achieved by the present invention having a user interactive system that provides access to interactive controls.

  The system starts by playing music preselected for the first exercise. The type of exercise is communicated to the user via speech synthesis and / or text / display cues. The music information is transmitted to the user by display / speech synthesis, for example. The exercise can have a predetermined fixed duration or can have a variable duration. In the latter case, the exercise is extended until a predetermined level of user performance is achieved or as long as the predetermined level of user performance needs to be maintained.

  When exercising and listening to music, if the user needs full attention to other tasks, such as crossing the road or talking to passers-by, the user can temporarily play and exercise the music. Can be stopped. The music and exercise can be resumed later. If the exercise corresponds to reaching or keeping a certain level of performance, resuming can mean starting the exercise again. The user receives information about the start of the exercise by display and / or by speech synthesis.

  When reproduction of a specific music is disliked, the user can skip the music currently being reproduced. In that case, the next pre-selected music is played. The selection of this new music can depend on the playback mode. Information about new music is communicated to the user. The skipped music can be played repeatedly until a favorite music is found. The selection of new music is cyclical, allowing the user to recover errors and full navigation through all of the music pre-selected for exercise.

  The user is given information when the exercise ends. The end of the exercise means that the predetermined duration has ended, or that the user has successfully reached or has maintained a certain level of performance. This information can be conveyed by auditory cues, vibration cues, text / display cues or LED light cues, or by changing (eg, slowing, calming) the characteristics of music playback. In that case, the user decides to proceed to the next exercise by skipping the current exercise. The user can also decide to continue the current exercise, and after a short period of time, the device resumes the current exercise. At any time, the user can decide to skip the current exercise as if the exercise has not yet ended. Moving to the next exercise is cyclic. If the exercise is the last exercise in the training program, the next exercise is the first exercise in the program. This cyclic nature allows for error recovery and full navigation by the user. Other embodiments of the system may have only some of the features described above.

  The electronic device 21 of the present invention has an electronic circuit 23, for which reference is made to FIG. The electronic circuit 23 operates to select a subset of the content items from the collection of content items depending on the feedback collected during the pre-physical activity of the human, and the subset stores information about the pre-physical activity. Less than the feedback and content item collection you have. The electronic device 21 can be fixed or portable. The electronic device 21 is, for example, a content playback device for playing back content when a human is performing physical activity, such as a portable music player, a memory used in the content playback device, or a content playback device. It can be a device for transferring items. The electronic device 21 may have an input 25 for receiving a subset of content items, for example a USB port or a network connection. The electronic device 21 can have an output unit 25, for example a USB port or a network connection, to transmit a subset of the content items to the memory used by the content playback device or to the content playback device. The electronic device 21 can comprise a storage means 31 (fixed or removable), for example a hard disk, a solid state memory and / or an optical disk reader. The storage means 31 stores, for example, a collection of content items, a subset of content items and / or collected feedback. The electronic device 21 may have playback means 29, for example a speaker, a haptic feedback generator, a display or a lamp, for playing a subset of the content items. Software can be used to operate electronic circuit 23 to perform the method of the present invention.

  Although the present invention has been described above in connection with a preferred embodiment, it will be apparent to those skilled in the art that modifications within the principles outlined above are possible and thus the present invention. Are not intended to be limited to those preferred embodiments, but are intended to encompass such modifications. The invention belongs to each and every novel characteristic characteristic and to a combination of each and every characteristic characteristic. The use of the expression “comprises” and its derivatives does not exclude the presence of elements other than those listed in a claim. Using the singular representation of an element does not exclude the presence of multiple such elements.

  As will be appreciated by those skilled in the art, a 'means' can be a separate function or a separate element so that a function is performed alone or associated with another function. Means having any hardware (separate or integral circuit or electronic element) and software (program or part of a program) designed to work or executed to operate To do. The present invention can be implemented by hardware having several distinct elements and by a suitably programmable computer. 'Software' is any software that can be downloaded over a network such as the Internet and stored on a computer readable medium such as a floppy disk that is marketable in any other way. It is understood as meaning.

FIG. 4 shows an example of a training scheme used in an embodiment of the method of the present invention. FIG. 4 shows a flowchart of an embodiment of the method of the present invention. It is a block diagram of the electronic apparatus of this invention.

Claims (9)

  An electronic device having an electronic circuit, said electronic circuit operating to select a subset of content items from a collection of content items in dependence on feedback collected during a human pre-physical activity; An electronic device, wherein the subset is smaller than the collection of the feedback and content items with information about the pre-physical activity.   The electronic device according to claim 1, wherein the feedback comprises the human physical state during the pre-physical activity.   The electronic device of claim 1, wherein the feedback comprises the human pace during the pre-physical activity.   The electronic device according to claim 1, wherein the feedback is related to the human physical condition during the pre-physical activity.   The electronic device of claim 1, wherein the feedback comprises information identifying which one of the subset of content items was skipped during a particular type of physical exercise.   The electronic device of claim 1, wherein the feedback comprises information identifying one or more effects of the subset of the content items on the human physical state.   The electronic device according to claim 1, wherein the feedback has a specific type of length of physical exercise. A method for selecting content items:
Selecting a subset of content items from a collection of content items depending on feedback collected during a human pre-physical activity, the subset comprising information about the pre-physical activity And smaller than the collection of said content items;
Having a method. Software that enables a programmable device to perform the method of claim 8.

Images