JP2011034366A - Music selection method and music selection device of in-vehicle audio apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a music selection method and a music selection device of an in-vehicle audio apparatus, for selecting music depending on the mood of a driver, on the basis of a plurality of vehicle environments. <P>SOLUTION: A degree of an adjective indicating the mood of the driver in each of a plurality of environmental states of a vehicle is registered as first registration data for each adjective, and a genre of music desired in the mood of each adjective and the genre of music desired in the mood opposite to the adjective are registered as second registration data. Actual environment of the vehicle is identified, a value corresponding to the actual environment for each adjective is acquired, and the total value of acquired values corresponding to each environment is set as the value of a music genre desired in the mood of the adjective. A value obtained by inverting the sign of the total value is set as the value of a music genre desired in the mood opposite to the adjective. Then, total values are calculated for each music genre, a ratio of the level of the total value is set as selection probability of the genre to select the genre, and music to be played back is selected from the selected genre. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a music selection method and a music selection device in an in-vehicle audio device, and in particular, estimates a genre of a song that a user would like to hear according to an actual environment around the vehicle, for example, a driving environment, and selects a music from the genre. The present invention relates to a music selection method and a music selection device in an on-vehicle audio device to be reproduced.

Recent vehicle-mounted devices have a navigation function, search for a route to the destination, display the route on a map around the vehicle, and guide the user to the destination. The in-vehicle device has an audio function, receives AM / FM radio broadcast, reproduces a CD, or reproduces music recorded on a recording medium such as a hard disk. In order to reproduce a desired music recorded on a hard disk or the like, an operation for selecting the music is required. However, since a large amount of music can be recorded on a recording medium such as a hard disk, the music selection operation becomes complicated.
For this reason, there has been proposed a method of automatically estimating and reproducing a song that the user wants to listen to according to the travel environment (congested road, highway, specific place or area, weather, etc.) (Patent Document 1). For example, music selection such as up-tempo music on a highway, classical music that calms down on a congested road, music of a specific singer in a place where the sea can be seen, and the like is automatically performed according to the vehicle environment.
In addition, the impression value (X, Y), which is a two-dimensional coordinate value, is associated with music, and depending on the vehicle position (highway, service area, interchange, ordinary road, park, seaside, airfield, etc.) A method has been proposed in which an impression value is obtained and a music piece close to the impression value is selected and reproduced (Patent Document 2).
Although these conventional music selection methods determine the playback music based on the traveling environment or the vehicle position of the vehicle, there is a problem that the music selection that reflects the actual mood when the user is driving is not performed.

JP 2003-150176 A JP 2004-86189 A

  In view of the above, an object of the present invention is to automatically estimate the driver's mood based on a plurality of vehicle environments and to perform music selection that reflects the driver's mood.

The present invention is a music selection method and a music selection device in an in-vehicle audio device.
-Music selection method The music selection method in the in-vehicle audio apparatus of the present invention prepares a plurality of adjectives expressing the user's mood, and first registers a value indicating the degree of mood of each adjective in each of a plurality of environmental states of the vehicle. Register as data, and for each adjective, register the genre of the song that you want to hear when you feel the adjective and the genre of the song that you want to hear when you feel opposite the adjective as second registered data, sensor signal, map A step of determining the actual plurality of environments of the vehicle using information, etc .; for a predetermined adjective, a value corresponding to each actual environment is obtained using the first registration data, and according to each obtained environment The total value is the value of the music genre that you want to hear when you feel the adjective, and you hear the inverted value of the total value when you feel opposite the adjective. The step of setting the value of the desired music genre, similarly, for each of the adjectives, the step of obtaining the value of the musical genre that the user wishes to hear when the mood is in the adjective and the value of the musical genre that the user wants to hear when the mood is opposite to the adjective; Adding the values of the genres for each genre, selecting a genre using the ratio of the magnitudes of the total values as a genre selection probability, and selecting a song to be played from the selected genre.
The plurality of vehicle environmental conditions are presence / absence of a curve, presence / absence of a slope, width of a waterside, a ratio of a building to a plant, and presence / absence of a shield.

Music selection device The music selection device in the in-vehicle audio device of the present invention prepares a plurality of adjectives expressing the user's mood, and stores a value indicating the degree of mood of each adjective in each of a plurality of environmental states of the vehicle. 1 storage unit, for each adjective, a second storage unit that stores the genre of the music that the user wants to hear when the adjective feels and the genre of the music that the user wants to hear when the adjective feels opposite, and photographs the scenery in front of the vehicle A first memory for a camera, a sensor for detecting a running state of the vehicle, an environment discriminating unit for discriminating a plurality of actual environmental states of the vehicle using a photographed image of the camera, map information and sensor signals, and a predetermined adjective The value corresponding to each actual environment is acquired using the stored data of the part, the values corresponding to each acquired environment are summed, and the total value is listened to when the adjective feels. As the value of the channel, the value obtained by inverting the sign of the total value is the value of the music genre that you want to hear when you feel the opposite of the adjective. Similarly, you want to hear each adjective when you feel the adjective The value of the music genre and the value of the music genre to be heard when feeling in the opposite of the adjective are obtained, the values of the genre are summed for each genre, and the genre is selected with the ratio of the total value as the genre selection probability A music selection section for selecting and selecting music to be played back from the selected genre.
In the music selection method and the music selection apparatus, when the total values are summed, the total value can be calculated by regarding the negative values among the values of the music genres as 0. Moreover, the value of the said music genre can be recalculated using a Gaussian function, and the said total value of each music genre can be calculated using the recalculated value.

According to the present invention, a plurality of adjectives expressing the driver's mood are prepared, and for each adjective, the genre of the music to be heard when the adjective feels and the music to be heard when the mood is opposite to the adjective Register the genre, calculate the degree of mood of a predetermined adjective in consideration of multiple vehicle environments, set the degree as the value of the music genre that you want to hear when the mood of the adjective, The value of the music genre that you want to hear when you feel the opposite of the adjective, and similarly the value of the music genre that you want to hear when you feel the adjective and the music genre that you want to hear when you feel the opposite of the adjective The value of the genre is calculated for each genre, and the music to be reproduced is selected from the genre having the maximum total value, so that the music selection that automatically reflects the actual mood during the user's driving is selected. Do Door can be.
Further, according to the present invention, a value indicating the degree of mood of each adjective in each of a plurality of vehicle environments is registered, and the actual plurality of vehicle environments is determined using sensor signals, map information, etc. For the adjective, using the registration data to acquire the value according to each actual environment, summing the values according to each acquired environment, the total value as the degree of mood of the adjective, The degree of mood of each adjective can be determined accurately, and music selection reflecting the actual mood during the user's driving can be performed.

It is a block diagram of the vehicle-mounted audio apparatus of this invention. It is explanatory drawing of the adjective score correspondence table classified by vehicle environment. It is explanatory drawing of an adjective-music genre correspondence table. It is a music recording example of a hard disk. It is a music selection processing flow of the present invention. It is a degree explanatory drawing of the mood of each adjective in the actual environmental state of a vehicle. It is score explanatory drawing of the genre which he wants to listen according to the mood of each adjective in the present vehicle environment. It is explanatory drawing of the score of the genre which he wants to listen when the mood of each adjective and the mood of the adjective which has a meaning opposite to each adjective are the mood. It is score explanatory drawing of the degree which wants to listen of each genre. It is normalization explanatory drawing. FIG. 9 is an explanatory diagram of scores of each adjective in Example 2 (scores of genres that the adjectives want to listen to when they feel) and normalized genre selection probabilities. FIG. 10 is an explanatory diagram of scores of each adjective in Example 3 (scores of genres that the adjectives want to listen to when they feel), and normalized genre selection probabilities. FIG. 10 is an explanatory diagram of a score of each adjective of Example 4 (score of a genre to be listened to when the adjective feels) and a normalized selection probability of the genre.

FIG. 1 is a configuration diagram of an in-vehicle audio apparatus according to the present invention, which is a case where music recorded on a recording medium such as a hard disk is selected and reproduced. Audio sources such as an AM / FM tuner and a CD player are illustrated. Omitted.
The music selection unit 11 includes a music selection processing unit 21, a first storage unit 22 that stores a vehicle environment / adjective score correspondence table EVSC, and a second storage unit 23 that stores an adjective-music genre correspondence table AJMC. Yes. As will be described in detail later, the music selection processing unit 21 calculates the degrees of a plurality of adjectives expressing the driver's mood in consideration of a plurality of current vehicle environments, and based on the mood degree of each adjective. Then, the genre of the music to be reproduced is determined, and control is performed so that the music of the genre is reproduced.
As shown in FIG. 2, the vehicle environment / adjective score correspondence table EVSC prepares a plurality of adjectives expressing the user's mood, and scores indicating the degree of mood of each adjective in each of a plurality of vehicle environmental states (scores) ) Is obtained and registered by a statistical method. This vehicle environment / adjective score correspondence table EVSC is created using a multivariate analysis method, which is a statistical method, and is commonly used by users. As a multivariate analysis method, for example, there is a quantification theory class I (referred to by Tamaro Tatsumi, “Easy Illustration, Questionnaire Analysis Classroom” pages 116-128).

In this embodiment, the following eight adjective pairs are prepared as adjectives expressing the mood of the user during driving. However, the present invention is not limited to such adjective pairs, and other adjective pairs are appropriately adopted. be able to. In 8 adjective pairs 1-8,
Adjective pair 1 is "can't relax-can't"
Adjective pair 2 is "closed-open"
The adjective pair 3 is “Yuutsuna-exhilarating”
Adjective pair 4 is "realistic-fantasy"
Adjective Pair 5 is “boring-exciting”
The adjective pair 6 is "dangerous-safe"
Adjective Pair 7 is “Lonely – Lively”
The adjective pair 8 is "noisy-calm"
It is. In the present embodiment, the following five environments are considered as the vehicle environment state, but the present invention is not limited to such an environment, and other environments can be adopted as appropriate. Each of the five environments is 1. Whether there is a curve on the road,
2. Whether there is a slope on the road,
3. The waterside,
4). The proportion of buildings and plants,
5). The presence or absence of high shielding.

Vehicle Score / Adjective Score Correspondence Table Each score in the EVSC indicates the mood of an adjective (relaxed, open, refreshing, fantastic, exciting, safe, lively, calm). For example, paying attention to the adjective pair 6 “dangerous-safe”,
1) The degree of feeling safe when “curved” is −0.071, and the degree of feeling safe when “without curve” is 0.073. A safe feeling of -0.071 is equivalent to a dangerous feeling of 0.071. 2) The degree of feeling safe when “with gradient” is −0.061, and the degree of feeling safe when “without gradient” is 0.054. A safe mood of -0.061 is equivalent to a dangerous mood of 0.061. 3) The degree of feeling safe while driving on a road with wide waterside is 0.059, the degree of feeling safe while driving on a road with narrow waterside is -0.056, driving on a road without waterside The degree to feel safe inside is -0.011. A degree of safe feeling of −0.056 and −0.011 is equivalent to a degree of dangerous feeling of 0.056 and 0.011. This means that everyone feels safe when driving on roads where the sea and lakes look spacious.
4) The degree of feeling safe when the ratio of buildings to plants is “10: 0”, “7: 3”, “5: 5”, “3: 7”, “0:10” , −0.489, −0.278, 0.088, 0.108, and 0.384. This means that you feel safer as you travel in places with more plants.
5) The degree of safe feeling when the high shield is “existing on both sides”, “existing on one side”, and “not present” is −0.462, 0.023, and 0.417, respectively. This means that it feels safer without a high shield.

As shown in FIG. 3, the adjective-music genre correspondence table AJMC shows, for each of the above 8 adjective pairs, the genre of the music to be heard when the adjective feels and the music to be heard when the mood is opposite to the adjective. Are registered in advance by the user. In the example of FIG.
1) The genre that the user wants to hear when “feeling relaxed” in adjective pair 1 is “Fusion”, and the genre that the user wants to hear when “feels unable to relax” is “rock”.
2) The genre that the user wants to hear when “closed mood” in adjective pair 2 is “jazz”, and the genre that the user wants to hear when “open mood” is “rock”,
3) The genre that the user wants to listen to in the adjective pair 3 is “Jazz”, and the genre that the user wants to hear in the “exhilarated mood” is “Enka”.
4) The genre that the user wants to hear when “realistic mood” in the adjective pair 4 is “rock”, and the genre that the user wants to hear when “fantasy” is “classic”,
5) The genre that the user wants to hear when “feeling bored” in the adjective pair 5 is “Rock”, and the genre that the user wants to hear when “I feel excited” is “Fusion”.
6) The genre that the user wants to hear when “dangerous” in the adjective pair 6 is “rock”, and the genre that the user wants to hear when “safe” is “classic”,
7) The genre that the user wants to listen to in the adjective pair 7 is “Jazz”, and the genre that the user wants to hear in the “lively mood” is “pops”.
8) The genre that the user wants to hear when “noisy” in the adjective pair 8 is “pops”, and the genre that the user wants to hear when “gentle” is “classic”.

Returning to FIG. 1, the music source unit 12 is, for example, a hard disk, which compresses and stores a large number of music pieces, reads out the compressed data of the music piece requested from the music selection unit 11, and transmits audio through the music selection unit. Input to the playback unit 13. FIG. 4 is an example of music recording on the hard disk 12, and FIG. 4A is an example of storing the artist name, genre, music name of the album and music data (music data storage destination address) in association with the album name. (B) is an example of storing the album name, artist name, music name of the album and music data (music data storage destination address) for each genre. Note that the storage example in FIG. 4 is an example, and various storage methods are conceivable.
The audio playback unit 13 expands the compressed data of the input music, converts the obtained PCM audio data into an analog signal, performs predetermined sound quality control, and outputs sound from the speaker. The operation unit 14 includes various keys, and operates the keys to input data and commands to the music selection unit 11, and the display unit 15 displays a song name, performance time, and the like of the reproduced music, and a predetermined input that the music selection unit 11 inputs. A plurality of genre music names are displayed, whereby the user can select a desired music using the operation unit 14 or a touch panel (not shown). The display unit 15 can actually display a navigation map.

The navigation device 16 performs known navigation control. As a sensor, the gyroscope 31 detects the direction of the vehicle, the vehicle speed pulse sensor 32 generates a pulse every time the vehicle travels a predetermined distance, and the angular velocity detects the angular velocity of the vehicle. A sensor 33 and a GPS (not shown) are provided. The navigation control unit 30 detects the position of the vehicle, the inclination angle of the vehicle, and the traveling direction using the output signals from these sensors, reads a map around the vehicle from a map database (not shown), and displays it on the display unit 15. At the same time, a guidance route to the destination is searched and displayed. Further, the navigation control unit 30 identifies the state of the sea and lake around the current driving road from the map data, the water side data WFR indicating the width of the waterside, the data CUV indicating the presence or absence of the curve of the currently driving road, Data SLP indicating the presence or absence of the slope of the currently traveling road is input to the music selector 11.
The in-vehicle camera 17 captures, for example, a landscape in front of the vehicle, and the image processing unit 18 performs image processing on a captured image of the camera, and a ratio of a part other than green, for example, a building and a green part, that is, a plant (building: plant ratio). CPR is calculated and input to the music selection unit 11, and it is detected whether there is a high shielding unit on both sides or one side of the vehicle, and the data SBH is input to the music selection unit 11. Note that the width of the waterside can be identified by the image processing unit 18 processing the captured image and calculating the size of the blue area.

FIG. 5 shows a music selection processing flow according to the present invention. It is assumed that the storage unit 22 stores a vehicle environment / adjective score correspondence table EVSC, and the storage unit 23 has already registered an adjective-music genre correspondence table AJMC.
The music selection processing unit 21 is based on sensor signals and map information input from the navigation device 16 and the image processing unit 18 and includes a plurality of environmental conditions of the vehicle (presence / absence of a curve on the traveling road, presence / absence of a gradient of the traveling road, , The ratio of the building to the plant, the presence or absence of a high shield) is identified (step 101). The actual environmental state of the identified vehicle is as shown in FIG.
“Curved, no slope, wide waterside, building: plant ratio = 10: 0, high shields on both sides”,
Suppose that

When a plurality of actual environments of the vehicle can be identified, the music selection processing unit 21 uses the vehicle environment / adjective score correspondence table EVSC for the first adjective among the eight adjectives, and values corresponding to each actual environment And the values corresponding to the acquired environments are totaled (step 102). For example, for “Relaxing”, the score with a curve is 0.038, the score without a slope is -0.071, the score is 0.063 when the water is wide, the score when the building: plant ratio is 10: 0 Is -0.616, and the score when a high shield exists on both sides is -0.456, so the total value is -1.042. This total value is the degree of mood of the adjective “relaxed” in the current vehicle environment. Similarly, for all other adjectives, the degree of adjective mood in the current vehicle environment is calculated. This
Degree of adjective “relaxing” = 1.042
The degree of adjective “open” mood ＝ −0.680
Degree of adjective “exhilarating” ＝ −0.560
Degree of adjective “fantasy” ＝ −0.999
Degree of adjective “exciting” mood = 0.070
The degree of adjective "safe" mood = -0.909
Degree of adjective “lively” mood = 0.861
The degree of adjective “gentle” mood ＝ −1.180
It becomes.

Next, the music selection processing unit 21 sets the score of each adjective as the score of the music genre to be heard when the adjective feels (step 103). For example, the music genre that you want to listen to when you feel “relaxed” is fusion, so the degree (score) you want to listen to fusion music is −1.042. As a result, as shown in FIG. 7, the genre score that the user wants to listen to can be obtained according to the mood of each adjective in the current vehicle environment.
By the way, when the music genre (fusion) that you want to listen to when you feel “relaxed” is −1.042, you cannot relax, and the degree you do not want to listen to the fusion song is 1.042. In other words, it means that the degree of the genre (jazz) that you want to listen to when you feel “cannot relax” is +1.042. The same is true for other adjectives.
Therefore, the music selection processing unit 21 sets the score obtained by inverting the sign of the score of each adjective as the score of the adjective having the opposite meaning to the adjective, and the score of the genre to be heard according to the adjective having the opposite meaning. (Step 104). As a result, as shown in FIG. 8, it is possible to obtain scores of genres that the adjectives want to listen to when they feel the adjectives and the adjectives having the opposite meaning to the adjectives.

Next, the music selection processing unit 21 adds up the values of the genre for each genre in FIG. 8 (step 105). For example, with respect to the fusion, the sum of −1.042 at the time of “relaxing” and −0.070 at the time of “exciting” gives a total value of −1.112. Similarly, when the total value is calculated for the music genres of rock, jazz, enka, classical music, and pop music, the result is as shown in FIG.
Next, the total value of each genre is normalized (step 106). As a normalization method, there are a method of normalizing only a positive value with a negative value as 0% (see A in FIG. 10), and a method of normalizing with a minimum value as 0% (see B in FIG. 10). .
Next, the maximum genre among the normalized values is regarded as the genre that the user wants to listen to most in the current vehicle environment, and the music source unit 12 selects and reproduces the music to be reproduced from the genre (step) 107). In this case, if the music source unit 12 stores data indicating whether or not it has been played before or the number of times of playback in association with the music, the music selection processing unit 21 has played back or has played many times. Can be preferentially selected.
It is also possible to inquire the music source section 12 for the title of the genre that he wants to listen to most, display the list on the display section 15, let the user select from the list, and play the selected song. .

  Although normalized in the above, normalization processing is not necessarily required, and the genre having the maximum total value in step 105 without normalization is regarded as the genre that the user wants to listen to most in the current vehicle environment. It is possible to select a song to be reproduced from the genre.

In the first embodiment, the genre selection probability is calculated in consideration of the negative adjective score in FIG. 8, but the genre selection probability can be calculated by regarding the negative score as 0.
FIG. 11 (A) is an explanatory diagram of the score of each adjective when the negative score is 0 (the score of the genre that the adjective wants to listen to), and FIG. 11 (B) is FIG. 11 (A). This is an example in which the values of the genre are totaled for each genre based on the above, and the total value of each genre is normalized. In the second embodiment, as in the first embodiment, the classic is determined as the genre that the user wants to listen to most.

を用いる。ガウス関数ｐにおいて、eは元の形容詞得点(図８の得点)であり、このガウス関数で形容詞得点を再計算することにより道路景観からの心理的影響を強調することができる。
図１２（A）はガウス関数を用いて再計算した各形容詞の得点（該形容詞の気分のときに聴きたいジャンルの得点）説明図であり、図１２（B）は図１２（A）に基づいてジャンル毎に該ジャンルの値を合計し、各ジャンルの合計値を正規化した例である。なお、負の値を0％として正の値のみで正規化した場合と、最小値を0％として正規化した場合の両方を示している。
実施例３でも、実施例１と同様にクラシックが最も聴きたいジャンルであるとして決定されているが、その度合は実施例１に比べて大きくなっている。 In Example 1, the score of each adjective in FIG. 8, that is, the score of the genre that the adjective wants to listen to, was used as it was, but in Example 3, the psychological influence from the roadscape was calculated based on the Gaussian function. Recalculate the adjective scores so that they can be emphasized. Then, based on the recalculated adjective score, the values of the genre are totaled for each genre, and the total value of each genre is normalized.
As an example, the Gaussian function p is given by

Is used. In the Gaussian function p, e is the original adjective score (score in FIG. 8). By recalculating the adjective score with this Gaussian function, the psychological influence from the roadscape can be emphasized.
FIG. 12A is an explanatory diagram of the scores of the adjectives recalculated using the Gaussian function (scores of the genre that the adjective wants to listen to when they feel), and FIG. 12B is based on FIG. In this example, the values of the genre are totaled for each genre, and the total value of each genre is normalized. In addition, both the case where the negative value is normalized as 0% and normalized only with the positive value and the case where the minimum value is normalized as 0% are shown.
In the third embodiment as well, the classic is determined to be the most desired genre as in the first embodiment, but the degree is larger than that in the first embodiment.

In Example 3, the negative adjective score in FIG. 12 was also considered. However, the negative score is regarded as 0, only the positive adjective point is recalculated by a Gaussian function, and the recalculated score of each adjective is used. The genre selection probability can be calculated.
FIG. 13 (A) is an explanatory diagram of the scores of each adjective obtained by recalculating only the positive adjective points with a Gaussian function, assuming that the negative score is 0 (the score of the genre that the adjective wants to listen to). FIG. 13B is an example in which the values of the genre are totaled for each genre based on FIG. 13A, and the total value of each genre is normalized. In the fourth embodiment as well, the classic is determined as the most desired genre as in the first embodiment. In the above embodiment, the music recorded on the hard disk or other large-capacity recording medium is selected and reproduced. However, the present invention is not limited to such a case, and requests to download the selected music to an external server via a communication means such as a mobile communication network, wireless LAN, and the Internet, and plays the downloaded music The present invention can also be applied.

DESCRIPTION OF SYMBOLS 11 Music selection part 12 Music source part 13 Audio reproduction part 14 Operation part 15 Display part 16 Navigation apparatus 17 Car-mounted camera 18 Image processing part 21 Music selection part 22 The 1st memory | storage part 23 which memorize | stores vehicle-specific / adjective score correspondence table EVSC Adjective-music genre correspondence table AJMC second storage unit 31 Gyroscope 32 Vehicle speed pulse sensor 33 Angular speed sensor

Claims (7)

In the music selection method in the in-vehicle audio device,
A plurality of adjectives expressing a user's mood are prepared, a value indicating the degree of mood of each adjective in each of a plurality of environmental states of the vehicle is registered as first registration data, and the mood of the adjective for each adjective Register the genre of the song you want to hear sometimes and the genre of the song you want to hear when you feel the opposite of the adjective as second registration data,
Using the sensor signal to determine the actual environment of the vehicle,
With respect to a predetermined adjective, a value corresponding to each actual environment is acquired using the first registration data, the values corresponding to each acquired environment are summed, and the total value is heard when the adjective feels. The value of the desired music genre, and the value obtained by inverting the sign of the total value is the value of the desired music genre when the mood is opposite to that of the adjective. Similarly, for each adjective, Find the value of the music genre you want to hear and the value of the music genre you want to hear when you feel the opposite of the adjective.
Summing up the values of the genre for each genre, selecting a genre with the ratio of the total value as a genre selection probability, and selecting a song to be played from the selected genre.
The music selection method in the vehicle-mounted audio apparatus characterized by the above-mentioned. The environmental conditions of the plurality of vehicles are the presence or absence of a curve, the presence or absence of a slope, the width of a waterside, the ratio of a building to a plant, the presence or absence of a shield,
The music selection method according to claim 1, The negative value of the music genre values is regarded as 0, and the total value of each music genre is calculated.
The music selection method according to claim 1. Recalculating the value of the music genre using a Gaussian function, and calculating the total value of each music genre using the recalculated value;
The music selection method according to claim 1. In a music selection device in an in-vehicle audio device,
A plurality of adjectives expressing a user's mood, and a first storage unit for storing a value indicating the degree of mood of each adjective in each of a plurality of environmental states of the vehicle;
A second storage unit that stores, for each adjective, a genre of a song that the user wants to hear when the adjective feels and a genre of a song that the user wants to hear when he feels the opposite of the adjective;
A camera that captures the scenery in front of the vehicle,
A sensor that detects the running state of the vehicle
An environment discriminating unit that discriminates the plurality of actual environmental states of the vehicle using the captured image of the camera and sensor signals;
For a predetermined adjective, using the stored data of the first storage unit, obtain a value corresponding to each actual environment, sum the values according to each acquired environment, and add the total value to the mood of the adjective The value of the music genre that is sometimes desired to be heard, and the value obtained by inverting the sign of the total value is the value of the music genre that is desired to be heard when the mood is opposite to the adjective. Similarly, for each adjective, the mood of the adjective The value of the music genre that the user wants to listen to and the value of the music genre that they want to hear when they feel the opposite of the adjective are obtained. Select a genre as the selection probability and select a song to play from the selected genre,
A music selection device comprising: The music selection unit considers negative values of the music genre values to be 0, and calculates the total value of each music genre.
The music selection method according to claim 5. The music selection unit recalculates the value of the music genre using a Gaussian function, and calculates the total value of each music genre using the recalculated value.
The music selection method according to claim 5.

Images