Audio playback control method for intelligent speakers

Abstract

The present invention discloses an audio playback control method for smart speakers, and relates to the field of smart speaker control. By collecting the adjustment behavior data of the user on the playback parameters of the smart speakers during use, the usage time data of the user after any adjustment behavior is obtained, and the sampling mean of the usage time data is used as the classification basis, the user's satisfaction after adjusting the playback parameters is classified from high to low, and the playback parameters are used as user needs. According to the classification results of the satisfaction, the playback parameters are classified into user demand attributes using the KANO model, and the satisfaction influence SI value and the dissatisfaction influence DSI value are obtained. The sensitivity of the obtained playback parameter is used as the evaluation basis of the user's preference for this type of playback parameter at the current moment, and the adjustment strategy for this type of playback parameter is generated according to the evaluation results of the user's preference, and the playback parameter of the smart speaker is adjusted according to the adjustment strategy.

Description

Technical Field

[0001] The present invention relates to the field of intelligent speaker control, and in particular to an audio playback control method for an intelligent speaker. Background Art

[0002] The development of economy and technology and the birth of the Internet and the Internet of Things have given birth to many intelligent products. Smart speakers have become an indispensable smart device in many families, providing a lot of convenience and fun for people's lives. In particular, with the continuous development of intelligent technologies, such as voice interaction technology, current smart speakers can play audio according to users' actual needs and user habits, meeting users' personalized needs.

[0003] However, at present, the personalized demand control strategy for smart audio playback is still based on command input control methods such as voice recognition and voice interaction, and the audio equipment itself does not have the function of autonomous adjustment and control. Therefore, in order to improve the intelligence level of smart speakers, so that they can perform automatic control of audio playback based on the autonomous analysis results of individual users' actual needs and usage habits, and better serve users, we propose an audio playback control method for smart speakers. Summary of the invention

[0004] The main purpose of the present invention is to provide an audio playback control method for an intelligent speaker, which can effectively solve the problems in the background technology.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is:

[0006] An audio playback control method for a smart speaker, comprising:

[0007] Collecting the user's behavior data on adjusting the playback parameters of the smart speaker during use, including first behavior data on adjusting the equalizer, second behavior data on adjusting the playback volume, third behavior data on adjusting the sound focus point position, and fourth behavior data on adjusting the surround effect;

[0008] Obtaining usage time data of the user after performing any of the adjustment behaviors, and using the sampled mean of the usage time data as a classification basis, classifying the user's satisfaction after adjusting the playback parameter of the i-th item into five levels from high to low, namely, the first level, the second level, the third level, the fourth level, and the fifth level;

[0009] The satisfaction classification process after adjusting the playback parameters includes the following steps:

[0010] Get the user's usage time of the smart speaker after the user performs the adjustment behavior described in item i for the kth time ;

[0011] Calculate the sample mean of the time the user uses the smart speaker after performing the adjustment behavior described in item i ,in, , where The total number of times the user performs the adjustment behavior described in item i;

[0012] Get the user The minimum usage time of the smart speaker after performing the adjustment behavior described in item i and maximum value ;

[0013] Calculate the sample mean Minimum usage time The distance between and the sampling mean Maximum value of usage time The distance between , the calculation formulas are: ; ;

[0014] According to distance and distance The k usage duration data are divided into different level intervals, and the division principle is:

[0015] when ≤ ,and ≤ , then the usage time Divided into the third level interval, among which, For usage time With the sampling mean The distance between

[0016] when < ,and < ≤ , then the usage time Divided into the fourth level interval;

[0017] when < ,and < , then the usage time Divided into intervals up to the fifth level;

[0018] when > ,and ≤ ≤ , then the usage time Divided into the second level interval;

[0019] when > ,and > , then the usage time Divided into first-level intervals;

[0020] when > ,and > , then the usage time Divided into first-level intervals;

[0021] The level gradient of user satisfaction after adjusting the playback parameters described in item i is determined based on the division results of the usage time data. The determination principle is:

[0022] When using the duration When divided into the first level interval, the level gradient of satisfaction is the first level;

[0023] When using the duration When divided into the first to second level interval, the level gradient of satisfaction is the second level;

[0024] When using the duration When divided into the third level interval, the level gradient of satisfaction is the third level;

[0025] When using the duration When divided into the fourth level interval, the level gradient of satisfaction is the fourth level;

[0026] When using the duration When divided into the fifth level interval, the level gradient of satisfaction is the fifth level.

[0027] The playback parameters are used as user needs. According to the classification results of satisfaction, the playback parameters are classified into user need attributes using the KANO model. The attributes of the user needs include four types: basic needs, expected needs, attractive needs, and indifferent needs. The classification process is as follows:

[0028] Constructing a forward problem and a reverse problem, wherein the forward adjustment of any playback parameter is used as the forward problem, and the reverse adjustment is used as the reverse problem, for example, for the second behavior data, increasing the playback volume is used as the forward problem, and decreasing the playback volume is used as the reverse problem;

[0029] The adjusted satisfaction classification is statistically analyzed, and a two-dimensional attribute classification matrix is ​​constructed based on the statistical results, as shown in the following table:

[0030]

[0031] Table 1 Satisfaction classification two-dimensional attribute classification matrix

[0032] In the table, M, O, A, and I correspond to the four types of demand attributes, namely, basic demand, expected demand, attractive demand, and indifferent demand. Q and R correspond to the suspicious demand and reverse demand in the KANO model theory, which are not involved in this solution:

[0033] Calculate the Better-Worse coefficient of the user for each of the user demand attributes, and obtain the satisfaction influence SI value and the dissatisfaction influence DSI value;

[0034] The calculation formulas for the satisfactory influence SI value and the unsatisfactory influence DSI value are: ; ; In the formula, A, O, M, and I are the frequencies of attractive demand, expected demand, basic demand, and indifferent demand respectively;

[0035] Using the formula , calculate the sensitivity of each playback parameter, where, It is expressed as the sensitivity of the i-th playback parameter; It is expressed as the SI value of the satisfactory influence of the i-th playback parameter; It is represented by the DSI value of the i-th playback parameter;

[0036] To obtain the sensitivity of the i-th playback parameter As a basis for evaluating the user's preference for the playback parameters at the current moment, an adjustment strategy for the playback parameters of the type is generated according to the evaluation result of the user's preference;

[0037] The evaluation basis of the user's preference for the i-th playback parameter at the current moment is:

[0038] When the sensitivity of the i-th playback parameter When the value is on the left side of the factor selection line L of the KANO model, it is determined that the user's preference for the i-th playback parameter at the current moment is favorable, and the playback parameter does not need to be adjusted;

[0039] When the sensitivity of the i-th playback parameter When the value is on the right side of the element selection line L of the KANO model, it is determined that the user's preference for the i-th playback parameter at the current moment is not preferred, and the playback parameter needs to be adjusted.

[0040] When it is determined that the user's preference for the i-th playback parameter at the current moment is dislike, the degree of dislike is related to the sensitivity of the i-th playback parameter The distance to the element selection line L is proportional to the sensitivity of the i-th playback parameter. The greater the distance to the element selection line L, the greater the degree of dislike.

[0041] The playback parameter of the smart speaker in the i-th item is adjusted and controlled according to the adjustment strategy.

[0042] The present invention has the following beneficial effects:

[0043] Compared with the prior art, the present invention collects the user's adjustment behavior data on the smart speaker playback parameters during use, including the first behavior data for adjusting the equalizer, the second behavior data for adjusting the playback volume, the third behavior data for adjusting the sound focus point position, and the fourth behavior data for adjusting the surround effect, and obtains the user's usage time data after performing any of the above adjustment behaviors. The sampling mean of the usage time data is used as the classification basis, and the user's adjustment behavior of the first and second behavior data are classified into the following categories: The satisfaction degree after adjusting the playback parameters is classified into five levels from high to low, namely the first level, the second level, the third level, the fourth level and the fifth level. The playback parameters are used as user needs. According to the classification results of the satisfaction degree, the playback parameters are classified into user demand attributes using the KANO model, and the Better-Worse coefficient of the user for each user demand attribute is calculated to obtain the satisfaction influence SI value and the dissatisfaction influence DSI value. The formula is used to calculate the satisfaction influence SI value and the dissatisfaction influence DSI value. , calculate the sensitivity of each playback parameter, where, It is expressed as the sensitivity of the i-th playback parameter; It is expressed as the SI value of the satisfactory influence of the i-th playback parameter; It is expressed as the unsatisfactory influence DSI value of the i-th playback parameter to obtain the sensitivity of the i-th playback parameter. As a basis for evaluating the user's preference for this type of playback parameters at the current moment, an adjustment strategy for the playback parameters of this type is generated according to the evaluation result of the user's preference, and the playback parameters of the i-th item of the smart speaker are adjusted according to the adjustment strategy. Based on the autonomous analysis results of the actual needs and usage habits of individual users, automatic control of audio playback can be performed, thereby improving the intelligent control level of the smart speaker and improving the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 This is a flow chart of an audio playback control method for smart speakers of the present invention;

[0045] Figure 2 Screening quartile plot for the features of the KANO model. DETAILED DESCRIPTION

[0046] The present invention will be further described below in conjunction with specific implementation methods, wherein the accompanying drawings are only used for exemplary descriptions and represent only schematic diagrams rather than actual drawings, and should not be understood as limiting the present invention. In order to better illustrate the specific implementation methods of the present invention, some parts of the accompanying drawings may be omitted, enlarged or reduced, and do not represent the size of the actual product.

[0047] The specific implementation process of the technical solution of the present invention includes the following steps:

[0048] Step 1: Collecting data on the user's adjustment behavior of the smart speaker playback parameters during use;

[0049] The adjustment behavior data includes first behavior data for adjusting the equalizer; specifically, the first behavior data may be frequency (high frequency, low frequency, etc.) adjustment for the equalizer;

[0050] Second behavior data for adjusting the playback volume;

[0051] A third behavior data for adjusting the sound focus position;

[0052] The fourth line is data for adjusting the surround effect;

[0053] Step 2: Obtain the usage time data of the user after performing any adjustment behavior, and use the sampling mean of the usage time data as the classification basis to classify the user's The satisfaction degree after adjusting the playback parameters is classified into five levels from high to low, namely the first level, the second level, the third level, the fourth level and the fifth level;

[0054] The satisfaction classification process after adjusting the playback parameters includes the following steps:

[0055] Get the user in How long the user uses the smart speaker after performing the i-th adjustment behavior ;

[0056] Calculate the sample mean of the time users spend using the smart speaker after performing the i-th adjustment behavior ,in, , where The total number of times the user performs the i-th adjustment behavior;

[0057] Get the user The minimum usage time of the smart speaker after the ith adjustment behavior is performed and maximum value ;

[0058] Calculate the sample mean Minimum usage time The distance between and the sampling mean Maximum value of usage time The distance between , the calculation formulas are: ; ;

[0059] According to distance and distance Will The usage time data is divided into different level intervals, and the division principles are:

[0060] when ≤ ,and ≤ , then the usage time Divided into the third level interval, among which, For usage time With the sampling mean The distance between

[0061] when < ,and < ≤ , then the usage time Divided into the fourth level interval;

[0062] when < ,and < , then the usage time Divided into intervals up to the fifth level;

[0063] when > ,and ≤ ≤ , then the usage time Divided into the second level interval;

[0064] when > ,and > , then the usage time Divided into first-level intervals;

[0065] when > ,and > , then the usage time Divided into first-level intervals;

[0066] The level gradient of user satisfaction after adjusting the i-th playback parameter is determined based on the division results of the usage time data. The determination principle is:

[0067] When using the duration When divided into the first level interval, the level gradient of satisfaction is the first level;

[0068] When using the duration When divided into the first to second level interval, the level gradient of satisfaction is the second level;

[0069] When using the duration When divided into the third level interval, the level gradient of satisfaction is the third level;

[0070] When using the duration When divided into the fourth level interval, the level gradient of satisfaction is the fourth level;

[0071] When using the duration When divided into the fifth level interval, the level gradient of satisfaction is the fifth level.

[0072] Step 3: Taking playback parameters as user needs, and based on the classification results of satisfaction, use the KANO model to classify playback parameters into user demand attributes;

[0073] The attributes of user needs include four types: basic needs, expected needs, attractive needs, and indifferent needs. The classification process is as follows:

[0074] Construct forward and reverse problems, where the forward adjustment of any playback parameter is used as the forward problem, and the reverse adjustment is used as the reverse problem; the specific process is as follows:

[0075] When the playback parameter is the high frequency of the equalizer, whether the user is satisfied after the high frequency of the equalizer is increased is used as a positive question, and whether the user is satisfied after the high frequency of the equalizer is decreased is used as a reverse question, and the satisfaction after adjustment is classified based on the usage time data;

[0076] When the playback parameter is the playback volume, whether the user is satisfied after the playback volume is increased is used as a positive question, and whether the user is satisfied after the playback volume is decreased is used as a reverse question, and the satisfaction after adjustment is classified based on the usage time data;

[0077] When the playback parameter is the sound focus point position, whether the user is satisfied after the sound focus point position is increased is used as the positive question, and whether the user is satisfied after the sound focus point position is decreased is used as the reverse question, and the satisfaction degree after adjustment is classified based on the usage time data;

[0078] When the playback parameter is a surround effect, the surround effect is set to include mode 1, and whether the user is satisfied after the surround effect mode 1 is turned on is used as a positive question, and whether the user is satisfied after the surround effect mode 1 is turned off is used as a reverse question, and the satisfaction after adjustment is classified based on the usage time data; the above classification process is all referred to step 2, which is not repeated here;

[0079] Step 4: Statistic the satisfaction classification after adjustment, and construct a two-dimensional attribute classification matrix based on the statistical results, such as Figure 2 As shown, in the two-dimensional attribute attribution classification matrix, the attributes of the playback parameters are classified with the Worse value (i.e., the satisfactory influence SI value) as the horizontal coordinate and the Better value (i.e., the unsatisfactory influence DSI value) as the vertical coordinate. Specifically, they are classified into four quadrants, namely: A, O, I, and M. In addition, any point in the two-dimensional attribute attribution classification matrix corresponds to a unique Better value and Worse value.

[0080] Step 5: Calculate the Better-Worse coefficient of each user demand attribute to obtain the satisfaction influence SI value and the dissatisfaction influence DSI value; the calculation formulas are: ; ; In the formula, A, O, M, and I are the frequencies of attractive demand, expected demand, basic demand, and indifferent demand respectively;

[0081] Step 6: Using the formula , calculate the sensitivity of each playback parameter, where, It is expressed as the sensitivity of the i-th playback parameter; It is expressed as the SI value of the satisfactory influence of the i-th playback parameter; It is represented by the DSI value of the i-th playback parameter;

[0082] It should be noted that: Figure 2 The factor selection line L in the KANO model is the dividing line for screening out the demand factors that need to be improved. It is an arc with a point where both the Worse value and the Better value are 0 as the center and a radius of 0.707. The calculation formula of sensitivity is: Correspondingly, therefore, all points on the element selection line L have a sensitivity of 0, that is, the element selection line L is a sensitivity R i When the value of is 0;

[0083] When the point element in the two-dimensional attribute attribution classification matrix selects the left side of line L, for any point, the sensitivity R calculated by the corresponding unique Worse value and Better value i The values ​​of are all negative, which indicates that the i-th playback parameter does not need to be adjusted;

[0084] When the point element in the two-dimensional attribute attribution classification matrix is ​​selected on line L, the sensitivity R calculated by the corresponding Worse value and Better value i The values ​​of are all 0;

[0085] When a point in the two-dimensional attribute attribution classification matrix is ​​on the right side of the element selection line L in the figure, it is determined that the user's preference for the i-th playback parameter at the current moment is not preferred, and the playback parameter needs to be adjusted. For any point on the right side of the element selection line L in the two-dimensional attribute attribution classification matrix, the sensitivity R calculated by the corresponding unique Worse value and Better value is i The values ​​of are all positive numbers, and for these points located on the right side of the element selection line L, the farther they are from the element selection line L, the higher the sensitivity R in the recorded content. i The greater the distance to the element selection line L, the greater the degree of dislike, and the more this playback parameter needs to be adjusted first.

[0086] Step 7: To obtain the sensitivity of the i-th playback parameter As the evaluation basis of the user's preference for this type of playback parameters at the current moment, an adjustment strategy for this type of playback parameters is generated according to the evaluation result of the user's preference; the preference evaluation basis is:

[0087] When the sensitivity of the i-th playback parameter When the value is on the left side of the element selection line L of the KANO model, it is determined that the user has The preference level of the playback parameter is "favorite", and the playback parameter does not need to be adjusted;

[0088] When the sensitivity of the i-th playback parameter When the value is on the right side of the factor selection line L of the KANO model, it is determined that the user's preference for the i-th playback parameter at the current moment is not preferred, and the playback parameter needs to be adjusted. When it is determined that the user's preference for the i-th playback parameter at the current moment is not preferred, the degree of dislike is related to the sensitivity of the i-th playback parameter. The distance to the element selection line L is proportional to the sensitivity of the i-th playback parameter. The greater the distance to the element selection line L, the greater the degree of dislike, and the more this playback parameter needs to be adjusted first.

[0089] Step 8: Adjust and control various playback parameters of the smart speaker according to the generated adjustment strategy.

[0090] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. An audio playback control method for smart speakers, characterized in that: include: Collecting the user's behavior data on adjusting the playback parameters of the smart speaker during use, including first behavior data on adjusting the equalizer, second behavior data on adjusting the playback volume, third behavior data on adjusting the sound focus point position, and fourth behavior data on adjusting the surround effect; Obtaining usage time data of the user after performing any of the adjustment behaviors, and using the sampled mean of the usage time data as a classification basis, classifying the user's satisfaction after adjusting the playback parameter of the i-th item into five levels from high to low, namely, the first level, the second level, the third level, the fourth level, and the fifth level; Taking the playback parameters as user needs, taking the positive adjustment of the playback parameters as a positive problem and the negative adjustment as a negative problem, according to the classification results of the satisfaction degree, using the KANO model to classify the playback parameters into user demand attributes, calculating the Better-Worse coefficient of the user for each of the user demand attributes, and obtaining the satisfaction influence SI value and the dissatisfaction influence DSI value; Using the formula , calculate the sensitivity of each playback parameter, where, It is expressed as the sensitivity of the i-th playback parameter; It is expressed as the SI value of the satisfactory influence of the i-th playback parameter; It is represented by the DSI value of the i-th playback parameter; To obtain the sensitivity of the i-th playback parameter As a basis for evaluating the user's preference for the playback parameters at the current moment, an adjustment strategy for the playback parameters of the type is generated according to the evaluation result of the user's preference; The playback parameter of the smart speaker in the i-th item is adjusted and controlled according to the adjustment strategy.

2. The audio playback control method for smart speakers according to claim 1, characterized in that: The satisfaction classification process after adjusting the playback parameters includes the following steps: Get the user's usage time of the smart speaker after the user performs the adjustment behavior described in item i for the kth time ; Calculate the sample mean of the time the user uses the smart speaker after performing the adjustment behavior described in item i ,in, , where K is the total number of times the user performs the adjustment behavior described in item i; Get the minimum usage time of the smart speaker after the user performs the adjustment behavior described in item i for k times and maximum value ; Calculate the sample mean Minimum usage time The distance between and the sampling mean Maximum value of usage time The distance between , the calculation formulas are: ; ; According to distance and distance Divide the K usage duration data into different level intervals; when ,and , then the usage time Divided into first-level intervals; The level gradient of the user's satisfaction after adjusting the playback parameter described in the i-th item is determined according to the division result of the usage time data.

3. The audio playback control method for smart speakers according to claim 1, characterized in that: The evaluation basis of the user's preference for the i-th playback parameter at the current moment is: When the sensitivity of the i-th playback parameter When the value is on the left side of the factor selection line L of the KANO model, it is determined that the user's preference for the i-th playback parameter at the current moment is favorable, and the playback parameter does not need to be adjusted; When the sensitivity of the i-th playback parameter When the value is on the right side of the factor selection line L of the KANO model, it is determined that the user's preference for the i-th playback parameter at the current moment is not favorable, and the playback parameter needs to be adjusted; The factor selection line L is the dividing line for screening out the demand factors that need to be improved. It is an arc with a point where the Worse value and the Better value are both 0 as the center and a radius of 0.

707.

4. The audio playback control method for smart speakers according to claim 3, characterized in that: When it is determined that the user's preference for the i-th playback parameter at the current moment is not preferred, the sensitivity of the i-th playback parameter is The distance to the element selection line L is used to determine the degree of dislike of the user for the playback parameter. The determination principle is: the degree of dislike is proportional to the sensitivity of the i-th playback parameter. It is proportional to the distance to the feature selection line L.

5. The audio playback control method for smart speakers according to claim 1, characterized in that: The attributes of user needs include four types: basic needs, expected needs, attractive needs, and indifferent needs; The calculation formulas for the satisfactory influence SI value and the unsatisfactory influence DSI value are: ; ;In the formula, A, O, M, and I are the frequencies of attractive demand, expected demand, basic demand, and indifferent demand respectively.

6. The audio playback control method for smart speakers according to claim 2, characterized in that: The principles for dividing the level intervals of usage duration data are as follows: when ,and , then the usage time Divided into the third level interval, among which, For usage time With the sampling mean The distance between when , then the usage time Divided into the fourth level interval; when , then the usage time Divided into intervals up to the fifth level; when , then the usage time Divided into the second level interval; when , then the usage time Divided into first level intervals.

7. The audio playback control method for smart speakers according to claim 2, characterized in that: The level of satisfaction is determined by the following principles: When using the duration When divided into the first level interval, the level gradient of satisfaction is the first level; When using the duration When divided into the first to second level interval, the level gradient of satisfaction is the second level; When using the duration When divided into the third level interval, the level gradient of satisfaction is the third level; When using the duration When divided into the fourth level interval, the level gradient of satisfaction is the fourth level; When using the duration When divided into the fifth level interval, the level gradient of satisfaction is the fifth level.

Images