5G wireless broadband equipment with call function

Abstract

The invention relates to the technical field of broadband equipment, and discloses 5G wireless broadband equipment with a call function, a voice gateway of the invention can better match the identity of a user with sound information, and after matching is completed, the serial number of a starting area is obtained based on the state information, position information and historical data of the user, and the starting area is started according to the serial number of the starting area. Compared with a traditional control mode, the method has the advantages that only one intelligent loudspeaker box responds to each time after correct starting, and the intelligent loudspeaker boxes are screened based on historical data of a user, so that the intelligent loudspeaker boxes are more convenient to use. And the system is more suitable for the living habit of a user and is more intelligent.

Description

Technical Field

[0001] This application relates to the technical field of broadband devices, and in particular, to a 5G wireless broadband device with a call function. Background Art

[0002] There are various wireless broadband devices, and a wireless gateway is one of them. A wireless gateway with a call function is also called a voice gateway and is also a type of smart gateway. Nowadays, the use of smart gateways has become an indispensable part of smart homes. However, for the transformation of smart homes, it is more convenient to use brand smart speakers that have already built an ecosystem.

[0003] The problem is that smart speakers exist independently. In a small-sized home environment, on the one hand, it is easy for users to wake up multiple smart speakers of the same brand at the same time. On the other hand, the same type of electrical appliances controlled by smart speakers, such as air conditioners and lights, are also prone to accidental startup problems. Currently, there is a technical solution to control multiple lower-level smart gateways through an upper-level smart gateway. However, it still controls different electrical appliances and different smart speakers by different naming or prefix editing, and then controls them through voice recognition, or independently starts smart speakers based on room positioning. It is difficult to solve the problem of multiple smart speakers starting at the same time and the problem that the control of multiple same type of electrical appliances is not intelligent enough. Often, users need to call out air conditioners with different prefixes or names, which is more troublesome to use.

[0004] In view of this, the present invention proposes a 5G wireless broadband device with a call function, which improves the intelligent use experience of smart speakers controlled by a voice gateway through the analysis of the user's identity and status. Summary of the Invention

[0005] In order to solve the problems of multiple smart speakers starting at the same time and the problem that the control of multiple same type of electrical appliances is not intelligent enough at the same time, this application provides a 5G wireless broadband device with a call function.

[0006] In a first aspect, this application provides a 5G wireless broadband device with a call function, adopting the following technical solution: A 5G wireless broadband device with a call function includes a voice gateway and a plurality of smart speakers controlled by the voice gateway. An infrared positioning unit is provided in the smart speakers. The smart speakers are respectively arranged in a first area and a plurality of second areas. The plurality of second areas are all connected to the first area. A common area is provided between the first area and the plurality of second areas, and the first area and the plurality of second areas are numbered respectively; The voice gateway further includes a voice module, an identity matching module, and a startup judgment module; The voice module is linked to a number of smart speakers and performs semantic analysis on the voice information received by the smart speakers to obtain control commands; The identity matching module includes a voice recording unit and an identity matching unit. The voice recording unit is used to record the voice information of the user, and the identity matching unit is used to match the personnel in the first area and the second area with the recorded voice information; The judgment module includes a status analysis unit and a start judgment unit. The status analysis unit judges whether the user is in a resting state and obtains the user's location information through the infrared image obtained by the infrared positioning unit. The start judgment unit obtains the user status information and location information of the status analysis unit after receiving the control command, and obtains the number of the start area based on the historical data of the user. The start area is one of the first area and a number of second areas, and the corresponding intelligent device in the start area executes the control command.

[0007] Through the above technical solutions, a solution for intelligent control of smart speakers and electrical appliances controlled by smart speakers under the control of a voice gateway is provided. Specifically, the voice gateway of the present invention can better match the user's identity with the voice information. After the matching is completed, based on the user status information, location information, and historical data, the number of the start area is obtained, and the smart speaker in the first area or the second area corresponding to the number is started through this number to complete the control command. Compared with the traditional control method, only one smart speaker responds at a time after the correct start of this method, and this smart speaker is a smart speaker screened based on the historical data of the user, which is more in line with the user's living habits and is also more intelligent.

[0008] Optionally, the process of matching the personnel in the first area and the second area with the recorded voice information includes: After receiving the voice information, obtain the infrared image of the infrared transmission positioning unit and judge the number of people in the first area and the second area; If the number of people is one, match the person with the voice information and record the body posture data of the person's contour corresponding to the infrared image; If the number of people is not one, skip the current matching process and mark the current voice information as unmatched. In the unmatched state, each time the corresponding voice information is received, the process of matching the person with the recorded voice information is performed once.

[0009] Optionally, the process of judging whether the user is in a resting state includes: Obtain the infrared image of the infrared transmission positioning unit and input the user contour image in the infrared image into the trained neural network model, and the neural network model determines whether the user is in a resting state; The trained neural network model uses an image recognition model, and various resting postures of the human body are used as input objects during training.

[0010] Optionally, the process of obtaining the user status information and location information of the status analysis unit and obtaining the number of the activation area based on the historical data of the user includes: If the user status information indicates that the user is in a resting state, obtain the number of the location where the user is located based on the location information, and use this number as the number of the activation area; If the user is in the common area between the first area and the second area, obtain the face orientation of the user through the infrared image of the infrared positioning unit, and determine the smart speaker pointed to by the face orientation, and use the numbers of the first area and the second area where the smart speaker is located as the numbers of the activation area; If the user status information indicates that the user is not in a resting state, obtain the user location information, obtain a judgment coefficient based on the historical data of the user, and then determine the number of the activation area through the judgment coefficient.

[0011] Through the above technical solution: a process of determining the number of the activation area through the judgment coefficient is provided. The number of the activation area of the present invention incorporates the moving direction of the user on the basis of the user location information and usage habits for reference. Especially for some transition time periods where the user's usage habits are not clear enough, the moving direction can be used to intelligently perceive the user's intention, realizing a more intelligent device control process.

[0012] Optionally, the process of obtaining the historical data of the user includes: Based on the time series, count the number of times the user stays in each area within a preset time interval; The location information is obtained through infrared images acquired once per second. The first area and the second area corresponding to more than half of the location information in the location information corresponding to the user infrared images acquired within the preset time interval are determined as the staying areas, and it is recorded that the user stays once. The number of stays in each preset time interval is the historical data.

[0013] Optionally, the process of obtaining the judgment coefficient includes: After receiving the control command, set the detection time. During the detection time, detect the moving direction of the user multiple times, and then obtain the number corresponding to the location where the user is located within the current time interval and the staying probability P of the user in the first area and the second area corresponding to the number. Then, through the formula Obtain the judgment coefficient Ju, where , is the total number of detections of the user's moving direction during the detection time. If the user's location is in the first area, then Take 0. If the user's location is in the first area, then Take 1. is the number of detections of the exit whose moving direction points to the user's location during the detection time. The exit of the user's location refers to the exit of the first area or the second area where the user is located. is the ratio of the current time interval duration to the total duration of the user's stay at the current location within the current time interval. is a preset base value. is about a piecewise function of the value.

[0014] Optionally, the process of determining the number of the activation area by the judgment coefficient includes: Set a judgment interval , and compare the judgment coefficient with the judgment interval, where ; If the judgment coefficient falls within the judgment interval, set the number corresponding to the user's location as the number of the activation area; If the judgment coefficient does not fall within the judgment interval, obtain the number corresponding to the maximum value of the user's stay probability in the historical data of the current preset time interval and set it as the number of the activation area; After determining the activation number, perform error correction detection according to the instruction content within the error correction time.

[0015] Optionally, the error correction detection process includes: Within the error correction time, if two consecutive instructions of the same type with different contents are received, the same type refers to commands for the same intelligent device, and different contents refer to start or stop, then judge that there is a start error, and then record the preset time interval where the error correction time is located as the error correction interval and record the previous judgment coefficient adjacent to the error correction time; Generate an error correction interval based on the recorded judgment coefficient.

[0016] Optionally, the error correction detection process further includes: Within the next preset time interval recorded as the error correction interval, if the obtained judgment coefficient falls within the error correction interval, only obtain the number of the activation area without controlling the intelligent speaker; If the numbers of the activation areas obtained in at least two consecutive error correction intervals are the same as the number of the actual activation area, delete the current error correction interval from the record.

[0017] Through the above technical solution, an error correction detection process is provided. The error correction detection process of the present invention is a supplement in the case where the user's intention is contrary to the behavior and habit. Specifically, if the user's intention is contrary to the behavior and habit, resulting in an incorrect intelligent judgment, the error behavior will be recorded, and when a similar behavior occurs at the next same time node, the control of the voice gateway will be stopped until the error correction interval is deleted from the record. The error correction mechanism of the present invention can avoid excessive speculation of the user's intention by intelligent control, and on the basis of providing sufficient intelligent experience, retain sufficient autonomous control process for the user.

[0018] Optionally, the voice gateway is connected to a number of smart speakers via a 5G wireless network.

[0019] In summary, the present application includes at least one of the following beneficial technical effects: The voice gateway of the present invention can better match the user's identity with the voice information. After the matching is completed, based on the user's status information, location information, and historical data, the number of the activation area is obtained, and the smart speaker in the first area or the second area corresponding to the number is activated through this number to complete the control command. Compared with the traditional control method, after the correct activation, only one smart speaker responds at a time, and this smart speaker is the smart speaker screened based on the user's historical data, which is more in line with the user's living habits and is also more intelligent.

[0020] The number of the activation area of the present invention incorporates the user's moving direction for reference on the basis of the user's location information and usage habits. Especially for some transition time periods, the user's usage habits are not clear enough, and the user's intention can be intelligently perceived through the moving direction, realizing a more intelligent device control process.

[0021] The error correction detection process of the present invention is a supplement in the case where the user's intention is contrary to the behavior and habit. If the user's intention is contrary to the behavior and habit, resulting in an incorrect intelligent judgment, the error behavior will be recorded, and when a similar behavior occurs at the next same time node, the control of the voice gateway will be stopped until the error correction interval is deleted from the record. The error correction mechanism of the present invention can avoid excessive speculation of the user's intention by intelligent control, and on the basis of providing sufficient intelligent experience, retain sufficient autonomous control process for the user. Brief Description of the Drawings

[0022] Figure 1 It is a schematic diagram of the internal module relationship of the voice gateway of the present invention.

[0023] Figure 2 It is a schematic diagram of the composition of the identity matching module of the present invention.

[0024] Figure 3It is a schematic diagram of the composition of the startup judgment module of the present invention. Specific implementation mode

[0025] The following details the implementation modes of the present application. Examples of the implementation modes are shown in the accompanying drawings.

[0026] In the description of this specification, the description with reference to terms such as "certain implementation modes", "one implementation mode", "some implementation modes", "schematic implementation modes", "examples", "specific examples", or "some examples" means that the specific features, structures, materials, or characteristics described in connection with the implementation mode or example are included in at least one implementation mode or example of the present application. In this specification, the schematic expressions of the above terms do not necessarily refer to the same implementation mode or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more implementation modes or examples.

[0027] The embodiments of the present application disclose a 5G wireless broadband device with a call function. Refer to Figures 1-3 , including.

[0028] A voice gateway and several smart speakers controlled by the voice gateway. An infrared positioning unit is provided in the smart speakers. The smart speakers are respectively arranged in a first area and several second areas. The several second areas are all connected to the first area. A common area is provided between the first area and the second areas, and the first area and the several second areas are numbered respectively. The division method of the first area and the second areas is not limited. It can be modeled based on the infrared image or divided according to the movement range of the people in the infrared image. The application scenario of the present invention is the situation of one living room and multiple bedrooms in a residential building. The first area is the living room, and the second areas are multiple living rooms; The voice gateway further includes a voice module, an identity matching module, and a startup judgment module; The voice module is linked to several smart speakers and performs semantic analysis on the voice information received by the smart speakers to obtain control commands; The identity matching module includes a voice recording unit and an identity matching unit. The voice recording unit is used to record the voice information of the user. The identity matching unit is used to match the people in the first area and the second areas with the recorded voice information. When the matching cannot be completed, the voice gateway does not control the smart speakers, and the generated control commands are independently fed back by the smart speakers; The judgment module includes a status analysis unit and a startup judgment unit. The status analysis unit judges whether the user is in a resting state and obtains the user's location information through the infrared image obtained by the infrared positioning unit. After receiving the control command, the startup judgment unit obtains the user status information and location information of the status analysis unit, and obtains the number of the startup area based on the user's historical data. The startup area is one of the first area and several second areas, and the corresponding intelligent device in the startup area executes the control command. The intelligent devices include intelligent speakers, intelligent lamps, etc.

[0029] In this embodiment, a solution for intelligent control of an intelligent speaker and the appliances controlled by the intelligent speaker under the control of a voice gateway is provided. Specifically, the voice gateway of the present invention can match the user's identity with the voice information. After the matching is completed, based on the user status information, location information and historical data, the number of the startup area is obtained, and the intelligent speaker in the first area or the second area corresponding to the number is started to complete the control command. Compared with the traditional control method, only one intelligent speaker responds at a time after the correct startup of this method, and the intelligent speaker is the one screened based on the user's historical data, which is more in line with the user's living habits and is also more intelligent.

[0030] The process of matching the personnel in the first area and the second area with the recorded voice information includes: After receiving the voice information, obtain the infrared image of the infrared transmission positioning unit and judge the number of people in the first area and the second area; If the number of people is one, match the person with the voice information, and record the body posture data of the person's contour corresponding to the infrared image. The body posture data includes height, hand length, leg length, etc.; If the number of people is not one, skip the current matching process and mark the current voice information as unmatched. In the unmatched state, each time the corresponding voice information is received, a process of matching the personnel with the recorded voice information is performed.

[0031] The process of judging whether the user is in a resting state includes: Obtain the infrared image of the infrared transmission positioning unit and input the user contour image in the infrared image into the trained neural network model, and the neural network model determines whether the user is in a resting state; The trained neural network model uses a picture recognition model, and various human resting postures are used as input objects during training.

[0032] The process of obtaining the user status information and location information of the status analysis unit and obtaining the number of the startup area based on the user's historical data includes: If the user status information indicates that the user is in a rest state, obtain the number of the location where the user is located based on the location information, and use this number as the number of the activation area; If the user is in the common area between the first area and the second area, obtain the face orientation of the user through the infrared image of the infrared positioning unit, and determine the smart speaker pointed to by the face orientation. The numbers of the first area and the second area where the smart speaker is located are used as the numbers of the activation area. The process of determining the smart speaker pointed to by the face orientation can be as follows: obtain the connection lines of the positions of the smart speakers in the first area and the second area adjacent to the location where the user is located in the common area, and then compare the direction vector of the face orientation with the two connection lines, and use the smart speaker corresponding to the connection line with a smaller included angle as the smart speaker pointed to by the face orientation. Of course, methods such as the optical flow algorithm can also be used for determination; If the user status information indicates that the user is not in a rest state, obtain the user location information, obtain the judgment coefficient based on the user's historical data, and then determine the number of the activation area through the judgment coefficient.

[0033] In this embodiment, a process of determining the number of the activation area through the judgment coefficient is provided. The number of the activation area of the present invention incorporates the moving direction of the user on the basis of the user location information and usage habits for reference. Especially for some transition time periods where the user's usage habits are not clear enough, the moving direction can be used to intelligently perceive the user's intention and achieve a more intelligent device control process.

[0034] In one embodiment, the process of obtaining the user's historical data includes: Based on the time series, count the number of times the user stays in each area within a preset time interval. The preset time interval is not limited and can be half an hour or one hour. Based on the time series means that the data is statistically analyzed on a weekly or monthly basis; The location information is obtained through the infrared image acquired once per second. The first area and the second area corresponding to more than half of the location information in the location information corresponding to the user infrared image acquired within the preset time interval are determined as the stay areas, and it is recorded that the user stays once. The number of stays in each preset time interval is the historical data.

[0035] In one embodiment, the process of obtaining the judgment coefficient includes: After receiving the control command, set the detection time. During the detection time, detect the moving direction of the user multiple times, and then obtain the number corresponding to the location where the user is located within the current time interval and the stay probability P of the user in the first area and the second area corresponding to the number. Then, through the formula Obtain the judgment coefficient Ju, where , is the total number of detections of the user's moving direction during the detection time. If the user's location is in the first area, then take 0. If the user's location is in the first area, then take 1. is the number of detections that the moving direction points to the exit of the user's location during the detection time. The exit of the user's location refers to the exit of the first area or the second area where the user is located. The determination of the exit can be realized based on the infrared image through the doorframe recognition algorithm. The doorframe recognition algorithm can be one of the neural network algorithm or the image recognition algorithm. is the ratio of the current time interval duration to the total duration that the user stays at the current location within the current time interval. is a preset base value, and the value range is , preferably 1. is about a piecewise function of the value. When and only when is greater than the preset threshold, the output is 1. Otherwise, the output is 0. The selection range of the preset threshold is , preferably in this embodiment, the preset value is selected as 1, that is, only when all moving directions point to the exit of the user's location the output is 1. The preset threshold is set based on the user's movement habits.

[0036] The process of determining the number of the activation area through the judgment coefficient includes: Set the judgment interval , and compare the judgment coefficient with the judgment interval, where ; If the judgment coefficient falls within the judgment interval, then set the number corresponding to the user's location as the number of the activation area; If the judgment coefficient does not fall within the judgment interval, then obtain the number corresponding to the maximum value of the user's stay probability in the historical data of the current preset time interval and set it as the number of the activation area; After determining the activation number, perform error correction detection according to the instruction content within the error correction time. The error correction time is not limited, generally within 5 - 10 s after the completion of the activation instruction.

[0037] The error correction detection process includes: Within the error correction time, if two consecutive instructions of the same type with different contents are received, the same type refers to the command for the same intelligent device, and the different contents refer to start or stop, then it is judged that the start is incorrect, and then record the preset time interval where the error correction time is located as the error correction interval and record the judgment coefficient adjacent to the error correction time; Generate an error correction interval based on the recorded judgment coefficient, preferably ±0.1 of the judgment coefficient.

[0038] In one embodiment, the error correction detection process further includes: Within a preset time interval that is the next record for the error correction interval, if the obtained judgment coefficient falls within the error correction range, only obtain the number of the activation area without controlling the smart speaker, and let the smart speaker give feedback independently; If the numbers of the activation areas obtained in at least two consecutive error correction intervals are the same as the number of the actual activation area, delete the current error correction interval from the record.

[0039] In this embodiment, an error correction detection process is provided. The error correction detection process of the present invention is a supplement for the situation where the user's intention is contrary to their behavior and habits. Specifically, if the user's intention is contrary to their behavior and habits, resulting in an incorrect intelligent judgment, the error behavior will be recorded, and the control of the voice gateway will be stopped when a similar behavior occurs at the next same time node until the error correction interval is deleted from the record. The error correction mechanism of the present invention can avoid excessive speculation of the user's intention by intelligent control, and on the basis of providing a sufficient intelligent experience, retain sufficient independent control process for the user.

[0040] In one embodiment, the voice gateway is connected to a number of smart speakers using a 5G wireless network.

[0041] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limitations to the present application. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present application.

Claims

1. A 5G wireless broadband device with a call function, including a voice gateway and several smart speakers controlled by the voice gateway. An infrared positioning unit is provided in each smart speaker. It is characterized in that the smart speakers are respectively arranged in a first area and several second areas. The several second areas are all connected to the first area. A public area is arranged between the first area and the second areas, and the first area and the several second areas are numbered respectively; the voice gateway further includes a voice module, an identity matching module and a startup judgment module; the voice module is linked to several smart speakers and performs semantic analysis on the voice information received by the smart speakers to obtain control commands; the identity matching module includes a voice recording unit and an identity matching unit. The voice recording unit is used to record the voice information of the user, and the identity matching unit is used to match the personnel in the first area and the second areas with the recorded voice information; the judgment module includes a status analysis unit and a startup judgment unit. The status analysis unit judges whether the user is in a resting state and obtains the user's position information through the infrared image obtained by the infrared positioning unit. The startup judgment unit obtains the user's status information and position information of the status analysis unit after receiving the control command, and obtains the number of the startup area based on the user's historical data. The startup area is one of the first area and the several second areas, and the corresponding smart device in the startup area executes the control command.

2. The 5G wireless broadband device with a call function according to claim 1, wherein The process of matching the personnel in the first area and the second areas with the recorded voice information includes: after receiving the voice information, obtaining the infrared image of the infrared positioning unit and judging the number of people in the first area and the second areas; if the number of people is one, matching the person with the voice information and recording the body posture data of the person's contour corresponding to the infrared image; if the number of people is not one, skipping the current matching process and marking the current voice information as an unmatched state. In the unmatched state, each time the corresponding voice information is received, the process of matching the personnel with the recorded voice information is performed once.

3. The 5G wireless broadband device with a call function according to claim 1, characterized in that, The process of judging whether the user is in a resting state includes: obtaining the infrared image of the infrared positioning unit and inputting the user's contour image in the infrared image into a trained neural network model, and the neural network model determines whether the user is in a resting state; the trained neural network model adopts a picture recognition model, and various resting postures of the human body are used as input objects during training.

4. The 5G wireless broadband device with a call function according to claim 1, characterized in that, The process of obtaining the user's status information and position information of the status analysis unit and obtaining the number of the startup area based on the user's historical data includes: if the user's status information is in a resting state, obtaining the number of the location where the user is located based on the position information, and using this number as the number of the startup area; if the user is in the public area between the first area and the second areas, obtaining the user's face orientation through the infrared image of the infrared positioning unit, and judging the smart speaker pointed to by the face orientation, and making the numbers of the first area and the second area where the smart speaker is located be the number of the startup area; If the user status information indicates that the user is not in a resting state, obtain the user location information, obtain a judgment coefficient based on the user's historical data, and then determine the number of the activation area through the judgment coefficient.

5. The 5G wireless broadband device with a call function according to claim 4, characterized in that, The process of obtaining the user's historical data includes: Based on the time series, count the number of times the user stays in each area within a preset time interval; The location information is obtained through an infrared image acquired once per second. The first area and the second area corresponding to more than half of the location information in the location information corresponding to the user infrared images acquired within the preset time interval are determined as the staying areas, which is recorded as the user staying once, and the number of stays in each preset time interval is the historical data.

6. The 5G wireless broadband device with a call function according to claim 5, characterized in that, The process of obtaining the judgment coefficient includes: After receiving the control command, set the detection time. During the detection time, detect the moving direction of the user multiple times, and then obtain the number corresponding to the user's location within the current time interval and the residence probabilities P of the user in the first area and the second area corresponding to the number. Then, through the formula obtain the judgment coefficient Ju, where , is the total number of detections of the user's moving direction during the detection time. If the user's location is in the first area, then takes 0. If the user's location is in the first area, then takes 1, is the number of detections of the moving direction pointing to the exit of the user's location during the detection time, is the ratio of the duration of the current time interval to the total residence duration of the user at the location within the current time interval, is a preset base value, is a piecewise function regarding the value.

7. The 5G wireless broadband device with a call function according to claim 6, characterized in that, The process of determining the number of the activation area through the judgment coefficient includes: Set a judgment interval , and compare the judgment coefficient with the judgment interval, where ; If the judgment coefficient falls within the judgment interval, set the number corresponding to the user's location as the number of the activation area; If the judgment coefficient does not fall within the judgment interval, obtain the number corresponding to the maximum user staying probability in the historical data of the current preset time interval and set it as the number of the activation area; After determining the activation number, perform error correction detection according to the instruction content within the error correction time.

8. The 5G wireless broadband device with a call function according to claim 7, characterized in that, The error correction detection process includes: Within the error correction time, if two consecutive instructions of the same type with different contents are received, it is determined that the activation is incorrect, then record the preset time interval where the error correction time is located as the error correction interval and record the judgment coefficient adjacent to the error correction time; Generate an error correction interval based on the recorded judgment coefficient.

9. The 5G wireless broadband device with a call function according to claim 8, characterized in that, The error correction detection process further includes: Within the next preset time interval recorded as the error correction interval, if the obtained judgment coefficient falls within the error correction interval, only obtain the number of the activation area without controlling the smart speaker; If the numbers of the activation areas obtained in at least two consecutive error correction intervals are the same as the number of the actual activation area, delete the current error correction interval from the record.

10. The 5G wireless broadband device with a call function according to claim 1, characterized in that, The voice gateway is connected to a number of smart speakers using a 5G wireless network.

Images