In-vehicle voice broadcast method, device and equipment and storage medium

Abstract

The invention relates to the technical field of man-machine conversation, and discloses an in-vehicle voice broadcasting method, device and equipment and a storage medium, and the method comprises the steps: carrying out the semantic analysis of each current to-be-broadcasted voice instruction when a plurality of current to-be-broadcasted voice instructions are received; selecting a corresponding target TTS model from a preset model pool according to each semantic analysis result, wherein a plurality of preset TTS models are arranged in the preset model pool; and broadcasting the corresponding current to-be-broadcasted voice instruction through each target TTS model. Compared with the prior art that only one TTS model is configured, a plurality of preset TTS models can be set, and the corresponding target TTS model is selected from the preset TTS models for broadcasting, so that simultaneous broadcasting of a plurality of current to-be-broadcasted voice instructions can be supported, the broadcasting timeliness is improved, and the user experience is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of human-computer dialogue, and in particular to an in-vehicle voice broadcasting method, device, equipment and storage medium. Background Art

[0002] At present, the car computer in the car is generally equipped with a text-to-speech (TTS) model, which converts the text information to be output by the application software (Application, APP) in the car computer into voice for broadcast, such as navigation broadcast, car fault broadcast, etc.

[0003] However, since only one TTS model is configured in the existing car computer, when multiple APPs need to be broadcast, they are generally broadcast according to priority, but some broadcasts are still not timely, resulting in poor user experience.

[0004] The above contents are only used to assist in understanding the technical solution of the present invention and do not constitute an admission that the above contents are prior art. Summary of the invention

[0005] The main purpose of the present invention is to provide a method, device, equipment and storage medium for in-vehicle voice broadcasting, aiming to solve the technical problem that in the prior art, only one TTS model is configured to broadcast according to priority, resulting in some untimely broadcasts and poor user experience.

[0006] To achieve the above object, the present invention provides a method for in-vehicle voice broadcasting, the method comprising the following steps:

[0007] When receiving a plurality of voice instructions to be currently broadcasted, performing semantic analysis on each of the voice instructions to be currently broadcasted;

[0008] Selecting a corresponding target TTS model from a preset model pool according to each semantic analysis result, wherein the model pool includes a plurality of preset TTS models;

[0009] The corresponding current voice command to be broadcasted is broadcasted through each of the target TTS models.

[0010] Optionally, the step of selecting a corresponding target TTS model from a preset model pool according to each semantic analysis result includes:

[0011] Selecting candidate TTS models from a preset model pool according to each semantic analysis result, and obtaining a distribution map of supported instruction ranges corresponding to each candidate TTS model;

[0012] Draw a global instruction range distribution map based on each of the supported instruction range distribution maps;

[0013] Determine the position information of each of the semantic analysis results in the global instruction range distribution map;

[0014] A target TTS model is selected from the candidate TTS models based on the position information.

[0015] Optionally, the step of selecting a target TTS model from the candidate TTS models based on the location information includes:

[0016] Generate a model combination list according to each of the candidate TTS models based on the location information;

[0017] Determine the quantity information of the voice instructions to be broadcasted currently, and determine the target model combination from the model combination list based on the quantity information;

[0018] A target TTS model is selected from the candidate TTS models according to the target model combination.

[0019] Optionally, the step of determining a target model combination from the model combination list based on the quantity information includes:

[0020] Screening the model combination list based on the quantity information to obtain a preliminary screening combination list;

[0021] Acquire the timbre information corresponding to each candidate TTS model in the preliminary screening combination list, and determine the current voice command to be broadcast corresponding to each candidate TTS model;

[0022] Performing timbre similarity analysis on each of the currently to-be-broadcasted voice commands based on each of the timbre information;

[0023] The target model combination is determined from the preliminary screening combination list according to the analysis results.

[0024] Optionally, the step of broadcasting the corresponding current voice instruction to be broadcasted through each of the target TTS models includes:

[0025] Determine the broadcast object of each of the currently to-be-broadcasted voice commands, and determine the broadcast parameters corresponding to each of the target TTS models based on each of the broadcast objects, wherein the broadcast parameters include a broadcast speed, a broadcast speaker, and a broadcast order;

[0026] The corresponding current voice instruction to be broadcasted is broadcasted through each of the target TTS models according to the corresponding broadcasting parameters.

[0027] Optionally, the step of determining the announcement parameters corresponding to each of the target TTS models based on each of the announcement objects includes:

[0028] Determining the importance of each of the currently to-be-announced voice instructions relative to the corresponding announcement object;

[0029] The broadcast parameters corresponding to each of the target TTS models are determined based on the importance.

[0030] Optionally, the step of determining the announcement parameters corresponding to each of the target TTS models based on the importance level includes:

[0031] Determining the broadcasting loudspeaker corresponding to each of the currently to-be-broadcasted voice instructions based on the importance;

[0032] Determine the broadcast speed and broadcast order corresponding to each of the broadcast speakers according to the importance;

[0033] The announcement parameters are generated according to each of the announcement speakers, each of the announcement speeds, and each of the announcement sequences.

[0034] In addition, to achieve the above-mentioned purpose, the present invention also proposes an in-vehicle voice broadcasting device, the device comprising:

[0035] A semantic analysis module, for performing semantic analysis on each of the currently to-be-broadcasted voice instructions when receiving a plurality of currently to-be-broadcasted voice instructions;

[0036] A model selection module, used to select a corresponding target TTS model from a preset model pool according to each semantic analysis result, wherein the model pool is provided with a plurality of preset TTS models;

[0037] The voice broadcast module is used to broadcast the corresponding current voice instruction to be broadcast through each of the target TTS models.

[0038] In addition, to achieve the above-mentioned purpose, the present invention also proposes an in-vehicle voice broadcast device, which includes: a memory, a processor, and an in-vehicle voice broadcast program stored in the memory and runnable on the processor, and the in-vehicle voice broadcast program is configured to implement the steps of the in-vehicle voice broadcast method described above.

[0039] In addition, to achieve the above-mentioned purpose, the present invention also proposes a storage medium, on which an in-vehicle voice broadcast program is stored. When the in-vehicle voice broadcast program is executed by a processor, the steps of the in-vehicle voice broadcast method described above are implemented.

[0040] The present invention provides a method, device, equipment and storage medium for in-vehicle voice broadcasting. The method performs semantic analysis on each of the current voice instructions to be broadcasted when receiving multiple current voice instructions to be broadcasted; selects a corresponding target TTS model from a preset model pool according to each semantic analysis result, wherein the model pool is provided with multiple preset TTS models; and broadcasts the corresponding current voice instructions to be broadcasted through each target TTS model. Because the present invention can first perform semantic analysis on each of the current voice instructions to be broadcasted when receiving multiple current voice instructions to be broadcasted, and then selects a corresponding target TTS model from multiple preset TTS models according to the analysis result, and broadcasts the corresponding current voice instructions to be broadcasted through each TTS model. Compared with the existing configuration of only one TTS model, the present invention can be provided with multiple preset TTS models, and the corresponding target TTS model can be selected from them for broadcasting, thereby supporting the simultaneous broadcasting of multiple current voice instructions to be broadcasted, improving the timeliness of the broadcasting and improving the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 A schematic diagram of the structure of an in-vehicle voice broadcast device in a hardware operating environment involved in an embodiment of the present invention;

[0042] Figure 2 It is a flow chart of the first embodiment of the in-vehicle voice broadcasting method of the present invention;

[0043] Figure 3 This is a schematic diagram of a table of identification positions in the first embodiment of the in-vehicle voice broadcasting method of the present invention;

[0044] Figure 4 It is a flow chart of the second embodiment of the in-vehicle voice broadcasting method of the present invention;

[0045] Figure 5 A distribution diagram of supported instruction ranges corresponding to the A model, the B model, and the C model in the second embodiment of the in-vehicle voice broadcasting method of the present invention;

[0046] Figure 6 This is a structural block diagram of the first embodiment of the in-vehicle voice broadcasting device of the present invention.

[0047] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0048] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0049] Reference Figure 1 , Figure 1 The present invention is a schematic diagram of the structure of an in-vehicle voice broadcast device in the hardware operating environment involved in an embodiment of the present invention.

[0050] like Figure 1 As shown, the in-vehicle voice broadcasting device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a wireless fidelity (Wireless-Fidelity, Wi-Fi) interface). The memory 1005 may be a high-speed random access memory (Random Access Memory, RAM), or a stable non-volatile memory (Non-Volatile Memory, NVM), such as a disk storage. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0051] Those skilled in the art will understand that Figure 1 The structure shown in the figure does not constitute a limitation on the in-vehicle voice broadcasting device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0052] like Figure 1 As shown, the memory 1005 as a storage medium may include an operating system, a network communication module, a user interface module, and an in-vehicle voice broadcast program.

[0053] exist Figure 1 In the in-vehicle voice broadcast device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the in-vehicle voice broadcast device of the present invention can be set in the in-vehicle voice broadcast device, and the in-vehicle voice broadcast device calls the in-vehicle voice broadcast program stored in the memory 1005 through the processor 1001, and executes the in-vehicle voice broadcast method provided by the embodiment of the present invention.

[0054] It should be noted that, at present, the car computer in the car is generally equipped with a text-to-speech (TTS) model, which converts the text information pre-output by the application software (Application, APP) in the car computer into voice for broadcast, such as navigation broadcast, car fault broadcast, etc.

[0055] However, since only one TTS model is configured in the existing car computer, when multiple APPs need to be broadcast, they are generally broadcast according to priority, but some broadcasts are still not timely, resulting in poor user experience.

[0056] Therefore, in order to solve the above defects, when receiving multiple voice commands to be broadcasted, this embodiment can first perform semantic analysis on each voice command to be broadcasted, and then select the corresponding target TTS model from multiple preset TTS models according to the analysis results, and broadcast the corresponding voice command to be broadcasted through each TTS model. Compared with the existing configuration of only one TTS model, this embodiment can be provided with multiple preset TTS models, and the corresponding target TTS model can be selected from them for broadcasting, and then multiple voice commands to be broadcasted can be supported to be broadcasted at the same time, which improves the timeliness of the broadcasting and improves the user experience.

[0057] For ease of understanding, refer to Figures 2 to 6 The embodiments of the present invention and the following embodiments are described in detail.

[0058] Reference Figure 2 , Figure 2 Schematic diagram of the flow chart of the first embodiment of the in-vehicle voice broadcasting method of the present invention.

[0059] In this embodiment, the in-vehicle voice broadcast method includes the following steps:

[0060] Step S10: when receiving a plurality of voice instructions to be currently broadcasted, semantic analysis is performed on each of the voice instructions to be currently broadcasted.

[0061] It is understandable that the method of this embodiment can be applied in the scenario of voice broadcasting by the vehicle computer. The execution subject of the method of this embodiment can be a computing service device with in-vehicle voice broadcasting, network communication and program running functions, such as a vehicle computer, etc., and can also be other electronic devices that realize the same or similar functions. The following is an explanation of this embodiment and the following embodiments using the above-mentioned in-vehicle voice broadcasting device (hereinafter referred to as the device).

[0062] It should be understood that the above-mentioned voice instructions currently to be broadcast may be voice instructions from different sources, for example, they may be voice instructions generated by different APPs, and may also include voice instructions generated by the vehicle computer itself. In this embodiment, several APPs may be installed in the above-mentioned device, such as navigation APP, music APP, and weather APP, etc. Corresponding voice instructions can be generated by different APPs to realize vehicle-computer interaction with users. Of course, the above-mentioned device can also be connected to the car to receive voice instructions generated by the car for abnormal broadcasts, such as "car tire pressure is too low", "door is not closed tightly" and other prompts.

[0063] In a specific implementation, after receiving multiple current voice instructions to be broadcast from different sources, the above-mentioned device can perform semantic analysis on each current voice instruction to be broadcast, and obtain the specific meaning corresponding to each current voice instruction to be broadcast as the semantic analysis result.

[0064] Step S20: selecting a corresponding target TTS model from a preset model pool according to each semantic analysis result, wherein the preset model pool includes a plurality of preset TTS models.

[0065] It should be noted that, in the present embodiment, a preset model pool may be provided in the above-mentioned device, and a number of preset TTS models may be provided in the above-mentioned preset model pool, and each preset TTS model can be used for voice broadcasting. For the convenience of subsequent description, the number of preset TTS models in the preset model pool of the present embodiment is described as three, and they are respectively recorded as model A, model B and model C.

[0066] It is understandable that in this embodiment, each preset TTS model can be used to broadcast different voice commands. Since this embodiment uses three preset TTS models for illustration, the three preset TTS models can be used to correspond to three voice commands with different functions. For example, model A can be used to broadcast abnormal voice commands (such as "the current vehicle speed is too high", "the tire pressure is too low", etc.), model B can be used to broadcast driving-related voice commands (such as "turn left two hundred meters ahead", etc.), and model C can be used to broadcast entertainment and voice commands related to vehicle parameters (such as "the current temperature inside the car is 18 degrees Celsius", etc.).

[0067] The above-mentioned preset TTS models and the corresponding voice commands can be set according to actual conditions, and this embodiment does not limit this.

[0068] In a specific implementation, the above-mentioned device can determine the function of each voice instruction to be broadcasted currently according to the semantic analysis result, and select the corresponding preset TTS model from the preset model pool as the target TTS model according to the function.

[0069] For example, if the voice commands to be broadcast include: "The current vehicle speed is too high" and "The current temperature inside the car is 18 degrees Celsius", after semantic analysis, the device can select model A and model C from model A, model B and model C as the above-mentioned target TTS models.

[0070] Step S30: broadcasting the corresponding current voice instruction to be broadcasted through each of the target TTS models.

[0071] After determining the specific target TTS model, the above device can broadcast the corresponding current voice command to be broadcasted through the target TTS model, that is, broadcast "the current vehicle speed is too high" through model A, and broadcast "the current temperature inside the vehicle is 18 degrees Celsius" through model C. Compared with the existing configuration of only one TTS model, this embodiment can be provided with multiple preset TTS models, and the corresponding target TTS model can be selected from them for broadcasting, and multiple current voice commands to be broadcasted can be supported to be broadcasted at the same time, which improves the timeliness of the broadcast and improves the user experience.

[0072] Furthermore, considering that if the corresponding currently to-be-announced voice instructions are simultaneously broadcasted through different target TTS models, the broadcasted sounds may be superimposed, and thus the user cannot accurately understand the broadcast content. Therefore, in this embodiment, the above defects can be solved by setting the broadcast speed, broadcast order and broadcast speaker corresponding to the target TTS model, specifically: the above step S30 includes:

[0073] Step S31: determining the broadcast object of each of the currently to-be-broadcasted voice instructions, and determining the broadcast parameters corresponding to each of the target TTS models based on each of the broadcast objects, wherein the broadcast parameters include a broadcast speed, a broadcast speaker, and a broadcast order.

[0074] It is understandable that the above-mentioned broadcast object can be the location of the main broadcast object of the current voice command to be broadcast, which can be determined according to the source user of the current voice command to be broadcast. For example, before driving, the driver sends the command "Navigate to City X" to the device. When it is necessary to broadcast "The current vehicle speed is too high" during driving, the above-mentioned device can determine that the current voice command to be broadcast "The current vehicle speed is too high" is mainly broadcast to the driver, and the corresponding broadcast object is the driver's seat; when the passenger on the co-pilot asks the device "What is the current temperature in the car", the above-mentioned device can determine that the current voice command to be broadcast "The current temperature in the car is 18 degrees Celsius" is mainly broadcast to the co-pilot passenger, and the corresponding broadcast object is the co-pilot seat.

[0075] After determining the broadcast object corresponding to each current voice instruction to be broadcast, the broadcast speed, broadcast speaker and broadcast order of each target TTS model for broadcasting the current voice instruction to be broadcast can be determined according to each broadcast object; the broadcast speed can be the speed at which the target TTS model broadcasts the current voice instruction to be broadcast, the broadcast speaker can be the speaker corresponding to the target TTS model broadcasting the current voice instruction to be broadcast, and the broadcast order can be the order in which the target TTS model broadcasts the current voice instruction to be broadcast.

[0076] Furthermore, in order to accurately determine the announcement parameters corresponding to each target TTS model, in this embodiment, the step of determining the announcement parameters corresponding to each target TTS model based on each announcement object includes:

[0077] Step S311: determining the importance of each of the currently to-be-announced voice instructions relative to the corresponding announcement object;

[0078] Step S312: determining the announcement parameters corresponding to each of the target TTS models based on the importance.

[0079] It should be understood that the above-mentioned importance level can be the importance level of each currently to-be-broadcasted voice command relative to the passengers at various positions in the vehicle. In this embodiment, the above-mentioned device can pre-detect the common passengers in the vehicle and the position of each passenger. For example, the positions in a 5-seater vehicle can be the driver's seat, the co-pilot's seat, the left rear seat, the middle rear seat and the right rear seat.

[0080] After determining the position of each passenger, the above-mentioned device can analyze the importance of the current voice instruction to be broadcast relative to the position. The present embodiment can divide the importance into three levels, namely "very important", "generally important" and "unimportant". Then, during the analysis, the importance of the broadcast object corresponding to the current voice instruction to be broadcast can be set to "very important", and other broadcast objects can be set to "generally important" or "unimportant".

[0081] It should be emphasized that when setting other broadcast objects to "generally important" or "unimportant", the above-mentioned device can consider whether other broadcast objects are currently due to the current voice command to be broadcast. If so, they can be set to "unimportant"; if not, they can be set to "generally important".

[0082] Exemplarily, after determining that the broadcast object corresponding to the current voice instruction to be broadcasted of "the current vehicle speed is too high" is the driving seat, and determining that the broadcast object corresponding to the current voice instruction to be broadcasted of "the current temperature in the car is 18 degrees Celsius" is the co-pilot seat, the importance of the current voice instruction to be broadcasted of "the current vehicle speed is too high" to the driving seat is "very important", and because there is a current voice instruction to be broadcasted of "the current temperature in the car is 18 degrees Celsius" to the co-pilot seat, the importance of the current voice instruction to be broadcasted of "the current vehicle speed is too high" to the co-pilot seat is "unimportant", and because there is no current voice instruction to be broadcasted for the rear passenger (for example, the left rear seat), the importance of the current voice instruction to be broadcasted of "the current vehicle speed is too high" to the left rear seat is "generally important"; similarly, the importance of the voice instruction to be broadcasted of "the current temperature in the car is 18 degrees Celsius" to the driving seat is "unimportant", the importance to the co-pilot seat is "very important", and the importance to the left rear seat is "generally important".

[0083] After determining the importance of each currently to-be-announced voice instruction relative to each announcement object, the announcement speed, announcement speaker, and announcement order corresponding to each target TTS model may be determined based on each importance.

[0084] Furthermore, in this embodiment, in order to accurately determine the broadcast speed, broadcast speaker and broadcast order of each target TTS model, the above-mentioned step S312 includes: determining the broadcast speaker corresponding to each of the currently to-be-broadcasted voice instructions based on the importance; determining the broadcast speed and broadcast order corresponding to each of the broadcast speakers through the importance; generating broadcast parameters according to each of the broadcast speakers, each of the broadcast speeds and each of the broadcast orders.

[0085] Continuing with the above example, after determining the importance of each broadcast object of model A broadcasting "the current vehicle speed is too high" and the importance of model C broadcasting "the current temperature inside the car is 18 degrees Celsius", the above device can first determine whether each broadcast object needs to be broadcast according to the importance, that is, whether the speaker at the location needs to work.

[0086] For example, since "the current vehicle speed is too high" is "very important" for the main driver's seat, the flag of the broadcast speaker at the main driver's seat can be set to "1"; for the co-driver's seat, it is "unimportant", and the flag of the broadcast speaker at the co-driver's seat is set to "0"; for the left side of the rear row, it is "generally important", and the flag of the broadcast speaker at the left side of the rear row is set to "1", and the broadcast speakers at other positions are set to "0" by default, where "1" indicates that the broadcast speaker is broadcasting and "0" indicates that the broadcast speaker is not broadcasting.

[0087] Similarly, for "the current temperature inside the car is 18 degrees Celsius", the flag bit of the broadcast speaker at the main driver's seat is set to "0", the flag bit of the broadcast speaker at the co-driver's seat is set to "1", and the flag bit of the broadcast speaker at the left rear seat is set to "1". As for the broadcast speakers at other positions, the default is "0". Then, a flag bit table of each voice command to be broadcast can be drawn according to the flag bit, refer to Figure 3 , Figure 3 Schematic diagram of the identification position table in the first embodiment of the in-vehicle voice broadcasting method of the present invention. Figure 3 As shown, after the above settings, the flag position of the broadcast speaker corresponding to each broadcast object of "the current vehicle speed is too high" and "the current temperature inside the car is 18 degrees Celsius" can be obtained.

[0088] After determining the announcement speaker, the above device can continue to determine the announcement speed and the announcement order according to the importance. When determining the announcement speed, the above device can determine it according to the flag bit of the announcement speaker, that is, if there is only one "1" in the announcement speaker, the announcement speed can be set to "default" and the announcement order can be "none", that is, Figure 3 In the "Current vehicle speed is too high" section, the corresponding announcement speed of the main driver's seat is set to "Default", and the announcement order is set to "None". The corresponding announcement speed of the co-pilot seat "Current vehicle temperature is 18 degrees Celsius" is set to "Default", and the announcement order is set to "None";

[0089] If the flag bit of the announcement speaker is "0", the announcement speed and announcement sequence can be set to "none", that is, the announcement speaker does not broadcast. Figure 3 In the middle, set the "Current vehicle speed is too high" for the front passenger seat, the middle rear seat, and the right rear seat to "None", and set the "Current vehicle temperature is 18 degrees Celsius" for the main driver's seat, the middle rear seat, and the right rear seat to "None";

[0090] If there is more than one "1" on the announcement speaker, the announcement speed can be set according to the importance of the current voice command to be announced relative to the main driver's seat. If the importance is "very important", it is set to "slower" and the announcement order is set to "high priority", that is, Figure 3 In the "Current vehicle speed is too high" report speed of the left rear seat is set to "slower", and the report order is set to "high priority". If the importance is "unimportant", it is set to "faster", and the report order is set to "low priority", that is, Figure 3 The reporting speed for the left rear seat of "The current temperature in the car is 18 degrees Celsius" is set to "Fast" and the reporting order is set to "Low priority".

[0091] Step S32: each target TTS model announces the corresponding current voice instruction to be announced according to the corresponding announcement parameters.

[0092] After determining the broadcasting speakers, broadcasting speeds and broadcasting sequences, the voice broadcasting parameters can be used to enable the target TTS model to broadcast according to the corresponding broadcasting parameters, that is, for "the current vehicle speed is too high", the speakers corresponding to the main driver's seat and the left rear seat will broadcast, and the speakers for the co-driver's seat, the middle rear seat and the right rear seat will not broadcast, and the left rear seat will give priority to broadcasting at a slower speed when broadcasting; for "the current temperature in the car is 18 degrees Celsius", the speakers corresponding to the co-driver's seat and the left rear seat will broadcast, and the speakers for the main driver's seat, the middle rear seat and the right rear seat will not broadcast, and the left rear seat will broadcast at a faster speed after broadcasting "the current vehicle speed is too high".

[0093] After receiving multiple current voice instructions to be broadcast from different sources, the above-mentioned device in this embodiment can respectively perform semantic analysis on each current voice instruction to be broadcast, and obtain the specific meaning corresponding to each current voice instruction to be broadcast as the result of semantic analysis; the above-mentioned device can determine the function of each current voice instruction to be broadcast according to the result of semantic analysis, and select the corresponding preset TTS model from the preset model pool as the target TTS model according to the function; after determining the specific target TTS model, the above-mentioned device can broadcast the corresponding current voice instruction to be broadcast through the target TTS model. Furthermore, compared with the existing configuration of only one TTS model, this embodiment can be provided with multiple preset TTS models, and the corresponding target TTS model can be selected from them for broadcast, and then multiple current voice instructions to be broadcast can be supported to be broadcasted at the same time, which improves the timeliness of the broadcast and improves the user experience.

[0094] refer to Figure 4 , Figure 4 Schematic diagram of the flow chart of the second embodiment of the in-vehicle voice broadcasting method of the present invention.

[0095] Considering that in this example, when selecting the target TTS model according to the semantic analysis results of the current voice command to be broadcasted, some of the current voice commands to be broadcasted may support broadcasting by multiple preset TTS models. For example, model A used to broadcast abnormal voice commands can support broadcasting "the current vehicle speed is too high", and model B used to broadcast driving-related voice commands can also support broadcasting "the current vehicle speed is too high". Therefore, in order to accurately determine the target TTS model, such as Figure 4 As shown, in this embodiment, the step of selecting the corresponding target TTS model from the preset model pool according to each semantic analysis result includes:

[0096] Step S21: selecting candidate TTS models from a preset model pool according to the semantic analysis results, and obtaining a distribution map of supported instruction ranges corresponding to each candidate TTS model.

[0097] It should be noted that the above-mentioned alternative TTS model can be a model that can support the broadcast of the voice command currently to be broadcast. Continuing based on the above example, for "the current vehicle speed is too high", model A and model B can be selected as alternative TTS models, and for "the current temperature in the car is 18 degrees Celsius", model B and model C can be selected as alternative TTS models.

[0098] After the candidate TTS models are determined, a distribution map of supported instruction ranges corresponding to each candidate TTS model may be obtained. The distribution map of supported instruction ranges may be a schematic diagram corresponding to the instruction ranges supported by each preset TTS model.

[0099] Step S22: Draw a global instruction range distribution map based on each of the supported instruction range distribution maps.

[0100] It is understandable that there may be an intersection between the supported instruction range distribution maps corresponding to the preset TTS models, that is, "current vehicle speed is too high" may be in the supported instruction range distribution map corresponding to model A, and may also be in the supported instruction range distribution map corresponding to model B. For ease of understanding, refer to Figure 5 To explain, Figure 5 This is a distribution diagram of the supported instruction ranges corresponding to the A model, the B model, and the C model in the second embodiment of the in-vehicle voice broadcast method of the present invention, as shown in FIG. Figure 5 As shown in FIG. 1 , after determining the supported instruction range distribution diagrams corresponding to model A, model B, and model C, the global instruction range distribution diagrams of the three models can be drawn, that is, Figure 5 The distribution diagram composed of the three models, among which the distribution diagrams of the supported instruction ranges of the three models have intersections between each other, and there are also intersections among the three models.

[0101] Step S23: Determine the position information of each of the semantic analysis results in the global instruction range distribution map.

[0102] Continue to combine Figure 5 As shown, after drawing the global command range distribution map, the above-mentioned device can determine the position (that is, the above-mentioned position information) in the global command range distribution map according to the semantic analysis results corresponding to each current voice command to be broadcast, that is, Figure 5 The M in the middle indicates the position of "the current vehicle speed is too high", and the N indicates the position of "the current temperature inside the vehicle is 18 degrees Celsius".

[0103] Step S24: selecting a target TTS model from the candidate TTS models based on the location information.

[0104] After determining the location information, the corresponding model can be selected as the target TTS model. However, it should be emphasized that there may be multiple combinations of the selected target TTS model, namely Figure 5 There are four combinations of target TTS models, namely 1. Model A and Model B; 2. Model B; 3. Model A and Model C; 4. Model B and Model C. In order to accurately select a suitable target TTS model, in this embodiment, the above step S24 includes:

[0105] Step S241: Generate a model combination list according to each of the candidate TTS models based on the location information.

[0106] The above model combination list can be a list consisting of the above four combinations: 1. Model A and Model B; 2. Model B; 3. Model A and Model C; 4. Model B and Model C.

[0107] Step S242: Determine the quantity information of the current voice instructions to be broadcast, and determine the target model combination from the model combination list based on the quantity information.

[0108] It is understandable that after generating the model combination list, the above-mentioned device can determine how many voice commands are currently received and waiting to be broadcast. If "the current vehicle speed is too high" and "the current temperature in the car is 18 degrees Celsius" are currently received, the corresponding quantity information is 2.

[0109] When determining the target model combination according to the quantity information, the present embodiment can filter the model combination list according to the quantity information, that is, generally, the number of current voice commands to be broadcast can be set, that is, the number of preset TTS models to be selected at least. For example, if the quantity information is 2, the combination method of selecting only the B model can be deleted, and the remaining three combination methods can be used as the target model combination.

[0110] Furthermore, in order to facilitate user distinction, the present embodiment can also set a corresponding timbre for each preset TTS model, that is, different preset TTS models can use different timbre for broadcasting, where the timbre may include but is not limited to: Mandarin, Cantonese, local dialect, male voice and female voice, etc. However, considering that the present embodiment involves a scenario where multiple preset TTS models are broadcasting at the same time, in order to enable users to understand the broadcast content more clearly, the present embodiment can analyze the similarity of each preset TTS model when broadcasting according to the corresponding timbre, and select the preset TTS model with smaller similarity as the target TTS model for broadcasting. Specifically: the above step of determining the target model combination from the model combination list based on the quantity information includes:

[0111] Step S2421: Filter the model combination list based on the quantity information to obtain a preliminary screening combination list.

[0112] It should be noted that after screening the above quantitative information, the initial screening combination list obtained is: 1. Model A and Model B; 2. Model A and Model C; 3. A list consisting of three combinations of Model B and Model C.

[0113] Step S2422: Acquire the timbre information corresponding to each candidate TTS model in the initial screening combination list, and determine the current voice instruction to be broadcast corresponding to each candidate TTS model.

[0114] It can be understood that if the timbre information corresponding to the A model is recorded as timbre a, the timbre information corresponding to the B model is recorded as timbre b, and the timbre information corresponding to the C model is recorded as timbre c, then according to the above-mentioned preliminary screening combination list, the above-mentioned device can determine the corresponding broadcasting scheme as follows: 1. The A model broadcasts M according to the timbre a, and the B model broadcasts N according to the timbre b; 2. The A model broadcasts M according to the timbre a, and the C model broadcasts N according to the timbre c; 3. The B model broadcasts M according to the timbre b, and the C model broadcasts N according to the timbre c.

[0115] Step S2423: performing timbre similarity analysis on each of the currently to-be-broadcasted voice instructions based on each of the timbre information;

[0116] Step S2424: Determine the target model combination from the preliminary screening combination list according to the analysis results.

[0117] After determining the broadcasting plan, the above-mentioned device can perform similarity analysis on the timbre of each broadcast according to the broadcasting plan, and can extract the timbre features of the corresponding content of each model broadcast in each broadcasting plan, and measure the similarity between the extracted timbre features according to the Euclidean distance as the above-mentioned analysis result.

[0118] After obtaining the analysis results, the above-mentioned device can select the broadcasting scheme with the smallest similarity and use the model in the broadcasting scheme as the above-mentioned target model combination. For example, if the similarity obtained by model A broadcasting M according to timbre a and model C broadcasting N according to timbre c is the smallest, then the combination of model A and model C can be used as the above-mentioned target model combination.

[0119] Step S243: selecting a target TTS model from the candidate TTS models according to the target model combination.

[0120] In a specific implementation, the above-mentioned device can use the A model and the C model as the above-mentioned target TTS models.

[0121] In this embodiment, the above-mentioned device can select candidate TTS models according to the semantic analysis results, and select the target TTS model from the candidate TTS models according to the distribution map of the supported instruction ranges corresponding to each candidate TTS model, thereby improving the accuracy of the selection.

[0122] In addition, an embodiment of the present invention further proposes a storage medium, on which an in-vehicle voice broadcast program is stored. When the in-vehicle voice broadcast program is executed by a processor, the steps of the in-vehicle voice broadcast method described above are implemented.

[0123] Reference Figure 6 , Figure 6 This is a structural block diagram of the first embodiment of the in-vehicle voice broadcasting device of the present invention.

[0124] like Figure 6As shown, the in-vehicle voice broadcasting device proposed in the embodiment of the present invention includes:

[0125] The semantic analysis module 601 is used to perform semantic analysis on each of the currently to-be-broadcasted voice instructions when receiving multiple currently to-be-broadcasted voice instructions;

[0126] A model selection module 602 is used to select a corresponding target TTS model from a preset model pool according to each semantic analysis result, wherein the preset model pool includes a plurality of preset TTS models;

[0127] The voice broadcasting module 603 is used to broadcast the corresponding current voice instruction to be broadcast through each of the target TTS models.

[0128] After receiving multiple current voice instructions to be broadcast from different sources, the above-mentioned device in this embodiment can respectively perform semantic analysis on each current voice instruction to be broadcast, and obtain the specific meaning corresponding to each current voice instruction to be broadcast as the result of semantic analysis; the above-mentioned device can determine the function of each current voice instruction to be broadcast according to the result of semantic analysis, and select the corresponding preset TTS model from the preset model pool as the target TTS model according to the function; after determining the specific target TTS model, the above-mentioned device can broadcast the corresponding current voice instruction to be broadcast through the target TTS model. Furthermore, compared with the existing configuration of only one TTS model, this embodiment can be provided with multiple preset TTS models, and the corresponding target TTS model can be selected from them for broadcast, and then multiple current voice instructions to be broadcast can be supported to be broadcasted at the same time, which improves the timeliness of the broadcast and improves the user experience.

[0129] As an implementation mode, the voice broadcast module 603 is also used to determine the broadcast object of each of the currently to-be-broadcasted voice instructions, and determine the broadcast parameters corresponding to each of the target TTS models based on each of the broadcast objects, the broadcast parameters including the broadcast speed, the broadcast speaker and the broadcast order; the corresponding currently to-be-broadcasted voice instructions are broadcasted through each of the target TTS models according to the corresponding broadcast parameters.

[0130] As an implementation mode, the voice broadcast module 603 is further used to determine the importance of each of the currently to-be-broadcasted voice instructions relative to the corresponding broadcast object; and determine the broadcast parameters corresponding to each of the target TTS models based on the importance.

[0131] As an implementation mode, the voice broadcast module 603 is also used to determine the broadcast speakers corresponding to each of the currently to-be-broadcasted voice instructions based on the importance level; determine the broadcast speed and broadcast order corresponding to each of the broadcast speakers through the importance level; and generate broadcast parameters according to each of the broadcast speakers, each of the broadcast speeds and each of the broadcast orders.

[0132] Based on the first embodiment of the in-vehicle voice broadcasting device of the present invention, a second embodiment of the in-vehicle voice broadcasting device of the present invention is proposed.

[0133] In this embodiment, the model selection module 602 is also used to select alternative TTS models from a preset model pool according to each semantic analysis result, and obtain a support instruction range distribution map corresponding to each of the alternative TTS models; draw a global instruction range distribution map based on each of the support instruction range distribution maps; determine the position information of each of the semantic analysis results in the global instruction range distribution map; and select a target TTS model from each of the alternative TTS models based on the position information.

[0134] As an implementation mode, the model selection module 602 is also used to generate a model combination list according to each of the alternative TTS models based on the position information; determine the quantity information of the current voice instructions to be broadcast, and determine the target model combination from the model combination list based on the quantity information; and select the target TTS model from each of the alternative TTS models according to the target model combination.

[0135] As an implementation method, the model selection module 602 is also used to screen the model combination list based on the quantity information to obtain a preliminary screening combination list; obtain the timbre information corresponding to each of the alternative TTS models in the preliminary screening combination list, and determine the current voice instructions to be broadcast corresponding to each of the alternative TTS models; perform timbre similarity analysis on each of the current voice instructions to be broadcast based on the timbre information; and determine the target model combination from the preliminary screening combination list based on the analysis results.

[0136] Other embodiments or specific implementations of the in-vehicle voice broadcast device of the present invention can refer to the above-mentioned method embodiments and will not be described in detail here.

[0137] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0138] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0139] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as a read-only memory / random access memory, a magnetic disk, or an optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0140] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for in-vehicle voice broadcasting, characterized in that: The method comprises the following steps: When receiving a plurality of voice instructions to be currently broadcasted, performing semantic analysis on each of the voice instructions to be currently broadcasted; Selecting a corresponding target TTS model from a preset model pool according to each semantic analysis result, wherein the model pool includes a plurality of preset TTS models; The corresponding current voice command to be broadcasted is broadcasted through each of the target TTS models.

2. The in-vehicle voice broadcast method according to claim 1, characterized in that: The step of selecting a corresponding target TTS model from a preset model pool according to each semantic analysis result includes: Selecting candidate TTS models from a preset model pool according to each semantic analysis result, and obtaining a distribution map of supported instruction ranges corresponding to each candidate TTS model; Draw a global instruction range distribution map based on each of the supported instruction range distribution maps; Determine the position information of each of the semantic analysis results in the global instruction range distribution map; A target TTS model is selected from the candidate TTS models based on the position information.

3. The in-vehicle voice broadcasting method as claimed in claim 2, characterized in that: The step of selecting a target TTS model from each of the candidate TTS models based on the location information includes: Generate a model combination list according to each of the candidate TTS models based on the location information; Determine the quantity information of the voice instructions to be broadcasted currently, and determine the target model combination from the model combination list based on the quantity information; A target TTS model is selected from the candidate TTS models according to the target model combination.

4. The in-vehicle voice broadcasting method as claimed in claim 3, characterized in that: The step of determining the target model combination from the model combination list based on the quantity information comprises: Screening the model combination list based on the quantity information to obtain a preliminary screening combination list; Acquire the timbre information corresponding to each candidate TTS model in the preliminary screening combination list, and determine the current voice command to be broadcast corresponding to each candidate TTS model; Performing timbre similarity analysis on each of the currently to-be-broadcasted voice commands based on each of the timbre information; The target model combination is determined from the preliminary screening combination list according to the analysis results.

5. The in-vehicle voice broadcast method according to any one of claims 1 to 4, characterized in that: The step of broadcasting the corresponding current voice instruction to be broadcasted through each target TTS model includes: Determine the broadcast object of each of the currently to-be-broadcasted voice commands, and determine the broadcast parameters corresponding to each of the target TTS models based on each of the broadcast objects, wherein the broadcast parameters include a broadcast speed, a broadcast speaker, and a broadcast order; The corresponding current voice instruction to be broadcasted is broadcasted through each of the target TTS models according to the corresponding broadcasting parameters.

6. The in-vehicle voice broadcasting method as claimed in claim 5, characterized in that: The step of determining the announcement parameters corresponding to each of the target TTS models based on each of the announcement objects includes: Determining the importance of each of the currently to-be-announced voice instructions relative to the corresponding announcement object; The broadcast parameters corresponding to each of the target TTS models are determined based on the importance.

7. The in-vehicle voice broadcasting method according to claim 6, characterized in that: The step of determining the broadcast parameters corresponding to each of the target TTS models based on the importance level includes: Determining the broadcasting loudspeaker corresponding to each of the currently to-be-broadcasted voice instructions based on the importance; Determine the broadcast speed and broadcast order corresponding to each of the broadcast speakers according to the importance; The announcement parameters are generated according to each of the announcement speakers, each of the announcement speeds, and each of the announcement sequences.

8. A voice broadcasting device in a car, characterized in that: The device comprises: A semantic analysis module, for performing semantic analysis on each of the currently to-be-broadcasted voice instructions when receiving a plurality of currently to-be-broadcasted voice instructions; A model selection module, used to select a corresponding target TTS model from a preset model pool according to each semantic analysis result, wherein the model pool is provided with a plurality of preset TTS models; The voice broadcast module is used to broadcast the corresponding current voice instruction to be broadcast through each of the target TTS models.

9. An in-car voice broadcasting device, characterized in that: The device comprises: a memory, a processor, and an in-vehicle voice broadcast program stored in the memory and executable on the processor, wherein the in-vehicle voice broadcast program is configured to implement the steps of the in-vehicle voice broadcast method as described in any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium stores an in-vehicle voice broadcast program, and when the in-vehicle voice broadcast program is executed by the processor, the steps of the in-vehicle voice broadcast method according to any one of claims 1 to 7 are implemented.

Images