Method for waking up nearby voice devices in whole-house intelligent system
By adopting the voice intensity arbitration mechanism between voice panels in the whole-house intelligent system, the problem of multiple panels responding at the same time is solved, ensuring that only one panel responds, improving user experience and simplifying the response process and reducing network dependence.
Patent Information
- Application Number
- CN202110653025.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-06-11
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2041-06-11
AI Technical Summary
In a whole-house intelligent system, when multiple voice panels receive the user's wake-up voice at the same time, the lack of arbitration mechanism causes multiple voice panels to respond simultaneously, affecting the user experience, and the existing solutions rely on network and third-party processing, resulting in slow response time and network bandwidth usage.
Through voice strength arbitration between voice panels in the LAN, the voice panel that receives wake-up voice is obtained and broadcasts the voice intensity of itself and other panels. After comparison, decide whether to respond. If the intensity of itself is greater than other panels, respond, otherwise it will not behave, avoiding Multiple panels respond simultaneously and distinguish different users' wake-up voices through tone/tooth tags.
Ensure that there is only one voice panel closest to the user to respond, improve user experience, reduce network dependence, simplify the response process, improve response speed, and avoid the problem of multiple panels responding at the same time.
Smart Images

Figure CN113506570B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of whole-house intelligent systems, and in particular relates to a method for waking up a voice device nearby in a whole-house intelligent system. Background Art
[0002] With the deepening of the concept of smart home, whole-house smart systems are becoming more and more common. In a whole-house smart system, in order to meet the voice control of users in different locations, multiple voice panels are usually arranged in the venue to capture the voice commands of users in various corners. Although there is a certain distance between each voice panel, it is inevitable that the voice issued by the user can be received by two or even more voice panels at the same time. At present, there is no centralized arbitration mechanism between the various voice panels. When the user issues a wake-up word, if multiple voice panels receive the voice signal, the multiple voice panels will respond at the same time, that is, the user will hear the response words of multiple voice panels. For example, the user issues the wake-up word "Xiao K Xiao K", and three panels receive the wake-up word at the same time. The three panels simultaneously issue the response word "Hello, I'm here", which greatly reduces the intelligence of the whole-house smart system and provides a poor user experience.
[0003] A Chinese national patent discloses a method and apparatus for waking up a voice device [application number: 2020115152996]. This solution determines the acoustic features of the voice signal collected by the voice device, determines the wake-up score of the voice device based on the acoustic features, and sends the wake-up score to the server. The server determines a voice device as the target voice device for wake-up based on the wake-up scores of multiple voice devices received, which can solve the problem of waking up multiple voice devices at the same time.
[0004] Although the above solution can solve the problem of voice devices being woken up at the same time, all voice devices in the above solution need to calculate the wake-up score and send the wake-up score. A third party determines which voice device will finally respond based on the wake-up score. This solution has complex steps, slow response time, and requires reliance on the network, occupying network bandwidth, and other problems. Summary of the invention
[0005] The purpose of the present invention is to provide a method for waking up a voice device nearby in a whole-house intelligent system in response to the above-mentioned problem.
[0006] In order to achieve the above object, the present invention adopts the following technical solutions:
[0007] A method for waking up a voice device in a whole-house smart system nearby includes multiple voice panels, and the multiple voice panels respond to the user's wake-up voice nearby through the following arbitration method:
[0008] S1. At least one voice panel in the local area network receives the wake-up voice;
[0009] S2. The voice panels receiving the wake-up voice obtain the voice strength of the wake-up voice received and broadcast the voice strength;
[0010] S3. The voice panel that receives the wake-up voice compares its own voice strength with the received broadcast voice strengths one by one. When the strength is greater than all the broadcast voice strengths compared with it, it responds, otherwise it does nothing.
[0011] In the above-mentioned method for waking up voice devices nearby in the whole-house smart system, in step S3, the voice panel compares its own voice intensity with the broadcast voice intensity received within the first preset time period one by one.
[0012] In the above-mentioned method for waking up a voice device nearby in a whole-house smart system, in step S3, if the broadcast voice intensity is not received within the first preset time period, the voice intensity comparison is not performed, and the voice panel directly responds;
[0013] And the first preset time period is less than or equal to 0.1s.
[0014] In the above-mentioned method for waking up voice devices nearby in the whole-house smart system, a communication table is stored in each voice panel. In step S3, when there is a broadcast voice intensity equal to the voice panel's own voice intensity and there is no broadcast voice intensity greater than its own voice intensity, it is detected whether a response result unicast signal from other voice panels is received. If not, a response is made and the response result is unicasted to the voice panel that broadcasts a broadcast voice intensity equal to its own voice intensity.
[0015] In the above-mentioned method for waking up a nearby voice device in the whole-house smart system, in step S2, the timbre and / or pitch of the wake-up voice are simultaneously acquired, and the pitch and / or timbre are used as label signals of the voice intensity, and in step S3, the voice intensities with the same label signals are compared with each other, and the voice intensities with different label signals are not compared with each other.
[0016] In the above-mentioned method for waking up a voice device nearby in the whole-house smart system, step S3 is specifically as follows:
[0017] The voice panel that receives the wake-up voice compares its own voice strength with the broadcast voice strengths with the same tag signal one by one. When the strength is greater than all the compared broadcast voice strengths, it responds, otherwise it does nothing.
[0018] In the above-mentioned method for waking up voice devices nearby in the whole-house smart system, in step S3, when a voice panel has multiple self-voice intensities carrying different label signals at the same time, each self-voice intensity is compared with the broadcast voice intensity with the same label. When there is at least one self-voice intensity greater than all the broadcast voice intensities compared with it, a response is made, otherwise no action is taken.
[0019] In the above-mentioned method for waking up voice devices nearby in the whole-house smart system, in step S3, when a voice panel simultaneously has multiple voice intensities of its own carrying different label signals, and there are multiple voice intensities of its own that are greater than all broadcast voice intensities compared with it, a response is made.
[0020] In the above-mentioned method for waking up voice devices nearby in the whole-house smart system, in step S2, when a voice panel simultaneously has multiple voice intensities carrying different label signals, the multiple voice intensities are broadcast simultaneously.
[0021] In the above-mentioned method for waking up a voice device nearby in the whole-house intelligent system, step S3 is specifically as follows:
[0022] The voice panel that receives the wake-up voice first determines whether its own voice intensity is greater than the intensity threshold. If so, it responds directly. Otherwise, it compares its own voice intensity with the received broadcast voice intensity one by one. When it is greater than all the compared broadcast voice intensities, it responds, otherwise it does not take any action.
[0023] The advantages of the present invention are that when multiple voice panels simultaneously receive a user's wake-up voice, it can ensure that only one voice panel closest to the user responds by self-determination, thereby avoiding the problem of multiple voice panels responding at the same time and causing poor user experience. At the same time, because the wake-up is nearby, it can ensure that the user can clearly hear the response word, thereby facilitating the interaction between the user and the intelligent system.
[0024] Multiple voice panels decide whether to respond or not through self-arbitration, without the need for third-party comparison, ensuring that only one voice panel closest to the user responds while ensuring the response speed;
[0025] The voice panel broadcasts the voice strength through broadcasting so that other voice panels can compare the strength and make the final decision of responding or not responding, which is independent of the network.
[0026] The voice strength is labeled according to the pitch / timbre. When multiple users send out wake-up voices at the same time, the voice strengths of the wake-up voices of multiple users can be compared separately to avoid the problem that only one voice panel responds when multiple voice panels are required to respond. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] Figure 1 This is a flow chart of the method for nearby wake-up in Embodiment 1 of the present invention;
[0028] Figure 2 This is a schematic diagram of a scenario of four voice panels in Embodiment 1 of the present invention;
[0029] Figure 3 This is a schematic diagram of a scene with five voice panels in Embodiment 1 of the present invention;
[0030] Figure 4 This is a schematic diagram of a scenario in which two users simultaneously control four voice panels in Embodiment 3 of the present invention;
[0031] Figure 5 This is a flow chart of the method for nearby wake-up in Embodiment 4 of the present invention. DETAILED DESCRIPTION
[0032] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.
[0033] The whole house intelligent system of this scheme refers to an intelligent system applied to the fields of home, hotel, office, elderly care, education, etc., such as a smart home system, which can be based on a Zigbee network or a network with a local area network such as Ethernet. This embodiment takes the Zigbee network as an example, and arranges a Zigbee gateway and a voice panel and controlled devices that access the Zigbee gateway through the Zigbee network in a whole house intelligent system. The voice panel is used to realize voice interaction with the user, so that the intelligent system can control the corresponding controlled device actions according to the user's voice command. In actual scenarios, users can interact with the voice panel in any local area network to realize voice control of any controlled device in the local area network. Since it is necessary to ensure that the normal decibel voice of users in every corner can be received in a whole house intelligent system, multiple voice panels will be installed in the venue. Taking a three-room family as an example, one or more voice panels are usually installed in the living room, dining room, two bathrooms, a kitchen, three bedrooms, etc. In actual applications, users may issue voice control commands at any location, so two or more voice panels may receive voice at the same time. Traditionally, there is no arbitration mechanism between the voice panels, and all voice panels that receive the wake-up voice will respond. Although control can be achieved, it will give users a very poor experience and affect the popularity of the whole-house smart system. If the user does not want to hear two voice panels responding at the same time, he needs to get as close to one of the voice panels as possible and use a relatively light voice to control the voice to avoid waking up the voice panel next door.
[0034] Embodiment 1
[0035] This solution proposes a nearby wake-up method for voice strength arbitration between voice panels. After the user issues a wake-up voice, only the nearest voice panel will respond, which can effectively avoid the embarrassing situation of multiple voice panels responding at the same time. At the same time, since the voice panel closest to the user responds, it can provide the user with the best user experience. Figure 1 As shown, the specific implementation method includes the following steps:
[0036] S1. At least one voice panel in the local area network receives the wake-up voice;
[0037] S2. The voice panels receiving the wake-up voice obtain the voice strength of the wake-up voice received by each voice panel and broadcast the voice strength, and all voice panels in the local area network receive all broadcast voice strengths;
[0038] S3. The voice panel that receives the wake-up voice compares its own voice strength with the received broadcast voice strength one by one. When the strength is greater than all the broadcast voice strengths compared with it, it responds that the voice panel is awakened. Otherwise, it does not take any action and the voice panel is not awakened.
[0039] Furthermore, when the user stands between two voice panels, it is possible that the voice intensities of the wake-up voices received by the two voice panels are equally strong. In this case, this solution solves the problem of the voice intensities of the two voice panels being equally strong in the following manner:
[0040] like Figure 2As shown, here, an example of a nearby wake-up scenario is taken as an example with four voice panels A, B, C, and D in the network: when the user sends a wake-up voice "Xiao K Xiao K", voice panels A and B receive the wake-up voice, and A obtains the voice strength of the received wake-up voice, which is called its own voice strength A, and broadcasts the voice strength, which is called broadcast voice strength A. It should be noted that the own voice strength A and the broadcast voice strength A are the voice strengths of the same voice by voice panel A, so they are the same in size, and voice panels B, C, and D will all receive the broadcast voice strength A; similarly, B also obtains the voice strength of the wake-up voice it receives, which is called its own voice strength B, and broadcasts the voice strength, which is called broadcast voice strength B. Similarly, the own voice strength B and the broadcast voice strength B are the same in size, and voice panels A, C, and D will all receive the broadcast voice strength B. Voice panels C and D do not have their own voice strength, so they do not act directly. Voice panel A compares its own voice intensity A with the broadcast voice intensity B, and the result is that its own voice intensity A is greater than the broadcast voice intensity B, so voice panel A responds and sends a voice response of "I am here". Voice panel B compares its own voice intensity B with the broadcast voice intensity A, and the result is that its own voice intensity B is less than the broadcast voice intensity A, so it does not act. It can be seen that through this method, in general scenarios, it can ensure that only one voice panel closest to the user responds to the user's wake-up voice, effectively improving the user experience. In addition, the method determines whether to respond by the voice panel's autonomous comparison, without the need for third-party processing. The execution process is simple and fast, which can effectively improve the response speed and further enhance the user experience. It does not rely on the network and does not occupy any network space.
[0041] Furthermore, in step S3, the voice panel compares its own voice intensity with the broadcast voice intensity received within the first preset time period one by one, and if no broadcast voice intensity is received within the first preset time period, no voice intensity comparison is performed, and the voice panel responds directly, which usually means that only one voice panel receives the wake-up voice.
[0042] Preferably, the first preset time period is less than or equal to 0.1s, such as 0.1s, 0.09s, 0.08s, or 0.07s. Since the closer the distance between two voice panels is, the more timely the broadcast signals between them are received, by setting the first preset time period, it is only necessary to compare the voice intensity broadcasted by the voice panel that is closer to it, which can effectively improve the processing speed.
[0043] In addition, it is preferred that each voice panel deletes its own voice intensity and broadcast voice intensity regarding the wake-up voice after receiving the second preset time period of the wake-up voice, and the second preset time period is greater than or equal to the first preset time period, and the second preset time period can be 3s, 5s, etc.
[0044] Combine the following Figure 3 The advantages of setting the first preset time period are described. Figure 3 As shown, assuming that the user is near B, the wake-up voice issued is received by five voice panels A, B, C, D, and E. The five voice panels are spaced a certain distance apart, and panel B is closest to the user. We need voice panel B to respond, and other voice panels do not act. Since voice panels A and C are closest to voice panel B, voice panel B will first receive broadcast voice intensity A and broadcast voice intensity C and compare them with broadcast voice intensity A and C. If the first preset time is not set, voice panel B needs to continue to compare with broadcast voice intensity D and broadcast voice intensity E before responding. This solution sets a first preset time period so that each voice panel compares the broadcast voice intensity received within the first preset time period after receiving the wake-up voice with its own voice intensity. The first preset time period can be set to be only enough to receive the broadcast voice intensity of the adjacent voice panel, that is, only enough for the B voice panel to receive the broadcast voice intensity A and the broadcast voice intensity C. In the B voice panel, it is only necessary to compare with the broadcast voice intensity A and the broadcast voice intensity C respectively, and there is no need to compare with the broadcast voice intensity D and the broadcast voice intensity E. In the scene where the voice panels are more densely distributed, the response speed of the B voice panel can be significantly improved, and the response speed is improved significantly. In the A voice panel, it can be concluded that the broadcast voice intensity B is greater than the own voice intensity A, so no action is taken; in the C voice panel, it can be concluded that the broadcast voice intensity B is greater than the own voice intensity C, so no action is taken; in the D voice panel, since the user is closer to C, it can be concluded that the broadcast voice intensity C is greater than the own voice intensity D, so no action is taken, and the E voice panel is similar to the D voice panel. So the final result is still that only the B voice panel responds, and the response speed can be significantly improved. When put into use, the first preset time period can be set according to the time required for each voice panel in the network to broadcast to the adjacent voice panel. The first time period needs to be at least greater than the time required for the two voice panels with the farthest adjacent relationship in the network to broadcast to the adjacent voice panels.
[0045] Embodiment 2
[0046] This embodiment is similar to the first embodiment, except that this embodiment takes into account the possibility that two voice panels have the same voice strength. Figure 2The voice strength of the wake-up voice received by voice panel A and voice panel B may be the same. In this case, the two voice panels may not respond or both respond. This solution avoids this situation by the following methods:
[0047] Each voice panel stores a communication table obtained from the gateway. In step S3, when there is a broadcast voice intensity equal to the voice panel's own voice intensity and there is no broadcast voice intensity greater than its own voice intensity, it detects whether a response result unicast signal from other voice panels is received. If not, it responds and unicasts the response result to the voice panel that broadcasts a broadcast voice intensity equal to its own voice intensity. Figure 2 In the example, when the voice intensity A is equal to the voice intensity B, if A is compared first, A responds and unicasts the response result to the B voice panel. The B voice panel receives the response result unicast signal but does not respond.
[0048] This method can ensure that under normal circumstances, two voice panels will not respond at the same time, effectively improving the user experience and contributing to the popularization and application of whole-house smart systems.
[0049] Embodiment 3
[0050] This embodiment is similar to the first embodiment, except that this embodiment takes into account the possibility of two users issuing voice commands at the same time. The probability of two users issuing commands at the same time is relatively low, but it does not absolutely not happen. Especially in a large-scale application scenario, the setting of the first preset time period in the first embodiment can avoid the problem of two users at a relatively far distance issuing wake-up voices at the same time to a certain extent, causing mutual influence (in terms of Figure 3 For example, assuming that user 1 is near voice panel A and user 2 is near voice panel E, the first preset time period can only allow the voice panel to receive the broadcast voice intensity of the adjacent voice panel, then the voice intensities of voice panel A and voice panel E will not be compared with each other, and each has its own voice intensity greater than the result of all broadcast voice intensities compared with it, so the two voice panels can respond separately), but it is impossible to avoid the interference problem caused by two users sending wake-up voices at the same time in all cases. In addition, everyone has different timbre and tone when speaking, so this embodiment is further optimized as follows:
[0051] In step S2, the timbre and / or pitch of the wake-up voice are obtained simultaneously, and the pitch and / or timbre is used as a label signal of the voice intensity. When a voice panel has multiple voice intensities carrying different label signals at the same time, the multiple voice intensities are broadcast simultaneously.
[0052] Step S3 is specifically as follows: the voice panel that receives the wake-up voice compares its own voice strength with the broadcast voice strengths with the same label signal one by one, and responds when it is greater than all the compared broadcast voice strengths, otherwise it does not act. When a voice panel has multiple voice strengths with different label signals at the same time, each voice strength is compared with the broadcast voice strength with the same label, and responds when at least one voice strength is greater than all the compared broadcast voice strengths, otherwise it does not act; and when a voice panel has multiple voice strengths with different label signals at the same time, and there are multiple voice strengths greater than all the compared broadcast voice strengths, it responds once.
[0053] like Figure 4 As shown, user 1 and user 2 send out the wake-up voice "Xiao k Xiao K" at the same time, and voice panels A and B receive the wake-up voice of user 1, and voice panels B and C receive the wake-up voice of user 2. Here, the label signal for the wake-up voice of user 1 is represented by "1", and the label signal for the wake-up voice of user 2 is represented by "2". Voice panel A has its own voice strength A1, and receives the broadcast voice strength B1 for the wake-up voice of user 1 and the broadcast voice strength B2 for the wake-up voice of user 2 from voice panel B, and receives the broadcast voice strength C2 for the wake-up voice of user 2 from voice panel C; voice panel B has its own voice strength B1 and its own voice strength B2, and receives the broadcast voice strength C2 and the broadcast voice strength A1; voice panel C has its own voice strength C1, and receives the broadcast voice strength A1, the broadcast voice strength B1, and the broadcast voice strength B2; voice panel D receives the broadcast voice strength C2, the broadcast voice strength B2, the broadcast voice strength B1, and the broadcast voice strength A1. In voice panel A, compare the own voice intensity A1 with the broadcast voice intensity B1; in voice panel B, compare the own voice intensity B1 with the broadcast voice intensity A1, and compare the own voice intensity B2 with the broadcast voice intensity C2. When the own voice intensity B1 is greater than the broadcast voice intensity A1 or the own voice intensity B2 is greater than the broadcast voice intensity C2, voice panel B responds. When the own voice intensity B1 is greater than the broadcast voice intensity A1 and the own voice intensity B2 is greater than the broadcast voice intensity C2, a response is made. The response is directed to both user 1 and user 2.
[0054] Through the above method, even if two people send out wake-up voices at the same time, the voice panel can distinguish the wake-up voices sent by different users, and finally one voice panel can respond to two users at the same time, or two voice panels can respond to two users respectively as needed. This can avoid the situation where when two people send out wake-up voices at the same time, only one voice panel closer to one of the users responds, while this voice panel is far away from the other user, giving the other user a poor user experience.
[0055] Preferably, in this embodiment, if the voice panel receives multiple wake-up voices at the same time, and if the voice panel is finally awakened, it means that in the subsequent control command, the voice panel is likely to receive the control voices of different users at the same time. In the control command issued by the subsequent user, the tone / timbre of the control command is recognized at the same time. If it has the same tone / timbre as the wake-up voice that wakes up the voice panel, it acts according to the command, otherwise it does not act. It is used to avoid the problem that the voice panel simultaneously receives and processes control commands from two users during the subsequent voice control process. For example, the two voice panels in the bedroom and the living room are awakened by two users at the same time. The living room user issues a control command of "turn on the air conditioner", and the bedroom user issues a control command of "turn on the lights". Both voices do not have positions. At this time, it is easy to cause the problem of control misalignment. And through the above method, it can be ensured that the voice panel in the bedroom only responds to the bedroom user, and the voice panel in the living room only responds to the living room user. Of course, if the voice panel only receives one wake-up voice and is finally awakened, in the subsequent control command, the timbre / tone is not recognized, and the control is directly made according to the control voice of any user.
[0056] In addition, it should be noted that the preset time periods mentioned here are all time periods starting from the time point when the corresponding wake-up voice is received.
[0057] Embodiment 4
[0058] like Figure 5 As shown, this embodiment is similar to the first embodiment, except that step S3 of this embodiment is:
[0059] The voice panel that receives the wake-up voice first determines whether its own voice intensity is greater than the intensity threshold. If so, it responds directly and the voice panel is awakened. Otherwise, it compares its own voice intensity with the received broadcast voice intensity one by one. When it is greater than all the compared broadcast voice intensities, it responds and the voice panel is awakened. Otherwise, it does not take any action and the voice panel is not awakened.
[0060] The intensity threshold here is set by technical personnel in this field according to actual conditions. Each whole-house smart system can set the intensity threshold according to specific scenarios, user conditions, switch panel spacing, etc. Generally, the intensity threshold set for scenarios with a larger switch panel spacing is larger, and the intensity threshold set for scenarios with a smaller switch panel spacing is smaller; the intensity threshold set for scenarios with a louder user voice is larger, and the intensity threshold set for scenarios with a softer user voice is smaller.
[0061] When the voice intensity of a certain voice panel is greater than the intensity threshold, it can be considered that the user is close to the voice panel, such as standing in front of the voice panel, facing or sideways to the voice panel. At this time, the voice panel responds directly without waiting for the comparison result of the broadcast voice intensity broadcast by other voice panels, which can improve the response speed. Other voice panels will receive the broadcast voice intensity of the voice panel as usual. Since the user is close to the voice panel, the voice intensity of other voice panels will be lower than the voice intensity of the voice panel, so other voice panels can still remain silent.
[0062] This embodiment sets a strength threshold for the system. When a user approaches a voice panel and issues a wake-up voice, the voice panel can be quickly woken up. At the same time, other voice panels continue to remain silent through strength comparison, further optimizing the nearby wake-up solution and effectively improving the user experience.
[0063] The specific embodiments described herein are merely examples of the spirit of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in similar ways, but they will not deviate from the spirit of the present invention or exceed the scope defined by the appended claims.
[0064] Although the terms such as voice panel, wake-up voice, own voice intensity, broadcast voice intensity, etc. are used more frequently in this article, the possibility of using other terms is not excluded. These terms are used only to more conveniently describe and explain the essence of the present invention; interpreting them as any additional limitation is contrary to the spirit of the present invention.
Claims
1. A method for waking up a nearby voice device in a whole-house intelligent system, characterized in that: It includes multiple voice panels, and the multiple voice panels respond to the user's wake-up voice in the following arbitration method: S1. At least one voice panel in the local area network receives the wake-up voice; S2. The voice panels receiving the wake-up voice obtain the voice strength of the wake-up voice received and broadcast the voice strength; S3. The voice panel that receives the wake-up voice first determines whether the voice intensity is greater than the intensity threshold. If so, it responds directly. Otherwise, it compares its voice intensity with the broadcast voice intensity received within the first preset time period. When it is greater than all the broadcast voice intensities compared with it, it responds, otherwise it does not act; If the broadcast voice intensity is not received within the first preset time period, the voice intensity comparison is not performed, and the voice panel directly responds; A communication table is stored in each voice panel. In step S3, when there is a broadcast voice intensity equal to the voice panel's own voice intensity and there is no broadcast voice intensity greater than its own voice intensity, it is detected whether a response result unicast signal from other voice panels is received. If not, a response is made and the response result is unicasted to the voice panel that broadcasts a broadcast voice intensity equal to its own voice intensity.
2. The method for waking up a nearby voice device in a whole-house intelligent system according to claim 1, characterized in that: The first preset time period is less than or equal to 0.1 s and greater than or equal to 0.07 s.
3. The method for waking up a voice device nearby in a whole-house intelligent system according to claim 1, characterized in that: The first preset time period is set according to the time required for the voice panel at the installation site to receive the broadcast voice intensity of the adjacent voice panel.
4. The method for waking up a voice device nearby in a whole-house intelligent system according to claim 1, characterized in that: In step S2, the timbre and / or pitch of the wake-up speech are obtained simultaneously, and the pitch and / or timbre are used as label signals of speech intensity, and in step S3, speech intensities with the same label signals are compared with each other, and speech intensities with different label signals are not compared with each other.
5. The method for waking up a voice device nearby in a whole-house intelligent system according to claim 4, characterized in that: Step S3 is specifically as follows: The voice panel that receives the wake-up voice compares its own voice strength with the broadcast voice strengths with the same tag signal one by one. When the strength is greater than all the compared broadcast voice strengths, it responds, otherwise it does nothing.
6. The method for waking up a voice device nearby in a whole-house intelligent system according to claim 5, characterized in that: In step S3, when a voice panel has multiple self-voice intensities carrying different label signals at the same time, each self-voice intensity is compared with the broadcast voice intensity with the same label. When there is at least one self-voice intensity greater than all the broadcast voice intensities compared with it, a response is made, otherwise no action is taken.
7. The method for waking up a nearby voice device in a whole-house intelligent system according to claim 6, characterized in that: In step S3, when a voice panel simultaneously has multiple own voice intensities carrying different label signals, and there are multiple own voice intensities that are greater than all broadcast voice intensities compared with them, a response is made.
8. The method for waking up a voice device nearby in a whole-house intelligent system according to claim 7, characterized in that: In step S2, when a voice panel simultaneously has multiple voice intensities carrying different tag signals, the multiple voice intensities are broadcast simultaneously.
Citation Information
Patent Citations
Smart Speaker Wake-Up Method and Device, Smart Speaker and Storage Medium
US20210151048A1
Cited By
Intelligent voice panel awakening method and system
CN117354077A
