Voice air conditioner response parameter automatic adjusting method and voice air conditioner system

Abstract

The invention relates to the technical field of air conditioners, discloses a voice air conditioner response parameter automatic adjusting method and a voice air conditioner system, and solves the problem that user experience is affected due to the fact that existing voice air conditioner response parameters are fixed, perception is simplified and interaction lacks pertinence. In the scheme of the invention, space perception and sound perception are respectively carried out based on a millimeter wave radar and a matrix microphone, target personnel information is obtained through signal processing, a voice control instruction is separated from environment sound, and a sound source direction is estimated; the position information of the target person is matched with the sound source direction, the position information of the current speaker is locked, a corresponding response parameter adjustment strategy is generated, and a volume adjustment control instruction and a directional sound beam control instruction are generated according to the response parameter adjustment strategy; and finally, according to the volume adjustment control instruction, controlling the audio amplifier to adjust the volume of the voice response, and according to the directional sound beam control instruction, controlling the multichannel loudspeaker array to adjust the sound direction of the voice response.

Description

Technical Field

[0001] This invention relates to the field of air conditioning technology, specifically to a method for automatically adjusting voice air conditioning response parameters and a voice air conditioning system. Background Technology

[0002] With the deep integration of artificial intelligence and Internet of Things technologies, intelligent voice interaction has become one of the core functions of home appliances. Voice-controlled air conditioners, as commonly used smart devices in the home, are seeing their popularity continue to rise due to their advantages of being manual-free and easy to control. When using a voice-controlled air conditioner, users can perform operations such as turning it on / off, adjusting the temperature, and switching modes by issuing voice control commands. The device's voice response function (such as command confirmation, status feedback, and operation prompts) is a key link in the interaction loop and directly affects the user experience.

[0003] Currently, voice-activated air conditioners on the market generally suffer from a lack of standardized and intelligent voice response functions. On the one hand, the response output parameters are mostly preset fixed values, or only support manual adjustment by the user, which cannot adapt to dynamically changing usage scenarios. For example, when a user issues a control command from a distance in the living room, the fixed volume may be too far away to hear clearly; while when the user moves closer to the air conditioner, the same volume may be too close and sound abrupt, even causing auditory discomfort; in noisy environments such as the kitchen or living room gatherings, the fixed volume response is easily masked by ambient noise, and the user cannot get timely feedback; while in quiet scenarios such as the bedroom, the fixed volume may disrupt the tranquility of the environment. This "one-size-fits-all" response mode requires users to frequently adjust the volume manually, which is cumbersome and results in a poor user experience.

[0004] On the other hand, existing voice-activated air conditioners mostly have omnidirectional or fixed unidirectional outputs, lacking specificity. When there are multiple people in the room, the omnidirectional output will spread the response sound throughout the entire space, which may both disturb those who are not interacting and fail to ensure that the speaking user receives clear auditory feedback; while the fixed unidirectional output response may become blurred due to the user moving, making the voice interaction process lack human touch and immersion.

[0005] In addition, some existing voice devices attempt to adjust the response volume by collecting ambient noise through microphones, but relying solely on acoustic perception cannot accurately obtain spatial information such as the distance between the user and the device and the real-time orientation, resulting in a lack of targeted volume adjustment. While solutions using visual sensors such as cameras can achieve spatial positioning, they are greatly affected by lighting conditions and pose a risk of user visual privacy leakage, making it difficult to meet the privacy protection needs of home scenarios.

[0006] In summary, existing voice-activated air conditioning systems suffer from fixed parameters, limited perception, and a lack of targeted interaction. They also fail to dynamically adapt to complex scenarios such as changes in user location and environmental noise levels, resulting in a subpar user experience. Therefore, there is an urgent need for a technical solution that can dynamically perceive user status and environmental information and automatically adjust response parameters. Summary of the Invention

[0007] The technical problem to be solved by the present invention is to provide a method for automatically adjusting voice air conditioning response parameters and a voice air conditioning system, so as to solve the problems of fixed response parameters, uniform perception, and lack of targeted interaction in existing voice air conditioning systems, which affect the user experience.

[0008] The technical solution adopted by the present invention to solve the above-mentioned technical problems is as follows: On one hand, the present invention provides a voice-activated air conditioning system, comprising: The sensing sensor unit includes a millimeter-wave radar and a matrix microphone. The millimeter-wave radar is used for spatial perception of people in the room, and the matrix microphone is used for collecting sound signals from the environment. The sensor data processing unit includes a radar signal processing module and a microphone array signal processing module; the radar signal processing module is used to process the raw signals acquired by the millimeter-wave radar to obtain the position and motion status information of the target personnel; the microphone array signal processing module is used to process the sound signals acquired by the matrix microphone, separate the voice control commands, and estimate the direction of the sound source. The SOC decision unit is used to match the location information of the target person with the direction of the sound source, lock the location information of the current speaker, generate a corresponding response parameter adjustment strategy, and generate volume adjustment control instructions and directional sound beam control instructions according to the response parameter adjustment strategy. The execution unit includes an audio amplifier and a multi-channel speaker array. The audio amplifier is used to execute volume adjustment control commands to adjust the volume of the voice response, and the multi-channel speaker array is used to execute directional beam control commands to adjust the sound direction of the voice response.

[0009] On the other hand, based on the above system, the present invention also provides a method for automatically adjusting voice air conditioning response parameters, applied to the above-mentioned voice air conditioning system, the method comprising the following steps: S1. Continuously perceive the space of people in the room based on millimeter-wave radar; collect sound signals in the environment based on matrix microphones; S2. Process the raw signals acquired by the millimeter-wave radar to obtain the location, motion status, and vital signs of the target personnel; process the sound signals acquired by the matrix microphone to separate voice control commands and estimate the direction of the sound source; S3. Match the location information of the target person with the direction of the sound source, lock the location information of the current speaker, generate a corresponding response parameter adjustment strategy, and generate volume adjustment control instructions and directional sound beam control instructions according to the response parameter adjustment strategy; S4. According to the volume adjustment control command, control the audio amplifier to adjust the volume of the voice response, and according to the directional sound beam control command, control the multi-channel speaker array to adjust the sound direction of the voice response.

[0010] Furthermore, in step S1, the raw data collected by the millimeter-wave radar for spatial perception of people in the room includes: the distance, speed, and angle of the target person.

[0011] Furthermore, in step S2, the processing of the raw signal acquired by the millimeter-wave radar includes: The DBSCAN clustering algorithm is used to remove interference points from the original signal, and then the Kalman filter algorithm is used to track the target person, outputting 3D coordinates, movement speed and vital signs information.

[0012] Furthermore, in step S2, the processing of the sound signals acquired by the matrix microphone includes: The GCC-PHAT algorithm is used to calculate the time difference of sound waves arriving at each microphone and estimate the direction of the sound source. The MVDR beamforming algorithm is used to suppress ambient noise from directions other than the sound source, obtain voice control command signals, and calculate the ambient noise decibel value.

[0013] Furthermore, in step S3, the location information of the target person is matched with the direction of the sound source to lock the location information of the current speaker, including: The direction of the sound source located by the microphone is matched one by one with the 3D coordinates of multiple target personnel tracked by the radar. The angular deviation between the direction of the sound source and the position of each target is calculated, and candidate targets with a deviation less than the threshold are selected. If there are multiple candidate targets, the current main interactive object is locked by combining the movement trajectory of each candidate target and its distance from the device, and the 3D coordinates of the target provided by the radar are used as the position of the current speaker.

[0014] Furthermore, in step S3, the response parameter adjustment strategy includes: the beam pointing angle and volume gain value of the response; wherein, the beam pointing angle is calculated based on the current speaker's position and direction information; and the volume gain value is calculated based on the current speaker's distance from the device and the ambient noise level in decibels.

[0015] Furthermore, in step S3, if the current speaker's position changes, a response parameter adjustment strategy is generated based on the latest position information collected by the millimeter-wave radar, and the volume and direction of the voice response are adjusted synchronously in step S4.

[0016] The beneficial effects of this invention are: (1) The response experience is humanized, completely eliminating human intervention: Millimeter-wave radar accurately captures the user's real-time location and distance. Combined with beamforming technology of the speaker array, the response sound is directed and focused on the user speaking, ensuring that the user receives feedback clearly while avoiding sound diffusion that may interfere with people in other areas.

[0017] In addition, the system dynamically adjusts the volume based on the distance between the user and the air conditioner. The volume automatically increases when the user moves away and decreases when the user moves closer, avoiding the problem of being startled by loud noise at close range and not being able to hear clearly at a distance. At the same time, combined with the ambient noise sensed by the matrix microphone, the system automatically increases the volume gain in noisy scenes and actively lowers the volume in quiet scenes, without requiring the user to manually adjust it.

[0018] (2) Multimodal perception complementarity, balancing accuracy and privacy: Millimeter-wave radar is unaffected by light and can accurately identify user location, movement trajectory, and stationary / micro-movement status; matrix microphones use beamforming technology to suppress environmental noise and accurately locate sound sources. After data fusion, the limitations of a single sensor are effectively compensated for, ensuring the accuracy of speaker locking and accurately identifying the main interaction object even when multiple people are together.

[0019] In addition, millimeter-wave radar does not require the acquisition of visual images; it senses the user's status only through spatial signals, thus fundamentally avoiding the privacy leakage risks associated with camera-based solutions.

[0020] (3) Intelligent adaptation to multiple scenarios, covering diverse usage needs: Millimeter-wave radar can track multiple people simultaneously. The system combines the direction of the sound source with the movement trajectory of the people to determine the main interaction target, responding only to the main user, adapting to complex scenarios where multiple people are together in a family. When the user moves, the radar updates the location data in real time, and the system adjusts the response direction and volume synchronously to ensure the continuity of interaction and achieve an immersive interaction where "sound follows the person." If the speaker changes, the "auditory focus" is quickly switched through sound source localization without the need to re-trigger commands, adapting to dynamic scenarios such as user movement and multiple people taking turns interacting. Attached Figure Description

[0021] Figure 1 This is a structural block diagram of the voice-activated air conditioning system in an embodiment of the present invention.

[0022] Figure 2This is a flowchart of the automatic adjustment method for voice air conditioning response parameters in an embodiment of the present invention. Detailed Implementation

[0023] This invention aims to provide a method and system for automatically adjusting voice air conditioning response parameters, addressing the problems of fixed response parameters, simplistic perception, and lack of personalized interaction in existing voice air conditioning systems, which negatively impact user experience. Its core idea is to construct an intelligent interaction system based on multimodal perception fusion and dynamic adaptive decision-making, abandoning the traditional design of fixed voice air conditioning response parameters. By leveraging the complementary advantages of millimeter-wave radar and matrix microphones, it accurately captures user status and environmental information, achieving automated and personalized adjustment of response parameters.

[0024] The technical means employed by this invention to achieve the above-mentioned core ideas include: (1) Dual-sensing dimension collaborative acquisition, breaking through the limitations of a single sensor: millimeter-wave radar is responsible for spatial dimension perception, accurately acquiring the user's real-time 3D coordinates, distance, motion trajectory and static / micro-motion state; matrix microphone is responsible for acoustic dimension perception, collecting environmental noise, locating the direction of the sound source and reducing noise through beamforming. The two are collected simultaneously to form a full-dimensional data input of space + acoustics, solving the problems of inaccurate positioning and weak anti-interference of a single sensor.

[0025] (2) Multimodal data fusion decision-making to accurately lock the interactive target: Through the SOC multimodal data decision-making unit, the spatial location data of the radar and the sound source direction data of the microphone are matched. For example, the radar identifies multiple people in the room. After the microphone locates the sound source direction, the system can accurately lock which person is speaking and use the centimeter-level 3D coordinates of the radar as the final interactive target to ensure the accuracy of target locking and avoid misjudgment.

[0026] (3) Dynamic adaptation of response parameter adjustment strategy, taking into account personalization and scenario compatibility: Based on the locked user coordinates and environmental data, a dual adaptive strategy is formulated: First, orientation adaptation, by using the beamforming technology of the speaker array to focus the response sound beam on the user's area, reducing interference to other people; Second, volume adaptation, by combining the distance between the user and the device (increase the volume when far away, decrease it when close) and the intensity of environmental noise (increase the gain when noisy, decrease the volume when quiet), the response volume is adaptively adjusted.

[0027] (4) Continuous tracking and optimization to adapt to dynamic interaction scenarios: The system is not limited to parameter adjustment for a single interaction, but uses the real-time tracking capability of millimeter-wave radar to update the response direction and volume in sync with the user's movement, so as to achieve "sound follows the person"; at the same time, it supports multi-target recognition and main user determination (such as prioritizing people facing the device), and quickly switches the interaction focus in multi-person scenarios to ensure the continuity and accuracy of interaction in dynamic scenarios.

[0028] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0029] This embodiment first provides a voice-activated air conditioning system, see [link]. Figure 1 It comprises several parts: a sensing sensor unit, a sensor data processing unit, a SOC decision-making unit, and an execution unit; detailed descriptions are as follows: The sensing sensor unit includes millimeter-wave radar and a matrix microphone.

[0030] Millimeter-wave radar: Responsible for spatial perception. It can accurately detect and track the real-time location (distance, angle) of one or more people in a room. Its advantages are that it is not affected by light and can provide accurate distance information while protecting the user's visual privacy.

[0031] Matrix microphones: Responsible for acoustic sensing, they use an array of multiple microphones to locate sound sources and determine the speaker's direction by utilizing the time difference between the arrival times of sound waves at different microphone heads. More importantly, they can use beamforming technology to create an "auditory focus," picking up sound only from a specific direction and suppressing ambient noise and reverberation.

[0032] The sensor data processing unit includes a radar signal processing module and a microphone array signal processing module.

[0033] Radar signal processing module: Used to process the raw signals received by the radar, including distance, speed and angle.

[0034] Output: Generate point cloud data and output the real-time 3D coordinates (x, y, z), motion speed, and vital signs information of one or more targets through clustering and tracking algorithms.

[0035] Microphone array signal processing module: used to process signals from multiple microphones.

[0036] Output: The sound source direction (θ, φ) is estimated using algorithms such as GCC-PHAT and MUSIC; a pickup beam is formed using algorithms such as MVDR and GSC, pointing towards the main speaker and suppressing noise from other directions, outputting a clear speech signal.

[0037] The SOC decision unit is used to locate the current speaker's position information through multimodal data fusion, generate corresponding response parameter adjustment strategies, and generate volume adjustment control commands and directional sound beam control commands based on the response parameter adjustment strategies.

[0038] Multimodal data fusion: This process receives and merges "position data" from radar and "direction data" from microphones. It correlates the direction of the sound source located by the microphone with the target location tracked by the radar to confirm "which person is speaking." For example, if the radar detects two people, A and B, and the microphone detects sound coming from the direction of A, the system identifies A as the current speaker, using the precise 3D coordinates of A provided by the radar as the final "auditory focus." Data fusion overcomes the limitations of single technologies (e.g., microphones may be interfered with by noise, and radar may not be able to directly identify who is speaking). After fusion, the system can uniquely determine the speaker's precise coordinates.

[0039] Response parameter adjustment strategy generation: Based on the target location output by the fusion engine, control commands are generated, including volume adjustment control commands and directional sound beam control commands; Azimuth control commands: Calculate the parameters required to drive the speaker array, form the emitted sound beam, and point it towards the target; Volume control commands: Dynamically calculate the required gain using formulas based on the speaker's current distance and the current ambient noise.

[0040] The execution unit includes an audio amplifier and a multi-channel speaker array.

[0041] An audio amplifier is used to execute volume adjustment control commands, thereby adjusting the volume of voice responses to achieve an "equal loudness" experience, meaning that the volume heard is similar regardless of the user's distance.

[0042] A multi-channel speaker array is used to execute directional sound beam control commands. By adjusting the phase and amplitude of each channel, a directional sound beam is formed to ensure that the sound energy is concentrated on the current speaker, making it easier for them to hear, while also reducing interference to other areas.

[0043] Based on the above-described voice-activated air conditioning system, this embodiment also provides a method for automatically adjusting voice-activated air conditioning response parameters, see [link to relevant documentation]. Figure 2 It includes the following implementation process: 1. Synchronous detection: Millimeter-wave radar continuously scans space to collect raw point cloud data; matrix microphones continuously monitor ambient sounds to achieve uninterrupted capture of acoustic information.

[0044] The radar signal processing module receives raw point cloud data, removes static interference points such as furniture and walls using the DBSCAN clustering algorithm, and then tracks the target personnel using the Kalman filter algorithm, outputting stable 3D coordinates, movement speed, and vital signs information. The microphone array signal processing module processes multi-channel audio data: it calculates the time difference of sound waves arriving at each microphone using the GCC-PHAT algorithm to estimate the direction of the sound source (θ, φ); it forms an "auditory focus" using the MVDR beamforming algorithm to suppress environmental noise from non-sound source directions, outputs clear voice control command signals, and calculates the environmental noise decibel value N.

[0045] 2. Target Locking: When someone starts speaking, the microphone array locates the direction of the sound source, and the radar provides information on the person's location. This information is then fused with the direction of the sound source to accurately pinpoint the speaker's coordinates.

[0046] In one exemplary implementation, data fusion includes: matching the sound source direction (θ, φ) located by the microphone with the 3D coordinates of multiple target individuals tracked by radar one by one; calculating the angular deviation between the sound source direction and the position of each target; and filtering out candidate targets with a deviation ≤10°. If there is only one candidate target, that candidate target is the current speaker, and the precise position of the current speaker can be obtained based on the position information provided by the radar. If there are multiple candidate targets, combining the movement trajectory of the candidate targets (prioritizing targets facing the air conditioner) and their distance from the device (prioritizing those closer), a unique primary interaction object is identified, and the precise 3D coordinates of that target provided by the radar are used as the final "auditory focus".

[0047] 3. Intelligent decision-making: The system calculates the required beam pointing angle of the speaker array based on the speaker's position and generates directional beam control commands; it also calculates the required output volume gain based on the real-time distance between the speaker and the device, as well as the ambient noise, and generates volume adjustment control commands.

[0048] In one exemplary implementation, the two instructions are generated in the following ways: Directional beam control command generation: Calculate the beam pointing angle (azimuth θ, elevation φ) based on the speaker's 3D coordinates, determine the phase and amplitude parameters of each channel speaker, and ensure that the beam is accurately focused on the speaker (main user).

[0049] Volume adjustment control command generation: Combining the speaker's current distance d and the ambient noise N, and substituting them into a preset volume calculation model, the output volume V is dynamically calculated. out V out =V base +k1×(d-d0)+k2×(N-N0); Among them, V based0 is the reference volume value, which can be the comfortable response volume value in a normal distance and quiet environment, such as 50dB; d0 is the reference distance, which can be the normal interaction distance between the user and the air conditioner, such as 2m; N0 is the reference noise, which can be the typical noise value in a quiet scene such as a bedroom, such as 30dB; k1 is the distance adaptation coefficient, which can be 5dB / m; k2 is the noise adaptation coefficient, which can be 0.8.

[0050] 4. Dynamic adjustment: The audio amplifier adjusts its output power according to the volume adjustment control command, driving the speaker array to output the corresponding volume; the multi-channel speaker array adjusts the phase and amplitude of each channel according to the directional sound beam control command to form a directional sound beam, ensuring that the sound energy is concentrated and directed towards the speaker (main user).

[0051] 5. Continuous tracking and optimization: If the main user moves, the millimeter-wave radar updates its 3D coordinates in real time, and the system adjusts the beam pointing angle and volume synchronously to achieve an immersive experience of "sound following the user". If a new sound source is detected (such as a family member taking over the speaking), the microphone array quickly repositions the sound source direction, and the system repeats the process of target locking-intelligent decision-dynamic adjustment to switch the main interaction object without the user having to manually trigger it.

[0052] Although embodiments of the present invention have been described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the present invention, and all such changes and alterations shall not depart from the protection scope of the present invention.

Claims

1. A voice air conditioning system, characterized by, Comprise: A perception sensor unit comprising a millimeter wave radar for spatial perception of personnel in a room and a matrix microphone; The matrix microphone is used to collect sound signals in the environment; A sensor data processing unit comprising a radar signal processing module and a microphone array signal processing module; the radar signal processing module is used to process the raw signals collected by the millimeter wave radar to obtain the position and motion state information of the target personnel; the microphone array signal processing module is used to process the sound signals collected by the matrix microphone to separate out the voice control instruction and estimate the sound source direction; The SOC decision unit is used to match the position information of the target personnel with the sound source direction, lock the position information of the current speaker, and generate a corresponding response parameter adjustment strategy, and generate a volume adjustment control instruction and a directional sound beam control instruction according to the response parameter adjustment strategy; The execution unit comprises an audio amplifier and a multi-channel speaker array, the audio amplifier is used to execute the volume adjustment control instruction to adjust the volume of the voice response, and the multi-channel speaker array is used to execute the directional sound beam control instruction to adjust the sound direction of the voice response.

2. The method for automatically adjusting the voice air conditioning response parameter is applied to the voice air conditioning system as claimed in claim 1, characterized in that, The method comprises the following steps: S1. Continuously perceive the personnel in the room based on the millimeter wave radar; collect the sound signals in the environment based on the matrix microphone; S2. Process the raw signals collected by the millimeter wave radar to obtain the position, motion state information and vital signs of the target personnel; process the sound signals collected by the matrix microphone to separate out the voice control instruction and estimate the sound source direction; S3. Match the position information of the target personnel with the sound source direction, lock the position information of the current speaker, and generate a corresponding response parameter adjustment strategy, and generate a volume adjustment control instruction and a directional sound beam control instruction according to the response parameter adjustment strategy; S4. According to the volume adjustment control instruction, control the audio amplifier to adjust the volume of the voice response, and according to the directional sound beam control instruction, control the multi-channel speaker array to adjust the sound direction of the voice response.

3. The voice air conditioning response parameter automatic adjustment method of claim 2, wherein In step S1, the raw data collected by the millimeter wave radar for spatial perception of personnel in the room includes the distance, speed and angle of the target personnel.

4. The voice air conditioning response parameter automatic adjustment method of claim 2, wherein In step S2, processing the raw signals collected by the millimeter wave radar comprises: Eliminate interference points in the raw signals by DBSCAN clustering algorithm, and track the target personnel by Kalman filtering algorithm to output 3D coordinates, motion speed and vital signs information.

5. The voice air conditioning response parameter automatic adjustment method of claim 2, wherein In step S2, processing the sound signals collected by the matrix microphone comprises: Calculate the time difference of sound waves arriving at each microphone by GCC-PHAT algorithm to estimate the sound source direction; The environmental noise in the direction other than the sound source is suppressed by the MVDR beamforming algorithm to obtain a voice control instruction signal, and a decibel value of the environmental noise is calculated. 6.The voice air-conditioning response parameter automatic adjustment method of claim 2, wherein, In step S3, the position information of the target person is matched with the sound source direction to lock the position information of the current speaker, including: The sound source direction of the microphone positioning is matched with the 3D coordinates of the multiple target persons tracked by the radar one by one, the angle deviation of the sound source direction and each target position is calculated, and the candidate target with a deviation less than a threshold value is screened out; If there are multiple candidate targets, the current main interactive object is locked by combining the motion trajectory of each candidate target and the distance from the device, and the 3D coordinates of the target provided by the radar are used as the position of the current speaker. 7.The voice air-conditioning response parameter automatic adjustment method of claim 2, wherein, In step S3, the response parameter adjustment strategy includes a sound beam pointing angle and a volume gain value; wherein the sound beam pointing angle is calculated according to the position and direction information of the current speaker; and the volume gain value is calculated according to the distance between the current speaker and the device and the decibel value of the environmental noise. 8.The voice air-conditioning response parameter automatic adjustment method of claim 2, wherein, In step S3, if the position of the current speaker changes, the latest position information of the current speaker collected by the millimeter wave radar is used to generate a response parameter adjustment strategy, and the volume and sound direction of the voice response are adjusted synchronously in step S4.

Images