Air conditioner control method, device, apparatus and medium

Abstract

The application discloses an air conditioner control method, device, equipment and medium, the air conditioner control method obtains user breathing state information in real time through a millimeter wave radar, generates a voice prompt associated with the breathing state when detecting that the breathing state is abnormal and broadcasts the voice prompt. Subsequently, a voice feedback signal of the user to the prompt is received, and a voice recognition strategy is dynamically adjusted in combination with the current breathing state, and is parsed into an effective instruction. Finally, a control instruction is generated by fusing the breathing state and the voice instruction, and the air conditioner is adjusted to run. The scheme enables the air conditioner to actively initiate human-computer interaction, avoids the interaction loss caused by the traditional system depending on the user to actively speak, and significantly improves the initiative, interaction friendliness and user comfort experience of the intelligent air conditioner.

Description

Technical Field

[0001] This invention relates to the field of air conditioner control technology, and in particular to an air conditioner control method, device, equipment and medium. Background Technology

[0002] With the development of artificial intelligence and the Internet of Things (IoT) technologies, smart air conditioners have generally acquired voice control capabilities. Users can adjust operating parameters such as temperature and fan speed using voice commands, improving ease of operation. However, existing smart air conditioning systems still have significant shortcomings in human-computer interaction. The interaction mode heavily relies on users actively initiating voice commands, and the system itself lacks the ability to perceive and respond to users' actual physiological states and potential needs, resulting in a poor user experience. Summary of the Invention

[0003] The embodiments of the present invention provide an air conditioner control method, device, equipment and medium, which aims to solve the technical problem of poor user experience caused by insufficient active interaction between existing smart air conditioners and target users.

[0004] In a first aspect, embodiments of the present invention provide an air conditioner control method, applied to an air conditioner control system, the air conditioner control system including millimeter-wave radar and a voice interaction device, the method including: acquiring the breathing status information of a target user in real time through the millimeter-wave radar; If an abnormal breathing state is detected in the target user's breathing status information, an active interactive prompt associated with the breathing state is generated and broadcast through the voice interaction device; the user's voice feedback signal in response to the broadcast through the voice interaction device is received; based on the voice feedback signal, it is parsed into a voice command using a dynamically adjusted voice recognition strategy; an air conditioning control command is generated according to the breathing status information and the voice command, and the air conditioning operation status is adjusted based on the air conditioning control command.

[0005] Secondly, embodiments of the present invention also provide an air conditioner control device for executing the air conditioner control method described above.

[0006] Thirdly, embodiments of the present invention also provide a computer device, the computer device including a memory and a processor connected to the memory; the memory is used to store a computer program; the processor is used to run the computer program stored in the memory to perform the steps of the above-described air conditioner control method.

[0007] Fourthly, embodiments of the present invention also provide a computer-readable storage medium storing a computer program, the computer program including program instructions, which, when executed by a processor, can implement the steps of the above-described air conditioner control method.

[0008] Compared with the prior art, the beneficial effects of the present invention are: In the technical solution of this invention, the air conditioner control method acquires the user's breathing status information in real time using millimeter-wave radar. When an abnormal breathing status is detected, it proactively generates and broadcasts a related voice prompt. Subsequently, it receives the user's voice feedback signal to the prompt and dynamically adjusts the voice recognition strategy based on the current breathing status, parsing it into a valid command. Finally, it merges the breathing status and voice command to generate a control command to regulate the air conditioner's operation. This solution enables the air conditioner to proactively initiate human-machine interaction, avoiding the lack of interaction caused by traditional systems relying on the user's active voice input, significantly improving the initiative, interactivity, and user comfort of the smart air conditioner. Attached Figure Description

[0009] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 A flowchart of the air conditioner control method provided by the present invention; Figure 2 This is a first sub-flowchart of the air conditioner control method provided by the present invention; Figure 3 This is a second sub-flowchart of the air conditioner control method provided by the present invention; Figure 4 The third sub-flowchart of the air conditioner control method provided by the present invention; Figure 5 The fourth sub-flowchart of the air conditioner control method provided by the present invention; Figure 6 The fifth sub-flowchart of the air conditioner control method provided by the present invention; Figure 7 The sixth sub-flowchart of the air conditioner control method provided by the present invention; Figure 8 A schematic block diagram of a unit of an air conditioner control device provided by the present invention; Figure 9 A schematic block diagram of a computer device provided for an embodiment of the present invention. Detailed Implementation

[0011] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0012] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0013] It should also be understood that the terminology used in this specification is for the purpose of describing embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0014] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0015] This invention addresses the technical problem of poor user experience caused by insufficient proactive interaction between existing smart air conditioners and the target user. It discloses an air conditioner control method applied to an air conditioning control system, which includes millimeter-wave radar and a voice interaction device.

[0016] Reference Figures 1 to 7 The air conditioner control method includes the following steps: S110. Real-time acquisition of the target user's respiratory status information via the millimeter-wave radar; S120. If an abnormal breathing state is detected in the breathing state information of the target user, an active interactive prompt associated with the breathing state is generated and broadcast through the voice interaction device. S130. Receive the user's voice feedback signal in response to the voice interaction device broadcast by the voice interaction device. S140. Based on the voice feedback signal, the voice is parsed into a voice command using a dynamically adjusted voice recognition strategy; S150. Generate an air conditioning control command based on the breathing state information and the voice command, and perform air conditioning operation state adjustment based on the air conditioning control command.

[0017] Specifically, the millimeter-wave radar module can be installed below the air outlet of the air conditioner or in the center of the front panel, with its transmission direction facing the user's usual comfort zone, for non-contact, real-time acquisition of the target user's respiratory status information. Alternatively, the millimeter-wave radar module can be installed in other indoor smart home devices, achieving linkage with the air conditioner through a whole-house smart home communication system. The millimeter-wave radar operates in 60GHz frequency band frequency-modulated continuous wave mode, continuously transmitting swept-frequency signals and receiving echo signals reflected by the user's chest or abdomen micro-movements. The central controller demodulates and bandpass filters the received signals to remove environmental clutter and interference from the device's own vibrations. Subsequently, the main frequency component is extracted by performing a fast Fourier transform on the filtered micro-movement time-series data to obtain the respiratory frequency, and the respiratory amplitude is estimated by combining it with a peak detection algorithm. Based on the combination relationship between respiratory frequency and amplitude, the controller classifies the respiratory status information into states such as stable breathing, rapid breathing, deep breathing, or weak breathing.

[0018] When the system continuously detects abnormal breathing patterns in the target user, such as prolonged rapid or weak breathing, the central controller determines that the user may have potential comfort needs that they haven't actively expressed. At this point, it generates proactive interactive prompts related to the current breathing state. For example, when rapid breathing is detected, the system proactively announces through the speaker of the voice interaction device: "Your breathing is rapid; would you like to speed up the cooling?"; when slow and weak breathing is detected, it announces: "You seem a little drowsy; would you like to switch to sleep comfort mode?" In this way, the system can proactively initiate human-computer dialogue, guiding the user to complete the interaction, greatly improving the ease of use and intelligent experience.

[0019] The microphone array in the voice interaction device collects user voice feedback signals from the environment. The collected signals first undergo echo cancellation and spectral subtraction noise reduction processing to eliminate noise generated by the air conditioner fan, resulting in a clear voice signal. This clear voice signal is then sent to a voice recognition processor, which parses the voice commands based on a dynamically adjusted voice recognition strategy.

[0020] The central controller integrates the breathing status information with the parsed voice commands to generate the final air conditioning control commands. For example, when receiving a "cool down" command during rapid breathing, the system generates a rapid cooling control command, controlling the compressor to increase its output power and the fan to increase its speed; when receiving a "reduce fan speed" command during stable breathing, it generates a normal adjustment command, reducing the fan speed by one level. All control commands are sent to the air conditioner's actuators to precisely adjust the air conditioner's operating status, including temperature, fan speed, mode, and airflow direction. During long-term use, the system can also record user feedback on proactive prompts, such as confirmation and subsequent manual adjustments, to dynamically optimize the classification thresholds for breathing status and the adjustment strategies for command word weights, achieving increasingly intelligent and personalized services over time.

[0021] In one embodiment, step S140 is preceded by: S160. Dynamically adjust the parameters of the speech recognition strategy based on the breathing state information.

[0022] During the parsing process, the system dynamically adjusts its speech recognition strategy based on the previously acquired breathing status information. Specifically, when the breathing status information indicates that the user is in a state of rapid breathing, the system automatically lowers the confidence threshold of the speech recognition model and increases the weight of command words such as "cool down," "cool," and "cool quickly" in the language model. When the user is in a state of weak breathing or drowsiness, the system lowers the rejection threshold and increases the recognition priority of words such as "comfort" and "turn down the fan speed," thereby ensuring that speech commands can still be accurately parsed even when the user is not in an ideal vocal state.

[0023] In one embodiment, step S110 includes: S111. Receive reflected signals from the chest cavity or abdomen of the target user through the millimeter-wave radar to obtain micro-motion timing data; S112. Filter the micro-motion timing data to remove interference signals caused by the operation of the air conditioner fan or environmental vibration. S113. Based on the filtered micro-motion time series data, the respiratory frequency is extracted by fast Fourier transform, and the respiratory amplitude is obtained by peak detection algorithm; S114. Based on the breathing frequency and breathing amplitude, the breathing state information is classified into at least one of the following states according to the preset breathing state classification rules: steady breathing state, rapid breathing state, deep breathing state, and weak breathing state.

[0024] Millimeter-wave radar continuously transmits electromagnetic wave signals towards a preset monitoring area and receives echo signals reflected from the minute displacements of the human chest cavity or abdomen caused by respiratory movements. Because breathing causes periodic fluctuations in the chest cavity, with displacements typically on the order of millimeters or even sub-millimeters, these minute movements modulate the phase of the echo signal, producing a detectable micro-Doppler effect. The central controller acquires and analyzes these echo signals, extracting the micro-motion time-series data directly related to the chest cavity displacement. This data characterizes the sequence of changes in the distance from the user's chest cavity surface to the radar's radial direction over time.

[0025] To ensure the accuracy of micro-motion timing data, filtering is required to remove various interference signals. Specifically, the system uses the current operating parameters of the air conditioner, such as fan speed, air deflector oscillation angle, and compressor operating status, to construct an interference signal template. For example, when the fan operates at a specific speed, the periodic wind field disturbances it generates will form interference components of specific frequencies in the echo signal. The central controller uses adaptive filtering algorithms, such as notch filtering, to suppress these frequency components. Simultaneously, the system also uses voice activity detection signals to generate synchronization gating logic. When a user is detected speaking, high-frequency vibration components related to vocalization, such as micro-movements of the throat and lips, are temporarily suppressed in the micro-motion timing data. This prevents chest rise and fall during speaking from being misinterpreted as deeper breathing or rhythm disorder, thus ensuring the accuracy of breathing status judgment.

[0026] For the filtered micro-motion time-series data, the central controller employs signal processing algorithms for feature extraction. Specifically, by performing a Fast Fourier Transform on a continuous segment of time-series data, the time-domain signal is converted to the frequency domain, and its spectral characteristics are analyzed. The dominant frequency of respiratory movements is typically within the physiological frequency range of 0.1Hz to 1Hz. By detecting the dominant peak frequency in the spectrum, the system can determine the user's real-time respiratory rate. In the time domain, the system uses a peak detection algorithm to identify the peaks and troughs in the micro-motion waveform and calculates their difference as a quantitative estimate of the respiratory amplitude, which reflects the depth of a single breath.

[0027] Based on the extracted respiratory rate and amplitude, the central controller classifies the user's respiratory state according to a preset respiratory state classification rule. This classification rule is based on typical respiratory physiological parameters of adults and can be optimized through individualized calibration. For example, when the respiratory rate is between 12 and 18 breaths per minute and the respiratory amplitude is within the individual's normal range, it is determined to be a stable respiratory state; when the respiratory rate exceeds 20 breaths per minute, the amplitude is moderate or slightly low, and this persists for a certain period of time, it is determined to be a rapid breathing state; when the respiratory rate is below 10 breaths per minute and the respiratory amplitude increases significantly, it is determined to be a deep breathing state; when the respiratory rate is below 12 breaths per minute and the respiratory amplitude is consistently below the individual's baseline level, it is determined to be a weak breathing state. Furthermore, the system can also introduce sliding window and hysteresis comparison mechanisms, such as consistency judgment between short and long time windows, to avoid misjudgments caused by instantaneous signal jitter and ensure the stability of the respiratory state information output. This classification result serves as the core basis for subsequent dynamic adjustment of the speech recognition strategy and generation of proactive interactive prompts.

[0028] In one embodiment, step S120 includes: S121. Determine the abnormality type based on the respiratory status information. The abnormality type includes at least rapid breathing or weak breathing. S122. When the abnormality type is determined to be the shortness of breath state, a first voice prompt is generated. The first voice prompt is used to ask whether it is necessary to increase cooling or lower the temperature. S123. When the abnormality type is determined to be the weak breathing state, a second voice prompt is generated. The second voice prompt is used to ask whether it is necessary to enter the comfort mode, reduce the fan speed, or increase the temperature. S124. Play the corresponding voice prompt through the voice interaction device and listen to the target user's feedback voice signal within a preset time period.

[0029] The system acquires and analyzes the target user's breathing status information using millimeter-wave radar, classifying it into states such as stable breathing, rapid breathing, deep breathing, or weak breathing. If no valid voice command with a confidence level exceeding a threshold is identified within a continuous monitoring period (e.g., 10 seconds), the system determines that the user may have potential physiological comfort needs that they have not actively expressed, thus initiating an active interaction process. First, the central controller determines the current abnormality type based on the breathing status information. This abnormality type includes at least rapid breathing or weak breathing. Rapid breathing typically indicates that the user may feel hot due to an overheated environment, recovery from exercise, or physical discomfort; weak breathing is often associated with the user being relaxed, drowsy, preparing to sleep, or having low physical energy.

[0030] When the system determines that the abnormality is a state of rapid breathing, it generates a first voice prompt based on intelligent inference of the user's potential needs. This first voice prompt is designed as a guiding question, such as, "We've detected that your breathing is rapid; would you like us to speed up the cooling?" or "You seem a little hot; would you like us to lower the temperature?". This prompt is generated by the central controller using a pre-stored speech synthesis template and played through the speaker in the voice interaction device, initiating the interaction with the user in a gentle and natural tone. Correspondingly, when the abnormality is determined to be a state of weak breathing, the system generates a second voice prompt, focusing on improving the user's comfort and sleep quality, such as, "Your breathing is slow; would you like us to switch you to a comfortable sleep mode?" or "Would you like us to lower the fan speed and avoid direct airflow?" By associating different types of breathing abnormalities with specific comfort adjustment intentions, the system can provide precise and contextualized proactive services, avoiding the interference caused by indiscriminate prompts.

[0031] After the corresponding voice prompt is played through the voice interaction device, the system immediately enters a listening state, continuously collecting ambient sound using a microphone array, focusing on listening to the target user's feedback voice signal within a preset duration, such as 5 to 8 seconds. This listening process also employs Voice Activity Detection (VAD) technology to distinguish user responses from ambient noise. If positive feedback from the user is received within the preset duration, such as "Yes," "Okay," or "Turn it on," even if the voice is weak, the system lowers the rejection threshold due to the current weak breathing state. The central controller recognizes this feedback as a valid confirmation command and immediately executes the air conditioning adjustment operation corresponding to the prompt content, such as starting rapid cooling or switching to comfort mode. If negative feedback is received, such as "No need" or "Wait a moment," the system cancels the adjustment and resumes normal monitoring. If no valid feedback is received within the preset duration, the system marks this prompt as "unresponsive" and may decide whether to prompt again or stop prompting to avoid disturbance based on subsequent changes in breathing state. Through this closed-loop proactive interaction mechanism, the air conditioning system can initiate intelligent dialogue based on the user's physiological state without the user's active operation, realizing truly humanized and proactive intelligent service, and greatly enhancing the user's comfort experience in seamless interaction.

[0032] In one embodiment, step S160 includes: S161. When the breathing state information indicates that the target user is in the state of rapid breathing, the confidence threshold of the speech recognition strategy is reduced and the weight of the first preset type instruction word is increased. The first preset type instruction word is used to express the user's need for cooling. S162. When the breathing state information indicates that the target user is in a weak breathing state, the rejection threshold of the voice recognition strategy is reduced, and the weight of the second preset type instruction word is increased. The second preset type instruction word is used to express the user's need to improve user comfort.

[0033] When the breathing status information is analyzed and determined to indicate that the target user is experiencing rapid breathing, the system infers that the user may be feeling unwell due to a hot environment or physical activity. Their vocalization may manifest as unstable breathing, weak volume, or intermittent speech. To ensure accurate recognition of user intent even under these less-than-ideal conditions, the system makes the following adjustments: First, the central controller sends a control signal to the speech recognition processor, instructing it to lower the confidence threshold of the speech recognition strategy. For example, the confidence threshold used to determine a "valid instruction" is lowered from the standard value of 0.85 to 0.70, thereby reducing the recognition threshold and allowing more candidate words to be considered, avoiding direct rejection of instructions due to degraded speech signal quality. Simultaneously, the system increases the weight of "first preset type instruction words," which are specifically used to express the user's need for cooling, including but not limited to words such as "cool down," "refrigerate," "cool down," "lower the temperature," and "cool down quickly." This weight increase is achieved by modifying the prior probability of the corresponding word units in the language model, resulting in higher scores for sentence paths containing these keywords during decoding, significantly increasing the probability of them being selected as the final recognition result. For example, when a user utters an indistinct voice saying "air conditioning...cool...", the system is more likely to correctly identify it as a valid command for "air conditioning cooling" because it increases the weight of words such as "cool" and "cooling".

[0034] Accordingly, when the breathing state information indicates that the target user is in a state of weak breathing, the system determines that the user may be in a relaxed, drowsy, or sleep-ready state, and their voice characteristics are typically low volume, slow speech, and weak breath. In this case, to adapt to the user's gentle, low-intensity vocalization pattern, the system adopts different adjustment strategies: On the one hand, it lowers the rejection threshold of the speech recognition strategy, which is used to determine whether a speech segment is valid speech rather than noise. Lowering this threshold allows the system to more sensitively capture the user's whispered instructions, avoiding misjudgment as silence due to excessively low volume. On the other hand, the system increases the weight of "second preset type instruction words," which are specifically used to express the user's need for improved comfort and cover words related to creating a relaxing and comfortable environment, such as "comfort mode," "sleep mode," "gentle breeze," "natural breeze," "lower fan speed," "avoid direct airflow," and "increase temperature." By increasing the weight of these words, the system prioritizes semantic combinations related to comfort adjustment during the recognition process. For example, when a user whispers "wind... turn down..." while drowsy, the system can more accurately identify and associate the command with reducing the fan speed because comfort-related commands such as "lower the fan speed" have a higher weight.

[0035] The aforementioned adjustment process is dynamically triggered by the central controller based on real-time breathing status tags. Adjustment parameters can be stored in a local configuration table and support self-learning optimization based on long-term user habits. For example, if the system detects that a user frequently uses "cool down quickly" instead of "cool" when breathing rapidly, it can automatically increase the weight of the "cool down quickly" command. Through this dynamic adjustment mechanism deeply coupled with physiological states, the voice recognition strategy evolves from a "static, general" mode to a "dynamic, personalized" mode, effectively solving the technical problem of interaction failure in traditional smart air conditioners under specific user physiological states, and significantly improving the robustness of human-computer interaction and user experience.

[0036] In one embodiment, step S140 includes: S141. Perform echo cancellation and background noise suppression on the collected voice feedback signal to obtain a clear voice signal; S142. Extract the Mel frequency cepstral coefficient features from the clear speech signal and input them into a preset deep neural network model for acoustic modeling and decoding to obtain the initial speech command; S143. Based on the confidence threshold adjusted by the breathing state information and the command word weight, the initial voice command is re-evaluated, and the final voice command is output.

[0037] The microphone array in the voice interaction device acquires the raw audio signal, which includes the user's voice and ambient noise, in real time. To improve voice quality, the system performs front-end processing on this raw signal, focusing on echo cancellation and background noise suppression. The echo cancellation module uses the output signal of the air conditioner speaker as a reference, and estimates and subtracts the echo components generated by the air conditioner's own sounds, such as prompts and music, from the voice signal using an adaptive filtering algorithm. Background noise suppression employs spectral subtraction or a deep learning-based noise reduction model to suppress non-speech components such as fan operation and ambient noise, ultimately outputting a clear voice signal with a high signal-to-noise ratio, laying the foundation for subsequent recognition.

[0038] Subsequently, the system performs feature extraction and pattern recognition on the clear speech signal. Specifically, the speech signal is processed frame by frame. After pre-emphasis and windowing of each frame, its short-time power spectrum is calculated. Then, the power spectrum is weighted using a set of Mel filter banks that conform to the characteristics of human hearing, simulating the differences in human ear sensitivity to different frequencies. After taking the logarithm of the filter bank output, a discrete cosine transform is performed to extract the Mel frequency cepstral coefficients (MFCC) features of each frame of speech. Typically, energy, first-order difference, and second-order difference are combined to form a multi-dimensional feature vector. This feature sequence is then input into a preset deep neural network model for processing. In this embodiment, the deep neural network model can be a hybrid architecture of deep neural network-hidden Markov model (DNN-HMM). The DNN part is responsible for learning the complex nonlinear mapping relationship between MFCC features and the state of basic speech units, and outputting the posterior probability of each state. The HMM part, based on the temporal dynamic characteristics of speech, combines a language model customized for the air conditioning control scenario, including command words such as "turn on", "cool", "adjust temperature", and "adjust fan speed", and their grammatical structures, to decode the probability sequence output by the DNN, search for the optimal word string path, and thus obtain the preliminary recognition result, i.e., the initial speech command.

[0039] Finally, to achieve intelligent optimization of "breathing-voice coupling," the system re-evaluates the initial voice command based on the acquired breathing state information to output the final voice command. Specifically, the central controller uses the current breathing state label as contextual information to dynamically adjust the decision parameters for voice recognition. When the breathing state information indicates that the user is in a state of rapid breathing, the system lowers the confidence threshold for voice recognition and simultaneously increases the prior weights of command words such as "cool down," "cool," and "cool down quickly" in the language model; when the user is in a state of weak breathing or drowsiness, the system lowers the rejection threshold and increases the weights of words such as "comfort mode," "lower fan speed," and "sleep." The voice recognition processor uses the adjusted parameters to reorder and filter the confidence of the initial voice command and the candidate word sequence. If the confidence of a candidate command exceeds the new threshold after adjustment, it is confirmed as the final voice command; if there is still no high-confidence result, it can be marked as "recognition fuzzy" and the subsequent active confirmation process can be triggered in conjunction with the breathing state. Through this re-evaluation mechanism, the system can adapt to the vocal characteristics of users under different physiological states, significantly improving the robustness and accuracy of speech recognition in non-ideal scenarios.

[0040] In one embodiment, step S150 includes: S151. When the breathing state is the rapid breathing state and the voice command includes the first preset type command word, generate an air conditioning rapid cooling control command. S152. When the breathing state is the deep breathing state and the voice command includes the second preset type command word, generate an air conditioning comfort adjustment control command.

[0041] When the user's breathing is rapid and the voice command contains a first preset type of command word, the central controller determines that the user has a strong need for cooling, and that their physiological state is highly consistent with the command intent. At this time, the system generates a rapid cooling control command for the air conditioner. This command not only triggers the conventional cooling mode but also includes enhanced adjustment parameters, such as instructing the compressor to increase its output power to 120% of its rated power, or 20% within the inverter range, increasing the fan speed to the highest setting or 30% higher than the current setting, and controlling the air deflector to adjust to a wide-angle airflow mode to achieve a rapid drop in indoor temperature. This "rapid cooling" logic breaks through the traditional air conditioner mode that only adjusts based on a fixed temperature difference, achieving a high-intensity response to the user's immediate physiological needs and effectively alleviating the user's discomfort caused by the heat.

[0042] Accordingly, when the breathing state is deep breathing and the voice command contains a second preset type of command word, the central controller generates an air conditioning comfort adjustment control command. This command aims to create a quiet and gentle indoor environment, avoiding strong airflow from interfering with the user's relaxed state. Specifically, the control command reduces the fan speed to the lowest setting or one level lower than the current setting (approximately 10-15%), and adjusts the target temperature to a preset comfort range, such as 24°C to 26°C. This value can be dynamically adjusted based on ambient temperature, seasonal mode, and user's historical preferences. Simultaneously, it controls the air deflector to swing upwards or horizontally to prevent cold or hot air from blowing directly onto the body, and can also automatically turn off the air conditioning panel's indicator lights to reduce light pollution. At night or when the system detects that the user has been stationary for an extended period, it can further activate the "sleep curve" function, allowing the temperature to gradually change over subsequent periods to maintain comfort throughout the night.

[0043] The generation process of the aforementioned control commands embodies a deep integration of user breathing and user voice coupling logic. Breathing state information is not only used to optimize front-end voice recognition but also serves as crucial context for back-end control decisions, interpreting and reinforcing the semantics of voice commands. For example, in a state of rapid breathing, even if the voice command is only a vague "cold spot," the system tends to execute rapid cooling due to the correlation between the breathing state and the first preset command word. Conversely, receiving a command to "lower the fan speed" during a deep breathing state triggers a more thorough comfort mode than conventional adjustments. The central controller sends the generated control commands to the air conditioning actuators via the internal communication bus, including the compressor control module, the variable frequency fan drive module, and the air guide servo motor. Through this integrated control mechanism, the system achieves a shift from being driven by environmental parameters to being driven by the user's physiological state, significantly improving the intelligence, proactivity, and personalization of air conditioning services.

[0044] Furthermore, the air conditioner control method of the present invention also includes: broadcasting the air conditioner's operating status adjustment result or the response result of the active interactive prompt to the target user through the voice interaction device; continuously collecting the target user's breathing status information, voice signal and the correspondence of control results, and iteratively updating the judgment threshold of the breathing status, the adjustment strategy of the command word weight and the triggering condition of the active prompt according to the user's feedback behavior and historical correspondence.

[0045] After the air conditioning actuators, such as the compressor, fan, and air deflector, have completed their operating status adjustments, or after the user has responded to a proactive interactive prompt, the central controller broadcasts the result of this operation back to the target user through the voice interaction device, forming a complete interactive loop. This voice feedback not only informs the user of the operation result, enhancing the transparency of the interaction, but also helps the user confirm whether their understanding of the system is correct, providing a basis for subsequent learning.

[0046] Based on this, the system performs continuous data acquisition and iterative updates. The central controller continuously collects the target user's breathing status information, acquired voice signals, generated control commands, user feedback behavior to proactive prompts (such as confirmation, rejection, no response), and the final adjustment result, and establishes a historical correspondence database including breathing status, voice commands, user feedback, and control actions. The system adaptively iteratively updates the core algorithm parameters based on user feedback behavior and the long-accumulated historical correspondence. Specifically, when the system detects that a user frequently issues "cooling" commands at a breathing rate of 19 breaths / minute, and the current threshold defining "rapid breathing" as >20 breaths / minute fails to cover this scenario, the system can automatically and dynamically lower the user's rapid breathing state judgment threshold to 18-19 breaths / minute, achieving individualized calibration. Similarly, if data analysis shows that users tend to use "silent" rather than "sleep mode" to express comfort needs when breathing is weak, the system will automatically increase the weight of the word "silent" in the second preset type of command word, optimizing the command word weight adjustment strategy. In addition, the system can optimize the triggering conditions for proactive prompts based on the user's response rate to the prompts: if a user has a high response rate to the cooling prompt when "breathing rapidly", the prompt triggering delay will be shortened; if a user does not respond to or rejects the mode switching prompt when "deep breathing" for a long time, the triggering conditions will be relaxed or the prompt frequency will be reduced to reduce unnecessary disturbances.

[0047] This iterative update process can be completed locally by a lightweight learning algorithm built into the central controller, such as rule-based statistical learning or a small-scale neural network, ensuring user data privacy and security without relying on the cloud. The updated parameters are stored in the air conditioner's non-volatile memory and continuously applied in subsequent operations. Through this feedback and iterative mechanism, the air conditioning control system can not only achieve intelligent response in a single interaction, but also continuously learn and adapt to the user's individual physiological characteristics, language habits, and preferences as usage time increases. This makes breathing state recognition more accurate, voice recognition more appropriate, and proactive interaction more considerate, ultimately achieving a truly personalized, intelligent service experience that becomes increasingly user-friendly with use.

[0048] Figure 8 This is a schematic block diagram of an air conditioner control device 600 provided in an embodiment of the present invention. Figure 8 As shown, corresponding to the above air conditioner control method, the present invention also provides an air conditioner control device 600. This air conditioner control device 600 includes a unit for executing the above air conditioner control method, and the device can be configured in a desktop computer, tablet computer, smartphone, or other terminal.

[0049] Specifically, please refer to Figure 8 The air conditioner control device 600 includes: The breathing state acquisition unit 610 is used to acquire the breathing state information of the target user in real time through the millimeter-wave radar. The active interaction unit 620 is used to generate an active interaction prompt associated with the breathing state if an abnormal breathing state is detected in the breathing state information of the target user, and to broadcast it through the voice interaction device. The voice acquisition unit 630 is used to receive voice feedback signals from the user in response to the voice interaction device. The instruction parsing unit 640 is used to parse the voice feedback signal into a voice instruction using a dynamically adjusted voice recognition strategy. The control and adjustment unit 650 is used to generate air conditioning control commands based on the breathing state information and the voice commands, and to perform air conditioning operation state adjustment based on the air conditioning control commands.

[0050] In one embodiment, the air conditioner control device 600 further includes: The strategy dynamic adjustment unit is used to dynamically adjust the parameters of the speech recognition strategy based on the breathing state information.

[0051] In one embodiment, the respiratory status acquisition unit 610 includes: The micro-motion timing data acquisition unit is used to receive reflected signals from the chest cavity or abdomen of the target user through the millimeter-wave radar to acquire micro-motion timing data. The filtering unit is used to filter the micro-motion timing data to remove interference signals caused by the operation of the air conditioner fan or environmental vibration. The amplitude and frequency acquisition unit is used to extract the respiratory frequency through fast Fourier transform based on the filtered micro-motion time series data, and to obtain the respiratory amplitude through a peak detection algorithm. The breathing state determination unit is used to classify the breathing state information into at least one of the following states according to the breathing frequency and breathing amplitude and a preset breathing state classification rule: steady breathing state, rapid breathing state, deep breathing state, and weak breathing state.

[0052] Furthermore, the active interaction unit 620 includes: An abnormality determination unit is used to determine the type of abnormality based on the respiratory status information, wherein the abnormality type includes at least a rapid breathing state or a weak breathing state. The first voice prompt unit is used to generate a first voice prompt when the abnormality type is determined to be the shortness of breath state. The first voice prompt is used to ask whether it is necessary to increase cooling or lower the temperature. The second voice prompt unit is used to generate a second voice prompt when the abnormality type is determined to be the weak breathing state. The second voice prompt is used to ask whether it is necessary to enter the comfort mode, reduce the fan speed, or increase the temperature. An interactive feedback unit is used to play the corresponding voice prompt through the voice interaction device and listen to the feedback voice signal of the target user within a preset time period.

[0053] Furthermore, the strategy dynamic adjustment unit includes: A confidence threshold adjustment unit is used to reduce the confidence threshold of the speech recognition strategy and increase the weight of a first preset type instruction word when the breathing state information indicates that the target user is in the state of rapid breathing. The first preset type instruction word is used to express the user's need for cooling. The rejection threshold adjustment unit is used to reduce the rejection threshold of the speech recognition strategy and increase the weight of the second preset type instruction word when the breathing state information indicates that the target user is in the weak breathing state. The second preset type instruction word is used to express the user's need to improve user comfort.

[0054] Furthermore, the instruction parsing unit 640 includes: The noise reduction unit is used to perform echo cancellation and background noise suppression on the acquired voice feedback signal to obtain a clear voice signal; The initial voice command acquisition unit is used to extract the Mel frequency cepstral coefficient features from the clear voice signal and input them into a preset deep neural network model for acoustic modeling and decoding to obtain the initial voice command. The final voice command acquisition unit is used to re-evaluate the initial voice command based on the confidence threshold and command word weight adjusted according to the breathing state information, and output the final voice command.

[0055] In one embodiment, the control adjustment unit 650 includes: The first instruction generation unit is used to generate an air conditioning rapid cooling control instruction when the breathing state is the rapid breathing state and the voice instruction includes the first preset type instruction word. The second instruction generation unit is used to generate an air conditioning comfort adjustment control instruction when the breathing state is the deep breathing state and the voice instruction includes the second preset type instruction word.

[0056] In one embodiment, the air conditioner control device 600 further includes: The voice feedback unit is used to broadcast the air conditioner's operating status adjustment result or the response result of the proactive interaction prompt to the target user through the voice interaction device. The logic iteration unit is used to continuously collect the correspondence between the target user's breathing state information, voice signals and control results, and iteratively update the judgment threshold of the breathing state, the adjustment strategy of the command word weight and the triggering condition of the active prompt based on the user's feedback behavior and historical correspondence.

[0057] The aforementioned air conditioner control device 600 can be implemented as a computer program, which can, for example... Figure 9 It runs on the computer device shown.

[0058] Please see Figure 9 , Figure 9 This is a schematic block diagram of a computer device 500 provided in an embodiment of this application. The computer device 500 can be a terminal or a server. The terminal can be an electronic device with communication functions, such as a desktop computer, tablet computer, or smartphone. The server can be a standalone server or a server cluster composed of multiple servers.

[0059] See Figure 9 The computer device 500 includes a processor 502, a memory, and a network interface 505 connected via a system bus 501. The memory may include a non-volatile storage medium 503 and internal memory 504.

[0060] The non-volatile storage medium 503 may store an operating system 5031 and a computer program 5032. The computer program 5032 includes program instructions that, when executed, cause the processor 502 to perform an air conditioner control method.

[0061] The processor 502 provides computing and control capabilities to support the operation of the entire computer device 500.

[0062] The internal memory 504 provides an environment for the operation of the computer program 5032 in the non-volatile storage medium 503. When the computer program 5032 is executed by the processor 502, the processor 502 can execute an air conditioner control method.

[0063] This network interface 505 is used for network communication with other devices. Those skilled in the art will understand that... Figure 9 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device 500 to which the present application is applied. The specific computer device 500 may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0064] The processor 502 is used to run a computer program 5032 stored in a memory to implement the steps of the above method.

[0065] It should be understood that in the embodiments of this application, the processor 502 may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0066] It will be understood by those skilled in the art that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program includes program instructions and can be stored in a storage medium, which is a computer-readable storage medium. The program instructions are executed by at least one processor in the computer system to implement the process steps of the embodiments of the above methods.

[0067] Therefore, the present invention also provides a storage medium. This storage medium can be a computer-readable storage medium. The storage medium stores a computer program, wherein the computer program includes program instructions. When executed by a processor, the program instructions cause the processor to perform the steps of the above-described method.

[0068] The storage medium can be any computer-readable storage medium capable of storing program code, such as a USB flash drive, portable hard drive, read-only memory (ROM), magnetic disk, or optical disk.

[0069] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0070] In the several embodiments provided by this invention, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For example, the division of each unit is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.

[0071] The steps in the method of this invention can be adjusted, merged, or reduced in order according to actual needs. The units in the device of this invention can be merged, divided, or reduced according to actual needs. Furthermore, the functional units in the various embodiments of this invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0072] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a terminal, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention.

[0073] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An air conditioner control method, characterized in that, Applied to an air conditioning control system, the air conditioning control system including millimeter-wave radar and voice interaction device, the method includes: The millimeter-wave radar acquires the target user's respiratory status information in real time. If an abnormal breathing state is detected in the target user's breathing state information, an active interactive prompt associated with the breathing state is generated and broadcast through the voice interaction device; The voice interaction device receives voice feedback signals from the user in response to the voice interaction device's broadcast. The parameters of the speech recognition strategy are dynamically adjusted based on the breathing state information. These adjustments include: when the breathing state information indicates that the target user is experiencing rapid breathing, the confidence threshold of the speech recognition strategy is lowered, and the weight of a first preset type of instruction word is increased. This first preset type of instruction word expresses the user's need for cooling, and the confidence threshold controls the recognition threshold for candidate words. Conversely, when the breathing state information indicates that the target user is experiencing weak breathing, the rejection threshold of the speech recognition strategy is lowered, and the weight of a second preset type of instruction word is increased. This second preset type of instruction word expresses the user's need for improved user comfort, and the rejection threshold determines whether a speech segment is valid. Based on the voice feedback signal, a dynamically adjusted voice recognition strategy is used to parse it into a voice command. The parsing into a voice command includes: performing echo cancellation and background noise suppression on the acquired voice feedback signal to obtain a clear voice signal; extracting Mel-frequency cepstral coefficient features from the clear voice signal and inputting them into a preset deep neural network model for acoustic modeling and decoding to obtain an initial voice command; re-evaluating the initial voice command based on the confidence threshold adjusted according to the breathing state information and the command word weight, and outputting the final voice command. The system generates air conditioning control commands based on the breathing status information and the voice commands, and adjusts the operating status of the air conditioner based on the air conditioning control commands.

2. The air conditioner control method according to claim 1, characterized in that, The step of acquiring the target user's respiratory status information in real time via the millimeter-wave radar includes: The millimeter-wave radar receives reflected signals from the chest or abdomen of the target user to obtain micro-motion timing data. The micro-motion timing data is filtered to remove interference signals caused by the operation of the air conditioner fan or environmental vibration; Based on the filtered micro-motion time series data, the respiratory frequency is extracted by fast Fourier transform, and the respiratory amplitude is obtained by peak detection algorithm; Based on the respiratory rate and respiratory amplitude, the respiratory state information is classified into at least one of the following states according to a preset respiratory state classification rule: steady breathing state, rapid breathing state, deep breathing state, and weak breathing state.

3. The air conditioner control method according to claim 2, characterized in that, The step of generating an active interactive prompt associated with the breathing state and broadcasting it through the voice interaction device if an abnormal breathing state is detected in the target user's breathing state information includes: The abnormality type is determined based on the respiratory status information, and the abnormality type includes at least rapid breathing or weak breathing. When the abnormality type is determined to be the shortness of breath state, a first voice prompt is generated. The first voice prompt is used to ask whether it is necessary to increase the cooling or lower the temperature. When the abnormality type is determined to be the weak breathing state, a second voice prompt is generated. The second voice prompt is used to ask whether it is necessary to enter comfort mode, reduce the fan speed, or increase the temperature. The corresponding voice prompts are played through the voice interaction device, and the feedback voice signals from the target user are monitored within a preset time period.

4. The air conditioner control method according to claim 2, characterized in that, The step of generating air conditioning control commands based on the breathing state information and the voice commands, and then adjusting the operating status of the air conditioner based on the air conditioning control commands, includes: When the breathing state is the rapid breathing state and the voice command includes the first preset type command word, an air conditioning rapid cooling control command is generated. When the breathing state is the deep breathing state and the voice command includes the second preset type command word, an air conditioning comfort adjustment control command is generated.

5. An air conditioner control device, characterized in that, Used to perform the air conditioner control method as described in any one of claims 1 to 4.

6. A computer device, characterized in that, The computer device includes a memory and a processor connected to the memory; the memory is used to store a computer program; the processor is used to run the computer program stored in the memory to perform the steps of the method as described in any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which includes program instructions that, when executed by a processor, can implement the steps of the method as described in any one of claims 1 to 4.

Images