A virtual reality interaction method, device and medium based on respiratory data
By using an individualized breathing baseline model and multi-level signal quality control, the problems of user feature adaptation and signal discrimination in breathing interaction are solved, achieving high-precision, low-false-touch breathing interaction control and improving the robustness and user experience of the virtual reality system.
Patent Information
- Application Number
- CN202611266849.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-20
- Publication Date
- 2026-11-17
AI Technical Summary
Existing breathing interaction technologies cannot dynamically adapt to individual user characteristics and lack effective signal quality identification, resulting in high false trigger rates, unstable continuous control, and insufficient overall interaction robustness.
By employing a guided standardized calibration process and a multi-level signal quality control mechanism, an individualized respiratory baseline model is constructed, a digital respiratory data stream is generated, and a respiratory feature vector is constructed through preprocessing. Combined with virtual reality interactive state recognition of breathing intention, the individualized respiratory baseline and control parameters are dynamically adjusted to achieve high-precision respiratory interactive control.
It effectively overcomes physiological differences among users and initial noise interference, reduces the error rate, and achieves smoothness and stability of continuous control, enabling breathing interaction to move from a conceptually feasible to a practical and reliable system.
Smart Images

Figure CN123114256A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of virtual reality interaction technology, and in particular to a virtual reality interaction method, device and medium based on breathing data. Background Technology
[0002] In recent years, virtual reality (VR) technology has become increasingly popular in fields such as education, healthcare, gaming, and rehabilitation training. People have also begun to explore new ways to operate VR without using their hands. Among these methods, breathing, as a natural and actively controllable bodily signal, is gradually being seen as a promising means of interaction—especially suitable for users with limited hand mobility or those who wish to be fully immersed without interruption. Currently, attempts have been made to use breathing to collect users' breathing signals, such as detecting breath-holding to "confirm" options or using deep breathing to zoom in / rotate virtual scenes.
[0003] However, these existing methods still have many problems in practical use. First, everyone's breathing habits are very different: some breathe fast, some slow; some breathe deeply, some shallowly, but most systems use a fixed set of rules to judge everyone's intentions, resulting in them "failing" when used by a different user. Second, when users first start using the system, they often experience tension or adjust their posture, causing a lot of noise to be mixed into the collected breathing signals. The system lacks an effective mechanism to identify which data is reliable and which should be discarded, causing subsequent judgments to be "off track" from the beginning. More importantly, in scenarios requiring continuous control (such as rotating the viewpoint), breathing itself is a relatively slow physiological process. If the system directly maps every change in breathing to an action, it can easily cause screen jitter, misoperation, or even make the user feel out of control. Existing solutions rarely intelligently adjust sensitivity based on whether the user is "quickly browsing" or "fine-tuning". Therefore, existing breathing interaction technologies still need to solve the problems of not being able to dynamically adapt to individual user characteristics, lacking effective discrimination of signal quality, resulting in high false trigger rates in noisy environments, unstable continuous control, and insufficient overall interaction robustness. Summary of the Invention
[0004] In view of the aforementioned existing problems, the present invention is proposed.
[0005] Therefore, the present invention provides a virtual reality interaction method, device and medium based on respiratory data, which solves the problems of being unable to dynamically adapt to individual user characteristics, lacking effective discrimination of signal quality, resulting in high false trigger rate in noisy environments, unstable continuous control and insufficient overall interaction robustness.
[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:
[0007] In a first aspect, the present invention provides a virtual reality interaction method based on respiratory data, comprising,
[0008] Respiratory physiological signals are collected, and an individualized respiratory benchmark model is constructed through a guided standardized calibration process and a multi-level signal quality control mechanism. A digital respiratory data stream is generated, and a respiratory feature vector is constructed based on the digital respiratory data stream after preprocessing.
[0009] The interaction state identifier in virtual reality is obtained, and the breathing interaction intention is identified by combining the breathing feature vector. The identified breathing interaction intention is then mapped to generate virtual reality interaction control commands.
[0010] The virtual reality interactive control commands are sent, and the corresponding interactive operations are executed in the virtual reality environment. The virtual environment status is updated, the interactive execution results are analyzed and judged, and the individualized respiratory baseline and respiratory interactive control parameters are dynamically adjusted.
[0011] As a preferred embodiment of the virtual reality interaction method based on respiratory data described in this invention, the following steps are taken: the collection of respiratory physiological signals, the construction of an individualized respiratory benchmark model through a guided standardized calibration process and a multi-level signal quality control mechanism, and the generation of digital respiratory data stream indicators. After the virtual reality system is started, the initialization calibration phase is entered, and the user is guided to perform standardized breathing through a graphical user interface.
[0012] In the initialization calibration phase, an analog voltage signal, i.e. a respiratory physiological signal, is collected by an impedance breathing band worn on the user's chest and converted into a digital signal to form the raw respiratory data stream of the calibration phase. The voltage value of each sampling point in the raw respiratory data stream of the calibration phase is converted into an equivalent thoracic impedance value to construct an impedance form respiratory signal sequence of the calibration phase. A sliding window is used to perform deterministic quality screening on the impedance form respiratory signal sequence of the calibration phase. The deterministic quality screening selects the window that meets the stability condition by calculating the low-frequency energy ratio and the first-order difference mean in each window, which is defined as the stable resting segment, and the duration is obtained.
[0013] During the duration of the stable resting period, the impedance baseline statistic is calculated. The impedance form respiratory signal sequence of the calibration phase is standardized using the impedance baseline statistic to obtain a dimensionless normalized sequence. The dimensionless normalized sequence is then subjected to deterministic quality screening based on a sliding window to obtain windows that meet the stability conditions. These windows are then merged to obtain a set of high-quality data segments. Respiratory cycle detection is performed within the time interval covered by the set of high-quality data segments to obtain a complete respiratory cycle sequence. Based on the respiratory cycle sequence, the user's baseline respiratory rate parameter, baseline respiratory depth parameter, and baseline respiratory rhythm stability index parameter are calculated respectively, and an individualized respiratory baseline model for the user is constructed.
[0014] After calibration, the system enters the normal virtual reality interaction phase, continuing to collect the user's respiratory physiological signals through the same impedance breathing band and generating a continuous digital respiratory data stream at the same sampling rate.
[0015] As a preferred embodiment of the virtual reality interaction method based on respiratory data described in this invention, the step of constructing respiratory feature vectors by preprocessing digital respiratory data streams refers to preprocessing the digital respiratory data stream based on the user's individualized respiratory baseline model and the digital respiratory data stream, and segmenting the preprocessed signal using a sliding time window mechanism. For each signal segment within a time window, a complete inspiratory-expiratory cycle is identified using a peak-valley detection method, and the current respiratory feature parameters, including average respiratory rate, average respiratory depth, and average inspiratory-expiratory ratio, are calculated. The current respiratory feature parameters are compared with the user's individualized respiratory baseline model to construct a relative change as a normalized respiratory feature vector. The corresponding normalized respiratory feature vectors are calculated sequentially for all sliding time windows, and finally, a set of normalized respiratory feature vectors is obtained.
[0016] As a preferred embodiment of the virtual reality interaction method based on respiratory data described in this invention, the step of obtaining the interaction state identifier in virtual reality and combining it with respiratory feature vectors to identify respiratory interaction intentions refers to obtaining a set of normalized respiratory feature vectors and then obtaining the current interaction state identifier State of the virtual reality system. The interaction state identifier State is maintained in real time by the virtual reality application logic module and includes "main menu browsing state", "object selection state", "environment adjustment state" or "immersive viewing state". The normalized respiratory feature vector corresponding to the current time window and the current interaction state identifier State are used as inputs, and a state-feature joint judgment mechanism based on a rule engine is used to identify respiratory intentions.
[0017] The rule engine-based state-feature joint determination mechanism pre-configures a set of state-related matching rules. Each rule consists of a combination of the current state and breathing features. That is, for the current time window, the corresponding determination rule is called according to the value of the current interaction state identifier State to determine the normalized breathing feature vector.
[0018] Ultimately, the breathing interaction intent label associated with the current virtual reality interaction state is obtained, and a "no intent" label is output when there is no match.
[0019] As a preferred embodiment of the virtual reality interaction method based on respiratory data described in this invention, the step of mapping the identified respiratory interaction intent to generate virtual reality interaction control commands refers to using the respiratory interaction intent label and the "no intent" label combined with the interaction state identifier State as a joint query key, calling a predefined state-intent-command mapping rule library, and generating the corresponding virtual reality interaction control commands.
[0020] The state-intention-command mapping rule base pre-configures corresponding virtual reality interaction command types for different breathing interaction intentions, converting abstract interaction intentions into control commands that can be executed by the virtual reality system. Each line corresponds to a specific mapping rule. Control commands include instantaneous triggering interaction commands, no-operation commands, and continuous control interaction commands.
[0021] When the interactive command is a momentary trigger, a one-time operation event is generated and submitted to the input event queue of the virtual reality system. When the interactive command is an empty operation, no control command is generated. When the interactive command is a continuous control type, the control readiness determination process is initiated. The control readiness determination process determines whether a complete respiratory cycle exists based on the timestamp of the detected expiratory valley value. Continuous control output is enabled only after the first complete expiratory valley value is detected. The process dynamically switches between high-gain and low-gain control modes according to the normalized respiratory feature vector of the current time window to obtain the target control speed. The target control speed is then smoothed by inertial attenuation coefficient to obtain the actual output control speed. Finally, a continuous control type interactive command containing the actual output control speed is generated.
[0022] As a preferred embodiment of the virtual reality interaction method based on respiratory data described in this invention, the step of sending virtual reality interaction control commands and executing corresponding interaction operations in the virtual reality environment, and updating the virtual environment state, refers to sending standardized VR interaction commands as input to the virtual reality rendering engine, which then calls the corresponding application logic module to execute the operation and synchronously updates the virtual environment state.
[0023] As a preferred embodiment of the virtual reality interaction method based on respiratory data described in this invention, the analysis and judgment of the interaction execution result refers to providing the user with multimodal feedback corresponding to the respiratory input behavior after the interaction control command is executed. The multimodal feedback includes visual feedback and tactile feedback. The feedback is presented synchronously after the command is executed, and it is detected whether the user actively cancels the operation or repeatedly tries the same command after the interaction. This is used as the basis for judging false triggers. If the user executes "cancel", "return" or repeated triggers with the same intention within a preset time after the command is executed, the interaction is marked as invalid; otherwise, it is marked as valid.
[0024] As a preferred embodiment of the virtual reality interaction method based on respiratory data described in this invention, the dynamic adjustment of individualized respiratory reference and respiratory interaction control parameters refers to updating the individualized respiratory reference model and inertial decay coefficient respectively according to the effectiveness of the interaction, and using the updated individualized respiratory reference model and inertial decay coefficient to generate new virtual control commands.
[0025] In a second aspect, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, it implements any step of the virtual reality interaction method based on respiratory data as described in the first aspect of the present invention.
[0026] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the virtual reality interaction method based on respiratory data as described in the first aspect of the present invention.
[0027] The beneficial effects of this invention are as follows: Through individualized modeling and multi-level signal quality control, this invention effectively overcomes the problems of large physiological differences among users and initial noise interference. With the help of state-aware intent recognition and control readiness mechanism, the error rate is greatly reduced. Meanwhile, the gain adaptive switching and inertial smoothing strategy realize the smoothness and stability of continuous control. Overall, without the need for additional hardware, breathing interaction has moved from "conceptually feasible" to "practical and reliable". Attached Figure Description
[0028] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0029] Figure 1 This is a flowchart of a virtual reality interaction method based on respiratory data in Example 1.
[0030] Figure 2 This is a flowchart of the virtual reality respiratory monitoring calibration process in Example 1. Detailed Implementation
[0031] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
[0032] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.
[0033] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.
[0034] Example 1, referring to Figure 1 and Figure 2 This is the first embodiment of the present invention, which provides a virtual reality interaction method based on respiratory data, including the following steps:
[0035] S1. Collect respiratory physiological signals, construct an individualized respiratory baseline model through a guided standardized calibration process and a multi-level signal quality control mechanism, and generate a digital respiratory data stream. Based on the digital respiratory data stream, perform preprocessing to construct a respiratory feature vector.
[0036] Specifically, respiratory physiological signals are collected, and an individualized respiratory baseline model is constructed through a guided standardized calibration process and a multi-level signal quality control mechanism. A digital respiratory data stream is generated. After the virtual reality system is started, the user is guided to perform standardized breathing through a graphical user interface. This includes the user maintaining a sitting or lying position and in a naturally relaxed resting state. First, the user performs a fixed duration of natural breathing, then completes a maximum deep inhalation and a slow, complete exhalation, and then resumes a fixed duration of natural breathing. This helps the user establish a stable resting breathing state and reduces signal interference caused by initial tension or posture adjustment. During the user's breathing process, the system enters the initialization calibration phase.
[0037] During the initial calibration phase, an analog voltage signal, i.e., a respiratory physiological signal, is acquired through an impedance-type breathing band worn on the user's chest (which measures the change in AC signal impedance caused by volume changes in the chest cavity during respiration and outputs an analog voltage signal proportional to respiratory motion). The acquired analog voltage signal is then processed by an analog-to-digital converter (ADC) at a fixed sampling rate. This is converted into a digital signal, forming the raw respiratory data stream for the calibration phase:
[0038]
[0039] in, It is the raw respiratory data stream during the calibration phase. is the voltage value at the i-th sampling point, and N is the total number of sampling points during the calibration phase;
[0040] To convert the acquired analog voltage signal into a physiologically meaningful thoracic impedance value and to eliminate the influence of different user body shapes, electrode attachment pressure, or wire contact differences on the signal amplitude, the equipment parameters obtained from the factory calibration were used to convert the voltage value of each sampling point in the raw respiratory data stream during the calibration phase into an equivalent thoracic impedance value:
[0041]
[0042] in, Here, is the equivalent thoracic impedance value at the i-th sampling point, 'a' is the voltage-to-impedance conversion proportionality coefficient, and 'b' is the offset, reflecting the basic thoracic impedance and circuit zero point. Both 'a' and 'b' are constants obtained through standard resistor calibration before the equipment leaves the factory and are fixed in the system. The typical range of 'a' is... The typical range of values for b is ;
[0043] By processing all sampling points sequentially, an impedance-form respiratory signal sequence is constructed for the calibration phase:
[0044]
[0045] in, It is an impedance-form respiratory signal sequence during the calibration phase;
[0046] To ensure that subsequent parameter calculations are based on reliable data, a sliding window is used for the impedance-form respiratory signal sequence during the calibration phase. Deterministic quality screening is performed by setting a window length L and calculating the proportion of low-frequency energy and the mean of the first-order difference within each window.
[0047] The proportion of low-frequency energy The frequency domain representation is obtained by performing a Discrete Fourier Transform (DFT) on the window signal.
[0048]
[0049] in, It is a frequency domain representation. It is a frequency index, and j is the imaginary unit. It is pi. It is a Hamming window function used to suppress spectral leakage, where u is the frequency domain sampling point index. It is the impedance value of the u-th sampling point in the current window;
[0050] Based on frequency domain representation Calculate the energy percentage of the frequency band:
[0051]
[0052] in, It is the maximum DFT index corresponding to the lower limit of the effective respiratory frequency band. It is the minimum DFT index corresponding to the lower limit of the effective respiratory frequency band;
[0053] and The calculation formula is:
[0054]
[0055] in, This is the effective respiratory frequency band, set according to physiological consensus and standards such as IEEE and ISO, with a typical value of [0.1, 0.5] Hz. It is a fixed sampling rate;
[0056] The formula for calculating the first-order difference mean (i.e., the average value of the changes between adjacent points) is as follows:
[0057]
[0058] in, It is the mean of the absolute values of the first-order differences within the w-th sliding window, where L is the window length. It is the impedance value of the (u+1)th sampling point in the current window;
[0059] To eliminate absolute impedance drift caused by hydration status, diet, or fat layer thickness, the impedance-form respiratory signal sequence was obtained during the calibration phase. A window that meets the following conditions is selected and defined as a stable resting segment, with a duration of [duration value missing]. The conditions are:
[0060]
[0061] in, This is the threshold for the proportion of low-frequency energy, set through physiological signal experiments, with a typical value range of [0.5, 0.7]. It is the differential intensity threshold, set through actual measurement and statistical analysis, with a typical value range of [0.7, 1.0].
[0062] Duration of the stable resting period Internal calculated impedance baseline statistics:
[0063]
[0064] in, and It is the impedance baseline statistic. It is the number of sampling points used to calculate the "stable segment" of the statistic. It is the starting index of a stable segment;
[0065] Through impedance baseline statistics and Impedance-form respiratory signal sequences during the calibration phase Performing Z-score standardization yields a dimensionless normalized sequence:
[0066]
[0067] in, It is a dimensionless normalized sequence. It is the standardized equivalent thoracic impedance value of the i-th sampling point;
[0068] For dimensionless normalized sequences Perform the sliding window-based quality filtering process again, retaining only windows that meet the following conditions, and merging them to form a high-quality data segment set. The conditions are:
[0069]
[0070] In high-quality data segment collection Peak and trough detection is performed within the covered time interval. Standard local extremum detection algorithms (such as finding the derivative sign change point or sliding window comparison) are used to identify all local maxima (inspiratory peaks) and local minima (expiratory troughs), and minimum cycle interval constraints are applied. ;
[0071] The minimum period interval constraint The time interval between any two adjacent valleys (or peaks) must not be less than the minimum period interval constraint. Only when a peak is immediately followed by a trough (or vice versa), and the interval is greater than or equal to the minimum period interval constraint. Only when this time is reached is it considered a complete respiratory cycle;
[0072] Suppose a total of n complete respiratory cycles are detected, and the starting feature point index (based on the trough of expiration) for each cycle is recorded as follows: Calculate the duration of the k-th respiratory cycle (k=1,2,...,n):
[0073]
[0074] in, It is the duration of the kth complete respiratory cycle. It is the sampling rate. It is the index of the sampling point in the original respiratory data stream corresponding to the feature point of the (k+1)th complete respiratory cycle. It is the sampling point index of the feature point corresponding to the kth complete respiratory cycle in the original respiratory data stream;
[0075] From dimensionless normalized sequences Obtain the peak inspiratory voltage of the kth complete respiratory cycle. Expiratory trough voltage Calculation period relative amplitude:
[0076]
[0077] in, It is the relative amplitude of the cycle in the kth complete respiratory cycle;
[0078] The baseline respiratory rate parameter is calculated by using the duration of a complete respiratory cycle and the relative amplitude of the cycle. Basic respiratory depth parameters and baseline respiratory rhythm stability parameters Build an individualized respiratory baseline model for users. ;
[0079] The basic respiratory rate parameter Based on the duration of a complete respiratory cycle, calculate the baseline respiratory rate parameter:
[0080]
[0081] in, It is the average respiratory cycle duration;
[0082] The baseline respiratory depth parameter (Reflecting the extent of chest cavity expansion, which is positively correlated with tidal volume in impedance respiratory monitoring) is the average of the difference between the peak inspiratory volume and the trough expiratory volume in each respiratory cycle:
[0083]
[0084] in, It is the relative amplitude of the cycle in the kth complete respiratory cycle;
[0085] The basic respiratory rhythm stability index parameters Rhythmic stability is characterized by the standard deviation of the duration of adjacent respiratory cycles, and rhythmic stability is expressed by the coefficient of variation:
[0086]
[0087] in, It is the standard deviation of the mean respiratory cycle duration;
[0088] After calibration, the system enters the normal virtual reality interaction phase, continuing to collect the user's respiratory physiological signals through the same impedance breathing band and generating a continuous digital respiratory data stream R(t) at the same sampling rate for subsequent real-time processing.
[0089] Through standardized guidance, impedance calibration, two-stage quality screening, Z-score normalization, and constrained periodicity detection, a highly robust individualized respiratory benchmark model was constructed. The model effectively eliminated the influence of physiological differences among users, electrode contact variations, and initial state interference on the signal. It accurately extracted frequency, depth, and rhythm stability parameters that reflect the true respiratory rhythm, significantly improving the accuracy and stability of subsequent respiratory intention recognition. It avoided false triggering or response delays caused by signal noise or individual drift. Ultimately, reliable respiratory interaction across users and scenarios was achieved without additional training, providing a low-burden, highly immersive natural interaction channel for virtual reality systems.
[0090] Furthermore, preprocessing of digital respiratory data streams to construct respiratory feature vectors refers to a user-based individualized respiratory baseline model. The digital respiratory data stream R(t) is preprocessed to eliminate interference. Specifically, a fourth-order Butterworth bandpass filter is used to filter the digital respiratory data stream R(t), preserving the normal respiratory frequency range (e.g., the adult respiratory rate is typically 6–30 breaths / minute, corresponding to 0.1–0.5 Hz), suppressing high-frequency electromyographic noise and low-frequency baseline drift. A moving window midpoint filter is used to further suppress sudden motion artifacts. The filtered signal is then zero-mean processed, i.e., the local mean of the signal within the current time window is subtracted to eliminate residual DC offset. A sliding time window mechanism is used to segment the preprocessed signal. For each signal segment within a time window, a peak-valley detection method is used to identify the complete inspiratory-expiratory cycle. Assuming that a signal is detected within the window... A complete cycle (i.e., a complete cycle of the signal segment within the l-th time window, if) If the result is not obtained, skip this window or use the result from the previous window. Calculate the current respiratory characteristic parameters, including average respiratory rate, average respiratory depth, and average inspiratory-expiratory ratio.
[0091] The formula for calculating the average respiratory rate is:
[0092]
[0093] in, It is the average respiratory rate. It is the length of the time window. It is a time unit conversion factor used to convert the ratio of the number of cycles to time into the target frequency unit;
[0094] The average respiratory depth is calculated by taking the average of the peak-to-trough difference for each detected cycle.
[0095]
[0096] in, It is the average respiratory depth;
[0097] The average inspiratory-to-expiratory ratio is defined as the ratio of the average inspiratory duration to the average expiratory duration:
[0098]
[0099] in, It is the average inspiratory-to-expiratory ratio. and These are the duration of inhalation and the duration of exhalation;
[0100] Compare current respiratory characteristic parameters with the user's individualized respiratory baseline model. By comparison, relative changes are constructed as normalized features:
[0101]
[0102] in, It is the relative change in average respiratory rate. It is the relative change in average respiratory depth;
[0103] The absolute value of the mean inspiratory-to-expiratory ratio (IPR) is directly used as the normalized characteristic of the mean IPR. The normalized features are combined into a normalized respiratory feature vector for the l-th time window. The normalized respiratory feature vectors are calculated sequentially for all sliding time windows, resulting in a final set of normalized respiratory feature vectors. ,in, It is the normalized respiratory feature vector of the l-th time window.
[0104] By employing bandpass filtering, median smoothing, zero-mean normalization, and sliding window periodic detection, noise is effectively suppressed and breathing signals are stabilized. Then, based on an individualized benchmark model, relative change features of frequency and depth are constructed, and the absolute value of the inspiratory-to-expiratory ratio is fused to form a normalized breathing feature vector sequence that is resistant to drift and has high discriminative power. This significantly improves the robustness of the features to changes in user state and environmental interference, providing a reliable foundation for high-precision, low-latency breathing intention recognition. It is suitable for virtual reality natural interaction in long-term, dynamic scenarios.
[0105] S2. Obtain the interaction state identifier in virtual reality, combine it with the breathing feature vector to identify the breathing interaction intention, and map the identified breathing interaction intention to generate virtual reality interaction control commands.
[0106] Specifically, the process involves acquiring interaction state identifiers in virtual reality, combining them with breathing feature vectors to perform breathing interaction intent recognition, and obtaining a normalized set of breathing feature vectors. Then, the current virtual reality system's interaction state identifier, State, is obtained. This State is maintained in real-time by the virtual reality application logic module and represents the current operational stage of the system. It is one of the predefined discrete states, including "main menu browsing state," "object selection state," "environment adjustment state," or "immersive viewing state (non-interactive state)." The normalized breathing feature vector corresponding to the current time window and the current interaction state identifier, State, are used as inputs. A rule-engine-based state-feature joint judgment mechanism is employed to identify breathing intentions. The mechanism pre-configures a set of state-related matching rules. Each rule consists of a combination of the current state and breathing features. Specifically, for the current time window, based on the value of the current interaction state identifier, State, the corresponding judgment rule is invoked to judge the normalized breathing feature vector. The judgment rules include:
[0107] When the interaction state identifier State is "object selection state", if the user's breath-holding behavior is detected, it is determined as a "confirm" intent, and a breathing interaction intent label "Confirm" is generated. The conditions for recognizing the breath-holding behavior are:
[0108]
[0109] in, This is the relative decrease threshold for breath-holding determination, set based on historical sample statistics, with a typical value range of [-0.8, -0.5]. It is the maximum permissible respiratory rate for breath-holding determination. It was set through experiments, and the typical value range is [1,3].
[0110] When the interaction state is "Main Menu Browsing State", if rapid, shallow breathing is detected, it is determined to be a "Return to Previous Level" intent, and the intent label "Back" is generated. The determination condition is:
[0111]
[0112] in, This is the relative increase threshold for rapid breathing frequency, set through statistical analysis, with a typical value range of [0.4, 1.0]. It is the relative decrease threshold of the depth of rapid shallow breathing, which is set through an adaptive mechanism, with a typical value range of [-0.5, -0.2].
[0113] When the interaction state identifier is "Environmental adjustment state", if continuous deep breathing is detected, it is determined to be an intent to "confirm current settings", and the intent label "Apply" is generated. The determination condition is:
[0114]
[0115] in, This is the relative rise threshold of deep breathing depth, set using an adaptive percentile method, with a typical value range of [0.5, 1.2]. It is the maximum permissible coefficient of variation for rhythm stability, set by the ROC optimization method, with a typical value range of [0.1, 0.25].
[0116] When the interaction state identifier State is "Immersive Viewing State (Non-Interactive State)," no interaction commands are generated, and it is only used for visual prompts of breathing state. If the current normalized breathing feature vector does not match the preset conditions of the current state (such as deep breathing instead of holding one's breath in "Object Selection State"), no interaction intent label is generated to avoid accidental operation. Finally, the breathing interaction intent label (Confirm Intent, Back Intent, Apply / Effective Intent) associated with the current virtual reality interaction state is obtained, and the "No Intent" label is output when there is no match.
[0117] By using a state-feature joint determination mechanism, normalized breathing features are dynamically bound to VR interaction states to achieve context-aware intent recognition. Dedicated determination rules are enabled in different states (such as breath-holding confirmation, shallow quick return, and deep stable application) to effectively distinguish semantic differences between similar breathing patterns. Output is actively suppressed in non-interactive states, and "no intent" is returned for mismatched cases. This constructs a highly robust breathing interaction logic with low false touches, which significantly improves the system's intent recognition accuracy and user experience security in multiple scenarios.
[0118] Furthermore, the identified breathing interaction intentions are mapped to generate virtual reality interaction control commands. The breathing interaction intention labels (Confirm, Back, Apply) and the "No Intention" label (outputting when no match) are used as the main input. The current interaction state identifier (State) synchronously obtained from the virtual reality application's runtime environment is used as the joint query key. A predefined state-intention-command mapping rule library is invoked to generate the corresponding virtual reality interaction control commands. This state-intention-command mapping rule library pre-configures corresponding virtual reality interaction command types for different breathing interaction intentions, used to convert abstract interaction intentions into control commands executable by the virtual reality system. Each line corresponds to a specific mapping rule, in the following format:
[0119]
[0120] In this context, Intent is the breathing interaction intent label, and CommandType is a predefined control instruction type identifier, selected from instantaneous control type, continuous control type, or no-operation type (such as confirming the current UI option (UI_Confirm), returning to the previous menu (Menu_Back), applying the current settings (Setting_Apply), initiating left / right / up / down rotation of the view modulated by breathing depth (View_Rotate), initiating scene zooming / scaling modulated by breathing depth (Scene_Scale), or no valid interaction intent (None)).
[0121] When the corresponding virtual reality interaction control command type CommandType is “UI_Confirm”, “Menu_Back”, or “Setting_Apply”, an instantaneous trigger interaction command is generated. The instantaneous trigger interaction command is used to trigger a one-time operation event, including user interface confirmation, menu return, or setting activation. After the command is generated, it is directly submitted to the input event queue of the virtual reality system, and the VR application logic module executes the corresponding operation in the next frame processing cycle (such as closing the pop-up window, returning to the previous menu, or application environment settings).
[0122] When the corresponding virtual reality interaction control command type CommandType is “None”, no virtual reality interaction control command is generated, or an empty operation command is output to maintain the current virtual reality environment state and thus avoid accidental triggering of interaction operations.
[0123] When the corresponding virtual reality interactive control command type CommandType is a continuous control type (such as "View_Rotate" or "Scene_Scale", that is, the command type returned by the mapping rule base is a continuous control operation, such as "view rotation" or "scene scaling"), the control signal is not generated immediately, but the control readiness judgment process is started to solve the physiological limitation that breathing signals cannot be started instantaneously.
[0124] The control readiness determination process utilizes the detected respiratory cycle start and end point information (i.e., expiratory trough timestamp) to check whether at least one complete expiratory trough value (marking the closure of a respiratory cycle) has been detected since the continuous control intention was triggered. If not, the control request is temporarily stored, the output rate is kept at zero, and respiratory cycle events are continuously monitored. Only when the first complete expiratory trough value is detected is the continuous control output officially activated, and this moment is recorded as the control start point. After control is ready (i.e.) From the set of normalized respiratory feature vectors In the process, obtain the feature vector corresponding to the current time window l. To achieve efficient and stable control, adaptive gain switching is implemented, and a first-order difference method is used to calculate the rate of change of the current respiratory depth.
[0125]
[0126] in, It is the rate of change of the current respiratory depth. It represents the relative change in the average breathing depth within the current window. It represents the relative change in average respiratory depth over the previous window. It represents the length of the time window, and |·| is the absolute value operation to ensure that the rate of change is non-negative.
[0127] Calculate the coefficient of variation of respiratory cycle duration within the current window based on the duration of a complete respiratory cycle:
[0128]
[0129] in, It is the coefficient of variation of the respiratory cycle. It is the average respiratory cycle duration within the current window. It is the standard deviation;
[0130] , The calculation formula is:
[0131]
[0132] in, It is the complete period of the signal segment within the l-th time window. It is the duration of the k-th complete respiratory cycle within the l-th time window;
[0133] The control phase is determined by a dual-condition decision; if the rate of change of the current respiratory depth is simultaneously satisfied... Greater than the preset threshold (The constants set during system initialization are determined through statistical analysis of measured user breathing control behavior, with typical values being...) ), and coefficient of variation Less than If the user is in the active exploration phase and requires a rapid response, the control gain is set to a high gain value. Otherwise, the user is determined to be in the fine-tuning stage and requires precise operation; therefore, the control gain is set to a low gain value. ,in, , These are high gain and low gain values, set using user usability testing and subjective evaluation tuning methods, with typical value ranges of [0.7, 1.2] and [0.1, 0.4], respectively. , This is the threshold for the coefficient of variation of the respiratory cycle. Through statistical analysis of measured respiratory physiological data, a typical value range is [0.2, 0.3]. In continuous control scenarios, it uses the relative change of the current average respiratory depth. As a normalized control source for continuous control, to improve the smoothness and immersion of continuous interaction, a virtual inertial mapping mechanism is used for continuous control commands to represent the relative change in the current average breathing depth. The target is to control the speed, that is:
[0134]
[0135] in, It is to control the speed of the target. It is the velocity gain coefficient;
[0136] To avoid jitter in the virtual view or object movement caused by instantaneous fluctuations in the breathing signal, a first-order low-pass filter is used to control the target speed. Perform inertial smoothing to obtain the actual output control speed:
[0137]
[0138] in, This is the actual output control speed of the l-th time window. It is the actual output control speed of the previous time window. It is the inertia decay coefficient, which is set based on the subjective evaluation and optimization method of user usability testing, and the typical value range is [0.1, 0.6].
[0139] The actual output control speed of each window is encapsulated into standardized VR interaction commands. These standardized VR interaction commands include instantaneous trigger-type interaction commands (such as "Confirm" and "Back") and continuous control-type interaction commands (such as view rotation and zoom, including actual output control speed). ).
[0140] By using a state-intention-command mapping rule base, precise semantic binding of breathing intentions is achieved in different VR interaction scenarios, avoiding cross-contextual misoperations. For continuous control, a readiness determination mechanism based on the expiratory valley value is used to solve the problem of premature triggering caused by respiratory physiological delay. Furthermore, by using a dual-condition criterion of the rate of change of breathing depth and the coefficient of variation of rhythm, high / low control gain is dynamically switched to balance the response speed in the exploration phase and the operational accuracy in the fine-tuning phase. Combined with a first-order low-pass filter of the inertial attenuation coefficient, control jitter is effectively suppressed, and motion smoothness and immersion are improved. At the same time, no operation is output for unmatched intentions to build a safety mechanism to prevent accidental touches. Overall, the accuracy, robustness and user experience of breathing interaction in complex VR applications are significantly improved.
[0141] S3. Send virtual reality interactive control commands, execute corresponding interactive operations in the virtual reality environment, update the virtual environment status, analyze and judge the interactive execution results, and dynamically adjust individualized respiratory benchmarks and respiratory interactive control parameters.
[0142] Specifically, sending virtual reality interaction control commands and executing corresponding interactive operations in the virtual reality environment, and updating the virtual environment state means sending standardized VR interaction commands (including instantaneous triggering interaction commands and continuous control interaction commands) as input to the virtual reality rendering engine (such as Unity XR or Unreal Engine), which then calls the corresponding application logic module to execute the operation and synchronously update the virtual environment state.
[0143] When a momentary trigger-type interactive command is received, the virtual reality rendering engine immediately triggers the corresponding user interface event or scene logic operation, including button click, menu switching or setting confirmation, and updates the current state parameters of the virtual scene after execution;
[0144] When receiving continuous control-type interactive commands, the virtual reality rendering engine reads the actual output control speed carried in the commands. This is then converted into control variables to drive the motion of the virtual object or camera, and the pose parameters of the virtual object or camera are updated based on discrete-time integration.
[0145]
[0146]
[0147] in, It is the converted control quantity used to drive the motion of virtual objects or cameras. These are preset conversion factors used to convert the actual output control speed into physical or logical control quantities used within the virtual reality system. They are set based on the virtual reality engine coordinate system and unit unification method, with a typical value range of [0.5, 2.0]. It is the camera yaw angle of the current frame. It's the camera yaw angle from the previous frame. It is the frame interval time.
[0148] By integrating standardized VR interaction commands into the rendering engine, precise linkage between breathing control and the virtual environment is achieved. Instantaneous commands trigger immediate UI responses, while continuous commands drive camera or object movement through velocity-attitude integral conversion. Combined with conversion coefficients and time normalization of frame intervals, cross-platform control consistency and motion smoothness are ensured, effectively bridging the semantic gap between physiological signals and virtual operations, and significantly improving the immersiveness, response accuracy, and system robustness of hands-free interaction.
[0149] Furthermore, analyzing and judging the results of interactive execution refers to providing users with multimodal feedback corresponding to breathing input behavior after the interactive control command is executed, so as to enhance users' perception of the causal relationship between "breathing behavior - virtual environment response". Specific feedback mechanisms include visual feedback and tactile feedback.
[0150] The visual feedback involves overlaying a breathing indicator element into the user's field of vision. This breathing indicator element is a semi-transparent ring, and its display parameters are dynamically adjusted according to changes in the user's breathing characteristics. The radius of the indicator ring is calculated using the following formula:
[0151]
[0152] in, It indicates the radius of the indicator ring. It is the preset normalization range boundary, if Then take ,like Then take Based on the statistical distribution during the calibration phase, The typical value range is [-0.3, -0.6]. The typical value range is [+0.3, +0.6]. These are the minimum / maximum radii of the indicator ring, calibrated based on visual comfort experiments. The typical value range is [10, 40]px. The typical value range is [60, 200]px;
[0153] The tactile feedback is provided to the user via a head-mounted display device or other tactile output device when an instantaneously triggered interactive command is executed. The vibration intensity is related to the amplitude of the change in breathing depth, and the calculation formula is as follows:
[0154]
[0155] Where I is the vibration intensity, This is the maximum vibration intensity, determined by the driving capability of the tactile output device or system configuration parameters, with a typical value range of [0,1]. It is the degree of breathing effort the user exerts when triggering the action;
[0156] Feedback is presented synchronously after the command is executed, ensuring that the user can clearly perceive the causal relationship between the breathing behavior and the virtual response. It also detects whether the user actively cancels the operation or repeatedly tries the same command after the interaction, which serves as the basis for judging false triggers. If the user executes "cancel", "return" or repeats the same intention within a preset time after the command is executed (based on the user behavior analysis method of usability testing, with a typical value range of [1.5, 3.0] seconds), the interaction is marked as invalid (i.e., false trigger). Otherwise, it is marked as valid interaction.
[0157] By using a dynamic visual indicator ring and tactile vibrations related to breathing depth, multimodal real-time feedback is constructed, enabling users to clearly perceive the causal relationship between breathing behavior and virtual response. At the same time, based on the undo or repeat operation behavior within a preset time window after command execution, false triggering events are automatically identified. This not only enhances the user's understanding and confidence in the operation of the hands-free interaction, but also provides the system with an objective basis for evaluating the interaction quality, significantly improving the learning efficiency, usability, and long-term robustness of breathing interaction.
[0158] Furthermore, dynamically adjusting individualized respiratory baselines and respiratory interaction control parameters refers to adjusting the individualized respiratory baseline model based on the effectiveness of the interaction. and inertial decay coefficient Update them separately;
[0159] The individualized respiratory baseline model An update occurs only when a valid interactive event occurs more than a preset number of times (selected through internal testing (different users, different scenarios) to statistically determine the minimum stable number of false triggers and normal triggers, with a typical range of [2,5] times) and the user is in a resting state (determined by State = "immersive viewing state (non-interactive state)" or stable breathing rhythm). ,in, This is a stability threshold, set by the statistical quantiles during the user calibration phase (typically ranging from [0.10, 0.30]). Only when this threshold is reached is the individualized respiratory baseline model updated to avoid erroneous drift during dynamic interactions. The update uses a first-order exponential smoothing formula.
[0160]
[0161]
[0162] in, and These are the updated baseline respiratory rate and baseline respiratory depth parameters. and These are the mean respiratory rate and depth within the effective resting window, respectively. It is the baseline update weight, which is adaptively adjusted according to the magnitude of changes in user posture / state, with a typical value range of [0.02, 0.20].
[0163] During the interaction, when the user is detected to have entered a new stable breathing state, the coefficient of variation is recalculated based on the duration of the respiratory cycle in the new resting segment, and the baseline respiratory rhythm stability index parameters are updated. The coefficient of variation for the duration of the new resting respiratory cycle is updated synchronously. Finally, an updated individualized respiratory baseline model was obtained. ;
[0164] The inertial attenuation coefficient The update mechanism is triggered when more than a preset number of invalid interactions (false triggers) are detected consecutively (determined through user usability testing and interaction stability analysis, typically ranging from [2,5] times), and the user intent is a continuous control intent. In this case, the inertia is automatically increased (i.e., the response speed is reduced) to decrease sensitivity. The update rules are as follows:
[0165]
[0166] in, This is the updated inertial decay coefficient. This is the minimum allowable value to prevent complete unresponsiveness. It is set through user usability testing and subjective evaluation tuning, with a typical value range of [0.1, 0.3]. The step size is fixed and set experimentally, with a typical range of [0.05, 0.1].
[0167] If a series of valid interactions exceeding a preset number are detected and the user does not exhibit hesitation (e.g., no repeated operations), then the inertia is slightly reduced (response is improved).
[0168]
[0169] in, This is the maximum allowable value, retaining necessary smoothness. It is set through statistical analysis of respiratory signals and user experience testing, with a typical value range of [0.5, 0.8].
[0170] The updated individualized respiratory baseline model and inertial decay coefficient Used to generate new virtual control commands.
[0171] Through a dual-parameter adaptive mechanism driven by interaction effectiveness, the individualized respiratory baseline model is updated only when the user is at rest and the operation is stable, preventing model drift caused by dynamic interference. At the same time, the inertial decay coefficient is dynamically adjusted according to the frequency of continuous false triggers or effective interactions, achieving an online balance between sensitivity and robustness. This significantly improves the adaptability and reliability of the system in long-term, multi-scenario use, and effectively solves the performance degradation problem of respiratory interaction caused by physiological drift and environmental changes.
[0172] This embodiment also provides a computer device applicable to a virtual reality interaction method based on respiratory data, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the virtual reality interaction method based on respiratory data as proposed in the above embodiment.
[0173] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.
[0174] This embodiment also provides a storage medium on which a computer program is stored. When executed by a processor, the program implements a virtual reality interaction method based on respiratory data as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.
Claims
1. A virtual reality interaction method based on respiratory data, characterized in that: include, Respiratory physiological signals are collected, and an individualized respiratory benchmark model is constructed through a guided standardized calibration process and a multi-level signal quality control mechanism. A digital respiratory data stream is generated, and a respiratory feature vector is constructed based on the digital respiratory data stream after preprocessing. The interaction state identifier in virtual reality is obtained, and the breathing interaction intention is recognized by combining the breathing feature vector. The recognized breathing interaction intention is then mapped to generate virtual reality interaction control commands. The virtual reality interactive control commands are sent, and the corresponding interactive operations are executed in the virtual reality environment. The virtual environment status is updated, the interactive execution results are analyzed and judged, and the individualized respiratory baseline and respiratory interactive control parameters are dynamically adjusted.
2. The virtual reality interaction method based on respiratory data as described in claim 1, characterized in that: The collected respiratory physiological signals are used to construct an individualized respiratory benchmark model through a guided standardized calibration process and a multi-level signal quality control mechanism, and a digital respiratory data stream is generated. After the virtual reality system is started, it enters the initialization calibration phase and guides the user to perform standardized breathing through a graphical user interface. In the initialization calibration phase, an analog voltage signal, i.e. a respiratory physiological signal, is collected by an impedance breathing band worn on the user's chest and converted into a digital signal to form the raw respiratory data stream of the calibration phase. The voltage value of each sampling point in the raw respiratory data stream of the calibration phase is converted into an equivalent thoracic impedance value to construct an impedance form respiratory signal sequence of the calibration phase. A sliding window is used to perform deterministic quality screening on the impedance form respiratory signal sequence of the calibration phase. The deterministic quality screening selects the window that meets the stability condition by calculating the low-frequency energy ratio and the first-order difference mean in each window, which is defined as the stable resting segment, and the duration is obtained. During the duration of the stable resting period, the impedance baseline statistic is calculated. The impedance form respiratory signal sequence of the calibration phase is standardized using the impedance baseline statistic to obtain a dimensionless normalized sequence. The dimensionless normalized sequence is then subjected to deterministic quality screening based on a sliding window to obtain windows that meet the stability conditions. These windows are then merged to obtain a set of high-quality data segments. Respiratory cycle detection is performed within the time interval covered by the set of high-quality data segments to obtain a complete respiratory cycle sequence. Based on the respiratory cycle sequence, the user's baseline respiratory rate parameter, baseline respiratory depth parameter, and baseline respiratory rhythm stability index parameter are calculated respectively, and an individualized respiratory baseline model for the user is constructed. After calibration, the system enters the normal virtual reality interaction phase, continuing to collect the user's respiratory physiological signals through the same impedance breathing band and generating a continuous digital respiratory data stream at the same sampling rate.
3. The virtual reality interaction method based on respiratory data as described in claim 2, characterized in that: The preprocessing of the digital respiratory data stream to construct respiratory feature vectors refers to preprocessing the digital respiratory data stream based on the user's individualized respiratory baseline model and the digital respiratory data stream. A sliding time window mechanism is used to segment the preprocessed signal. For each signal segment within a time window, a peak-valley detection method is used to identify the complete inspiratory-expiratory cycle, and the current respiratory feature parameters, including average respiratory rate, average respiratory depth, and average inspiratory-expiratory ratio, are calculated. These current respiratory feature parameters are compared with the user's individualized respiratory baseline model to construct a relative change as a normalized respiratory feature vector. The corresponding normalized respiratory feature vectors are calculated sequentially for all sliding time windows, ultimately resulting in a set of normalized respiratory feature vectors.
4. The virtual reality interaction method based on respiratory data as described in claim 3, characterized in that: The process of obtaining the interaction state identifier in virtual reality and combining it with the breathing feature vector to identify the breathing interaction intention refers to obtaining a set of normalized breathing feature vectors and then obtaining the current interaction state identifier (State) of the virtual reality system. The interaction state identifier (State) is maintained in real time by the virtual reality application logic module and includes "main menu browsing state", "object selection state", "environment adjustment state" or "immersive viewing state". The normalized breathing feature vector corresponding to the current time window and the current interaction state identifier (State) are used as inputs, and a state-feature joint judgment mechanism based on a rule engine is used to identify the breathing intention. The rule engine-based state-feature joint determination mechanism pre-configures a set of state-related matching rules. Each rule consists of a combination of the current state and breathing features. That is, for the current time window, the corresponding determination rule is called according to the value of the current interaction state identifier State to determine the normalized breathing feature vector. Ultimately, the breathing interaction intent label associated with the current virtual reality interaction state is obtained, and the "no intent" label is output when there is no match.
5. The virtual reality interaction method based on respiratory data as described in claim 4, characterized in that: The step of mapping the identified breathing interaction intent to generate virtual reality interaction control commands refers to using the breathing interaction intent label and the "no intent" label combined with the interaction state identifier State as a joint query key, calling the predefined state-intent-command mapping rule base, and generating the corresponding virtual reality interaction control commands. The state-intention-command mapping rule base pre-configures corresponding virtual reality interaction command types for different breathing interaction intentions, converting abstract interaction intentions into control commands that can be executed by the virtual reality system. Each line corresponds to a specific mapping rule. Control commands include instantaneous triggering interaction commands, no-operation commands, and continuous control interaction commands. When the interactive command is a momentary trigger, a one-time operation event is generated and submitted to the input event queue of the virtual reality system. When the interactive command is an empty operation, no control command is generated. When the interactive command is a continuous control type, the control readiness determination process is initiated. The control readiness determination process determines whether a complete respiratory cycle exists based on the timestamp of the detected expiratory valley value. Continuous control output is enabled only after the first complete expiratory valley value is detected. The process dynamically switches between high-gain and low-gain control modes according to the normalized respiratory feature vector of the current time window to obtain the target control speed. The target control speed is then smoothed by inertial attenuation coefficient to obtain the actual output control speed. Finally, a continuous control type interactive command containing the actual output control speed is generated.
6. The virtual reality interaction method based on respiratory data as described in claim 5, characterized in that: The process of sending virtual reality interactive control commands, executing corresponding interactive operations in the virtual reality environment, and updating the virtual environment state refers to sending standardized VR interactive commands as input to the virtual reality rendering engine, which then calls the corresponding application logic module to execute the operation and synchronously update the virtual environment state.
7. The virtual reality interaction method based on respiratory data as described in claim 6, characterized in that: The analysis and judgment of the interaction execution result refers to providing the user with multimodal feedback corresponding to the breathing input behavior after the interaction control command is executed. The multimodal feedback includes visual feedback and tactile feedback. The feedback is presented synchronously after the command is executed, and it is detected whether the user actively cancels the operation or tries the same command repeatedly after the interaction. This is used as the basis for judging false triggers. If the user executes "cancel", "return" or repeated triggers with the same intention within a preset time after the command is executed, the interaction is marked as invalid; otherwise, it is marked as valid.
8. The virtual reality interaction method based on respiratory data as described in claim 7, characterized in that: The dynamic adjustment of individualized respiratory baseline and respiratory interaction control parameters refers to updating the individualized respiratory baseline model and inertial decay coefficient respectively according to the effectiveness of the interaction, and using the updated individualized respiratory baseline model and inertial decay coefficient to generate new virtual control commands.
9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the virtual reality interaction method based on respiratory data as described in any one of claims 1 to 8.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the virtual reality interaction method based on respiratory data as described in any one of claims 1 to 8.
