A bionic robot breathing control method, system, device and robot

Abstract

This invention discloses a biomimetic robot breathing control method, system, device, and robot. The biomimetic robot breathing control method includes: acquiring external multimodal sensing data, internal state data, and voice data to be output; generating first breathing control parameters based on the multimodal sensing data and internal state data; generating basic breathing parameters after voice coordination based on the voice data to be output; generating second breathing control parameters through an emotion decision path based on the multimodal sensing data; fusing the basic breathing parameters after voice coordination with the second breathing control parameters to generate a breathing control command for the robot; and driving the breathing actuator to operate according to the breathing control command. The biomimetic robot breathing control device includes a deformation generating device, an airflow device, at least one biomimetic air outlet, and a gas passage. This invention has advantages such as improving the realism of robot breathing simulation.

Description

Technical Field

[0001] This application relates to the field of biomimetic robot technology, and more specifically, to a biomimetic robot breathing control method, system, device, and robot. Background Technology

[0002] With the rapid development of bionic robot technology, its applications in service, companionship, and medical rehabilitation are becoming increasingly widespread. To enhance the naturalness and friendliness of human-computer interaction, researchers have made significant progress in anthropomorphizing robot appearances, simulating facial expressions, and enhancing voice interaction. However, a crucial but long-neglected aspect is the significant gap in robots' ability to simulate basic human physiological behaviors. Breathing, a critical physiological phenomenon that accompanies life and is closely related to emotional states and language activities, has been almost entirely unsimulated and unintegrated in existing bionic robots.

[0003] Currently, the robotics field lacks a complete solution capable of intelligently simulating human breathing. Existing attempts are limited to simply using motors to drive the chest cavity in a rhythmic up-and-down motion, which has fundamental limitations: First, existing solutions are functionally fixed and cannot dynamically adjust breathing patterns based on the robot's motion and environmental changes, making it difficult to simulate the respiratory changes induced by movement and environmental factors in real physiological states. Second, existing technology is disconnected from emotional computing, failing to convey emotional states through changes in parameters such as breathing frequency and depth, thus losing the important function of breathing as a non-verbal emotional expression channel. Third, existing solutions operate in isolation from voice output, failing to achieve a natural breathing rhythm during dialogue, resulting in a lack of realism in voice interaction. Finally, existing hardware designs have low fidelity, making it difficult to reproduce the complex chest cavity dynamics and airflow characteristics of human breathing, including details such as different airflow velocities during inhalation and exhalation, and changes in nasal airflow temperature.

[0004] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0005] The purpose of this application is to provide a biomimetic robot breathing control method, system, device and robot, which has the advantages of being able to dynamically adjust the breathing mode according to the robot's motion state and environmental changes, achieving natural coordination of emotional expression and voice interaction, and improving the realism of breathing simulation.

[0006] This application provides a biomimetic robot breathing control method, the technical solution of which is as follows: A biomimetic robot breathing control method includes the following steps: S1: Acquiring multimodal sensing data, internal state data, and voice data to be output from the robot's external environment; the multimodal sensing data includes environmental data from the robot's external environment; the internal state data includes data on the robot's motion state; S2: Based on the multimodal sensing data and internal state data, generating a first breathing control parameter through a physiological normality path module, the first breathing control parameter being used to characterize a basic breathing pattern coordinated with the robot's motion state and environmental factors; and coordinating and optimizing the first breathing control parameter based on the voice data to be output to generate voice-coordinated basic breathing parameters; S3: Based on the multimodal sensing data, generating a second breathing control parameter through an emotion decision path; S4: Fusing the voice-coordinated basic breathing parameter with the second breathing control parameter to generate a breathing control command for the robot; S5: Driving the breathing actuator to work according to the breathing control command.

[0007] Furthermore, this application also proposes that the process of coordinating and optimizing the first breathing control parameters includes: planning the timing and airflow intensity of inhalation and exhalation at the corresponding time points of the basic breathing pattern based on the phrasing, pauses and volume changes of the speech data to be output.

[0008] Furthermore, this application also proposes that the physiological normal pathway module in step S2 further includes: matching preset respiratory frequency and tidal volume models under different exercise intensities based on the robot's motion sensor data.

[0009] Furthermore, this application also proposes that the physiological normal pathway module in step S2 is also used to: trigger a protective breathing mode when the environmental sensor detects abnormal gas, the protective breathing mode including breath-holding, slowed breathing or rapid breathing.

[0010] Furthermore, this application proposes that the emotion decision-making path is used to: interpret the context of multimodal sensor data based on large-scale model semantic analysis to obtain context labels that include scene semantics, relationships between interactive objects, and historical interaction context; determine the robot's emotion type and emotion intensity based on the context labels; and output the robot's emotion code based on the emotion type and emotion intensity through a preset multidimensional emotion model calculation.

[0011] Furthermore, this application also proposes a biomimetic robot breathing system for running the above method, comprising: a multimodal sensor group for collecting environmental and body data; a processor, communicatively connected to the multimodal sensor group for generating breathing control commands; and a breathing actuator, connected to the processor for receiving breathing control commands and executing breathing actions.

[0012] Furthermore, this application also proposes a biomimetic robot breathing device, comprising: a deformation generating device disposed inside the robot's torso, used to generate periodic physical deformation under gas drive to simulate the surface undulations caused by breathing; an airflow device for generating and regulating airflow according to breathing control commands; at least one biomimetic air outlet disposed at the nasal cavity and / or oral cavity model position of the robot's head; and a gas passage connecting the airflow device, the deformation generating device and the biomimetic air outlet.

[0013] Furthermore, this application also proposes that the deformation generating device is a biomimetic breathing airbag.

[0014] Furthermore, this application also proposes that at least one surface of the biomimetic breathing airbag is in contact with a support structure, the support structure being constructed to mimic the shape of a human rib.

[0015] Furthermore, this application also proposes that the biomimetic breathing airbag has an elastic biomimetic soft tissue layer covering the surface opposite to the supporting structure.

[0016] Furthermore, this application also proposes that the airflow device includes an air pump.

[0017] Furthermore, this application also proposes that a mechanism for adjusting airflow characteristics be provided at the biomimetic air outlet.

[0018] Furthermore, this application also proposes a robot including the aforementioned bionic robotic respiratory system.

[0019] As can be seen from the above, the bionic robot breathing control method, system, device and robot provided in this application generate breathing control commands that dynamically adapt to motion state, environmental factors and emotional expression through the collaborative processing of multimodal sensing data and voice data to be output, and realize realistic breathing actions using a bionic hardware structure. It has the advantages of being able to dynamically adjust the breathing mode according to the robot's motion state and environmental changes, realize the natural coordination of emotional expression and voice interaction, and improve the realism of breathing simulation. Attached Figure Description

[0020] Figure 1 A flowchart of a biomimetic robot breathing control method provided in this application;

[0021] Figure 2 This is a schematic diagram of the structure of a biomimetic robotic breathing device provided in this application;

[0022] In the diagram: 1-Deformation generating device; 2-Bionic air outlet. Detailed Implementation

[0023] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0024] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0025] In existing technologies, the field of bionic robots has long faced the challenge of incomplete physiological behavior simulation. Traditional solutions rely on motor-driven mechanical chest rise and fall, which cannot adjust breathing frequency according to exercise intensity and lacks adaptive response to environmental changes. When the robot is running, its breathing rhythm remains constant; it also fails to trigger protective breathing responses in smoky environments. During voice interaction, breathing movements are disconnected from sentence pauses, resulting in a lack of realism in the vocalization process.

[0026] To address these issues, researchers discovered that respiratory control requires integrating both physiological norms and emotional decision-making mechanisms. First, a multimodal data acquisition system needs to be established to translate environmental factors and bodily states into respiratory control parameters. Second, the timing of voice interaction and respiratory rhythm needs to be matched to prevent mechanical ventilation from disrupting the naturalness of the conversation. Finally, respiratory parameter fusion algorithms are being explored to achieve a dynamic balance between basic physiological patterns and emotional expression.

[0027] Therefore, this application proposes a technical solution to acquire multimodal sensor data, internal state data and voice data to be output from the robot, generate first breathing control parameters based on the multimodal sensor data and internal state data and perform voice coordination optimization, combine the second breathing control parameters generated by the emotion decision path, fuse the two, and finally generate a breathing control command to drive the execution device.

[0028] Multimodal sensing data refers to data collected by environmental sensors, such as temperature, gas composition, and light intensity, and carbon dioxide concentration detected by gas sensors. Multimodal sensing data provides a basis for environmental adaptation in adjusting breathing patterns.

[0029] The voice data to be output refers to the audio waveform data generated by the processor that will be expressed through the speaker.

[0030] Internal state data refers to acceleration and angular velocity data acquired through inertial measurement units, such as using a six-axis sensor to monitor the torso vibration frequency during robot walking. Internal state data reflects the impact of bodily motion load on breathing depth.

[0031] The physiological norm pathway module refers to a computational model based on a preset mapping relationship between exercise intensity and respiratory rate, such as establishing a linear correspondence between exercise speed and tidal volume. This module ensures that the basic breathing pattern conforms to the physiological laws of human exercise. The respiratory actuator refers to an airflow control system that includes an air pump. This device translates control commands into observable respiratory actions.

[0032] Specifically, when the robot is walking, the motion sensor detects an increased gait frequency, and the physiological normalization path module increases the breathing rate according to a preset motion intensity model. Simultaneously, when the environmental sensor detects excessive pollen concentration, a breathing slowing mode is triggered to reduce the risk of allergies. During voice interaction, by recognizing the end of sentences, the airflow intensity attenuation is planned during the exhalation phase, ensuring the breathing rhythm naturally matches the pauses in the dialogue. The emotion decision path generates rapid breathing parameters based on the dialogue content, which, when superimposed on the basic breathing pattern, drives the air pump to produce breathing airflow with emotional characteristics.

[0033] Compared to existing technologies, traditional solutions using fixed-frequency motor drives cannot achieve dynamic adjustment of breathing in relation to exercise intensity. This solution achieves dynamic optimization of breathing parameters through multi-sensor fusion. In existing technologies, breathing control and the voice system operate independently; this solution achieves precise synchronization between breathing and speech rhythm through voice signal analysis. Traditional hardware can only simulate chest rise and fall; this solution, through the coordinated control of the airflow device and deformation mechanism, reproduces the airflow dynamics characteristics during breathing.

[0034] Through the above technical solutions, this application achieves real-time adaptation of breathing patterns to movement states, automatically enhancing ventilation efficiency during strenuous exercise. When harmful gases are detected, protective breathing behaviors are triggered to prevent system damage. In dialogue interaction, breathing points are accurately matched with sentence pauses, eliminating the unnaturalness of mechanical breathing. By integrating emotional decision-making parameters, breathing movements can convey emotional states such as anxiety and calmness.

[0035] This application further proposes a process for coordinating and optimizing the first respiratory control parameters, including planning the timing and airflow intensity of inhalation and exhalation at the corresponding time points of the basic breathing pattern based on the phrasing, pauses, and volume changes of the speech data to be output.

[0036] Sentence segmentation refers to the natural segmentation points of sentences in a speech signal. Specifically, it can be achieved by detecting the silence intervals between sentences using speech recognition algorithms, and is used to determine the starting position of breathing and exchanging gases.

[0037] Pauses refer to brief silences within a sentence, which can be achieved by identifying low-energy regions in the speech waveform using the energy threshold method, and are used to control subtle adjustments to the breathing rhythm.

[0038] Volume change refers to the dynamic fluctuation of speech amplitude, which can be quantified by calculating the root mean square energy value of the speech frame to match the intensity changes of breathing airflow.

[0039] Specifically, during voice interaction, the air exchange window is determined by real-time monitoring of sentence segmentation points, triggering an inhalation action at the end of the sentence to replenish airflow reserves. When a semantically emphatic pause is detected within a sentence, the airflow maintenance time during the exhalation phase is simultaneously extended. Furthermore, the output power of the air pump is dynamically adjusted based on the current increase or decrease in speech amplitude, ensuring a positive correlation between breathing intensity and speech loudness.

[0040] Compared to existing technologies, current robotic breathing control solutions only employ a fixed-frequency chest rise and fall pattern, failing to dynamically adjust the breathing rhythm based on the content of voice interaction. This solution, by analyzing the prosodic features of the voice signal, establishes a temporal and intensity-based synergy between breathing actions and verbal expression, thus resolving the disconnect between mechanical breathing and voice output during dialogue.

[0041] Through the above technical solution, this application realizes the dynamic matching of breathing actions and language expression of bionic robots during voice interaction, so that the robot can automatically breathe at an appropriate position when uttering long sentences, and synchronously adjust the breathing intensity when expressing changes in emotional intensity, thereby eliminating the sense of incoordination between breathing rhythm and voice output during the robot's speech process.

[0042] This application further proposes a model that matches preset respiratory rate and tidal volume under different motion intensities based on the robot's motion sensor data.

[0043] Among them, motion sensor data refers to the robot limb motion parameters collected by accelerometers, gyroscopes or joint encoders. Specifically, it can be achieved by combining a three-axis accelerometer and an angular velocity sensor, which is used to quantify the robot's motion amplitude and energy consumption.

[0044] Among them, the preset respiratory rate and tidal volume model refers to the pre-established respiratory parameter database, which can be implemented by mapping the respiratory rate curve to the exercise intensity level. For example, the exercise intensity is divided into three levels: rest, walking and running. Each level corresponds to a different respiratory rate range and tidal volume value, which is used to simulate the physiological changes in breathing depth and rhythm during human exercise.

[0045] Specifically, when the robot performs actions of varying intensities, motion sensors collect data in real time, such as the rate of change of joint angles and trunk acceleration, and determine the current level of motion intensity by calculating the motion power value. Then, it calls upon the corresponding combination of respiratory frequency and tidal volume parameters from a preset model; for example, it uses a high-frequency, shallow breathing mode when running and a low-frequency, deep breathing mode when stationary. This process dynamically adjusts the airflow output frequency and single-breath volume of the respiratory actuator to achieve physiological coordination between the robot's breathing pattern and limb movements.

[0046] Compared to existing technologies, traditional solutions control breathing rhythms through fixed programs and cannot distinguish differences in exercise intensity. This solution, however, establishes a mapping relationship between exercise intensity and breathing parameters, enabling automatic switching of breathing modes based on the robot's actual movement amplitude. For example, it triggers rapid breathing when lifting heavy objects and switches to gentle breathing when resting, thus more realistically simulating the linkage mechanism between human movement and breathing.

[0047] Through the above technical solution, this application solves the problem of the disconnect between the robot's breathing mode and its movement state, enabling the breathing frequency and tidal volume to be dynamically adjusted according to the intensity of limb movements. For example, the breathing frequency is automatically increased when the robot moves quickly to match energy consumption, and the breathing frequency is reduced when the robot is stationary to simulate a relaxed state, which significantly enhances the physiological rationality and scene adaptability of the breathing behavior.

[0048] This application further proposes triggering a protective breathing mode when an environmental sensor detects an abnormal gas. The protective breathing mode includes breath-holding, slowed breathing, or rapid breathing.

[0049] Among them, environmental sensors refer to sensing devices used to detect changes in gas composition, specifically electrochemical sensors or semiconductor gas sensors, capable of real-time monitoring of gas concentration changes in the robot's environment. Abnormal gases refer to harmful or irritating gases exceeding preset safety thresholds, specifically determined by setting concentration thresholds for carbon monoxide, formaldehyde, or PM2.5. Protective breathing modes refer to breathing strategies dynamically adjusted according to the level of environmental threat, specifically implemented by using different breathing parameters corresponding to gas concentration gradients. For example, a breathing slowdown mode is triggered when mild pollution is detected, while a breath-holding action is performed when severe pollution is detected.

[0050] Specifically, during robot operation, environmental sensors continuously collect gas data and transmit it to the processor. When a specific gas concentration exceeds a preset threshold, the physiological normal pathway module immediately interrupts the basic breathing pattern generation process and instead invokes a preset breathing strategy library. For example, when excessive smoke concentration is detected, the breathing actuator stops inhalation and enters a breath-holding state, while simultaneously activating the internal air circulation system; when insufficient oxygen concentration is detected, a shallow and rapid breathing mode is switched to reduce oxygen consumption. The duration of the protective breathing mode can be dynamically adjusted based on environmental data until the sensor detection data returns to a safe range.

[0051] Compared to existing technologies, traditional robotic respiratory systems lack environmental perception and emergency response capabilities, maintaining a fixed breathing pattern even in hazardous environments, which may lead to damage to internal components or inappropriate interaction. This solution establishes a dynamic mapping relationship between environmental threats and breathing patterns, enabling the robot to make immediate physiological responses to environmental hazards, just like a living organism.

[0052] Through the above technical solution, this application effectively solves the technical defect of bionic robots' distorted breathing behavior in harmful gas environments, avoids the risk of mechanical structure corrosion caused by continuous inhalation of polluted air, and enhances the credibility of environmental response in human-computer interaction scenarios through adaptive adjustment of breathing patterns.

[0053] This application further proposes an emotion decision-making path for: interpreting multimodal sensor data based on large-model semantic analysis to obtain contextual labels containing scene semantics, interaction object relationships and historical interaction context; determining the robot's emotion type and emotion intensity based on the contextual labels; and calculating and outputting the robot's emotion code based on the emotion type and emotion intensity through a preset multidimensional emotion model.

[0054] Among them, large-scale model semantic analysis refers to using pre-trained language models to perform semantic understanding of multimodal sensor data. Specifically, it can be achieved using neural network models based on the Transformer architecture, such as using BERT or GPT series models to perform intent recognition on the output speech data and combining visual sensor data to analyze the interaction scenario.

[0055] Among them, contextual tags refer to metadata that provides a structured description of the interaction scenario. Specifically, this can be achieved by combining knowledge graph entity annotation with time series association analysis. For example, key entities in the dialogue content can be spatiotemporally associated with the robot's motion state.

[0056] Among them, the multidimensional emotion model refers to a mathematical model that quantifies the emotion dimension into computable parameters. Specifically, it can be implemented using a ring emotion model or a PAD three-dimensional emotion space model. For example, the emotion intensity can be mapped to the respiratory rate adjustment coefficient through the valence-arousal coordinate system.

[0057] Specifically, when the robot is in a service scenario, it identifies the user's urgent needs through large-scale semantic analysis and determines that the current dialogue is a high-priority event based on historical interaction data. The context labeling module marks such scenarios as "urgent service requests." The emotion decision path calculates the respiratory rate increase coefficient and respiratory depth decrease parameter based on a preset anxiety emotion parameter space. The multidimensional emotion model quantifies the emotion intensity into a modulation factor in the 0-1 range, and finally outputs an emotion code containing rapid breathing characteristics. This code will be used as the modulation parameter for breathing control commands.

[0058] Compared to existing technologies, traditional methods rely solely on simple sentiment keyword matching, failing to analyze the contextual relationships within a dialogue scenario, leading to a mismatch between breathing patterns and actual emotional states. This solution constructs dynamic contextual understanding capabilities through large-scale model semantic analysis, combined with multi-dimensional sentiment space modeling, enabling breathing control to accurately reflect emotional changes in complex interactive scenarios.

[0059] Through the above technical solution, this application solves the problem of disconnect between emotional computing and breathing control in the prior art, and realizes dynamic adaptation of breathing patterns to semantic scenarios. In medical care scenarios, when a patient's emotional fluctuations are detected, the breathing device can synchronously generate a soothing breathing rhythm; in educational companionship scenarios, it automatically matches corresponding breathing emotional expressions according to the development of the storyline, significantly improving the realism of emotional transmission in human-computer interaction.

[0060] This application further proposes a biomimetic robotic breathing system, including a multimodal sensor group, a processor, and a breathing actuator. The multimodal sensor group is used to collect environmental and body data, the processor communicates with the multimodal sensor group and generates breathing control commands, and the breathing actuator receives the breathing control commands and performs breathing actions.

[0061] The multimodal sensor array refers to a composite data acquisition device integrating multiple sensing units. Specifically, it can be implemented using a combination of temperature sensors, gas concentration sensors, motion acceleration sensors, and microphone arrays to simultaneously acquire the physical parameters of the robot's environment and its operational status. The processor is a computing unit with data fusion and decision-making capabilities. Specifically, it can be implemented using an embedded system with multi-threaded control algorithms to convert sensor data into respiratory rhythm regulation signals. The respiratory actuator is a driving mechanism capable of generating airflow and mechanical deformation. Specifically, it can be implemented using a linkage structure of an air pump, solenoid valve, and biomimetic airbag to generate physiologically characteristic breathing movements based on control commands.

[0062] Specifically, the multimodal sensor array continuously collects ambient temperature, oxygen content, motion acceleration, and voice signals, which are transmitted to the processor via a communication interface. The processor analyzes the data in real time and, combined with a preset breathing pattern algorithm, generates control commands that include inhalation duration, airflow intensity, and chest rise and fall amplitude. The air pump in the breathing actuator adjusts the air supply flow according to the commands, the solenoid valve controls the airflow direction, and the simulated airbags periodically expand and contract under air pressure, thereby simulating breathing behavior that matches the robot's current motion state and environmental conditions.

[0063] Compared to existing technologies, traditional robotic breathing devices only control motors via timers to achieve fixed-frequency chest cavity movements, failing to perceive environmental changes or the body's state. This solution utilizes a multimodal sensor array to achieve synchronous acquisition of environmental and bodily data. The processor generates breathing control commands based on dynamic data, enabling real-time adjustments to breathing movements according to external environmental stimuli and changes in movement intensity, thus solving the problem of fixed breathing patterns in existing technologies.

[0064] Through the above technical solutions, this application achieves dynamic adaptation of breathing patterns to environmental factors. For example, it automatically increases the breathing rate when an increase in motion acceleration is detected, and triggers a protective breathing response when harmful gases are detected. Simultaneously, the breathing actuator, through the synergistic effect of airflow conduction and mechanical deformation, can reproduce the physiological characteristics of synchronized chest rise and fall and nasal airflow during human breathing, improving the realism of the robot's breathing simulation and the naturalness of human-computer interaction.

[0065] This application further proposes a biomimetic robotic breathing device, such as Figure 2 As shown, the robot includes: a deformation generating device 1, an airflow device (not shown), at least one bionic air outlet 2, and a gas passage (not shown). The deformation generating device 1 is located inside the robot's torso and is used to generate periodic physical deformations under gas drive to simulate the surface undulations caused by breathing; the airflow device is used to generate and regulate airflow according to breathing control commands; the bionic air outlet 2 is located at the nasal cavity and / or oral cavity model position of the robot's head; the gas passage connects the airflow device, the deformation generating device 1, and the bionic air outlet 2.

[0066] The deformation generating device 1 refers to a component that induces periodic expansion and contraction of the mechanical structure through changes in gas pressure. Specifically, it can be implemented using a biomimetic breathing airbag, which simulates the rise and fall of the chest cavity through inflation and deflation. The airflow device refers to a device that generates and regulates gas flow, which can be implemented using a combination of an air pump and valves. It controls the gas flow rate and pressure to match different breathing modes. The biomimetic air outlet 2 refers to a structure that simulates the external morphology of the human respiratory organs, which can be implemented using a nasal cavity model with a microporous structure, its surface morphology matching human anatomical features. The gas passage refers to the gas transmission channel connecting the various components, which can be implemented using flexible silicone tubing, its inner diameter adaptable to different airflow rate requirements.

[0067] Specifically, breathing control commands drive the airflow device to generate gas flow. The gas enters the deformation generating device 1 through a passage, causing it to deform. The amplitude and frequency of the deformation are controlled by airflow parameters. Simultaneously, some gas is transported through the passage to the bionic exhaust port 2 for discharge, creating a perceptible airflow change. The periodic movement of the deformation generating device 1 is transmitted through the robot's shell, forming visible breathing fluctuations on the body surface. The gas distribution ratio in the passage can be adjusted by valves; for example, the deformation device is primarily driven during calm breathing, while the airflow output of the bionic exhaust port 2 is increased during deep breathing.

[0068] Compared to existing technologies, traditional methods rely on motor-driven rigid structures for simple reciprocating motion, achieving only mechanical fluctuations at a fixed frequency. This new method, through the coordinated action of a gas-driven deformation device and an airflow device, not only replicates the coordinated deformation of the chest cavity and abdomen during respiration but also synchronously generates airflow changes that conform to physiological characteristics. The integrated design of the gas pathway dynamically couples the deformation motion and airflow output during the breathing simulation, more closely resembling the synchronous relationship between airflow and body surface movement during real human respiration.

[0069] Through the above technical solutions, this application solves the problem of insufficient hardware realism in traditional robotic breathing devices, achieving coordinated simulation of body surface undulations and respiratory airflow. The gas-driven deformation generator 1 produces compliant movements that more closely resemble biological tissue, avoiding the mechanical feel of motor-driven systems. The topological design of the gas passage allows for independent control of chest cavity movement and nasal / oral airflow during breathing, providing a hardware foundation for subsequent coordinated control with voice interaction and emotional expression. The anatomical design of the biomimetic air outlet 2 enhances the external visibility of breathing behavior, contributing to improved natural perception in human-computer interaction.

[0070] This application further proposes that the deformation generating device 1 is a biomimetic breathing airbag.

[0071] Among them, the biomimetic breathing airbag refers to an inflatable and deflated structure made of flexible materials, specifically silicone or polymer composite materials, which generates deformation similar to the expansion and contraction of the lungs through gas drive. This airbag simulates the chest cavity movement during human breathing through periodic volume changes, and its flexibility and deformation trajectory can match the mechanical properties of different breathing modes.

[0072] Specifically, biomimetic breathing airbags are placed inside the robot's torso. When gas is injected or expelled controlled by an air pump and valves, the surface of the airbag undergoes undulating deformation. For example, during inhalation, the airbag inflates, causing the front of the robot's torso to bulge; during exhalation, the airbag contracts, restoring the torso to its original shape. The amplitude and speed of the airbag deformation can be controlled by adjusting the gas flow rate and pressure, thereby simulating different states such as deep breathing, shallow breathing, or rapid breathing.

[0073] Compared to existing technologies, traditional solutions use rigid motors to drive mechanical components to generate regular fluctuations, resulting in a single motion trajectory and a lack of biomechanical adaptability. In contrast, the biomimetic breathing airbag, through the elastic deformation properties of flexible materials, can reproduce the non-linear chest cavity movement pattern during human respiration. At the same time, the internal cavity structure of the airbag can work with an airflow device to achieve dynamic coupling between respiration and airflow.

[0074] Through the above technical solution, this application solves the problem of low hardware realism in existing robotic breathing devices, making the surface undulations of the robot's torso during breathing more closely resemble human physiological characteristics. The flexible deformation characteristics of the airbag can accommodate dynamic adjustments of different breathing frequencies and amplitudes, providing a physical basis for simulating complex breathing behaviors such as coughing and sighing.

[0075] This application further proposes that at least one surface of the biomimetic breathing airbag is in contact with a support structure, the support structure being constructed to mimic the shape of a human rib.

[0076] The supporting structure refers to the rigid or semi-rigid frame that provides physical support for the bionic breathing airbag. It can be made of metal alloys or engineering plastics through 3D printing, and its function is to maintain the stability of the breathing airbag's deformation trajectory. Simulating the morphology of human ribs means that the curvature distribution and connection method of the supporting structure conform to the anatomical characteristics of the human thoracic skeleton. This can be achieved through reverse engineering scanning of human rib data to create a model, and its function is to make the expansion and contraction path of the breathing airbag closer to real physiological movement patterns.

[0077] Specifically, the support structure is designed with multiple arc-shaped support units, which are interconnected by elastic connectors to form an adjustable linkage. When the bionic breathing airbag deforms under gas pressure, the arc-shaped units of the support structure elastically deform simultaneously, limiting the disordered expansion of the airbag in a single plane and guiding it to undulate along a preset breathing direction. For example, during inhalation, the elastic connectors of the support structure allow the spacing between adjacent arc-shaped units to increase, enabling the airbag to simulate the lateral expansion of the chest cavity; during exhalation, the elastic restoring force of the connectors pushes the arc-shaped units back to their original position, assisting the airbag in completing its contraction action.

[0078] In some specific embodiments, the surface of the arc-shaped unit of the support structure can be provided with sliding guide rails to form a sliding fit with the limiting protrusions on the outer wall of the airbag, further constraining the deformation direction of the airbag. In addition, the material hardness of the support structure can vary along the axial gradient, for example, the hardness near the spine simulation area is higher than that of the anterior chest area, in order to match the mechanical properties of different segments of the human rib.

[0079] Compared to existing technologies, which use planar support plates or simple spring structures, the deformation path of the breathing airbag lacks three-dimensional constraints, resulting in stiff breathing movements and significant deviations from human movement patterns. This solution, however, uses a biomimetic rib-shaped support structure to ensure that the airbag deformation process conforms to the three-dimensional movement characteristics of the chest cavity during human respiration. Furthermore, through the synergistic effect of elastic connectors, it achieves more natural breathing fluctuations while maintaining structural stability.

[0080] Through the above technical solution, this application solves the problem of insufficient hardware realism in existing breathing devices, enabling robots to reproduce the multi-dimensional motion characteristics of the human chest cavity when performing breathing actions, improving the realism and motion coordination of breathing simulation, and reducing the risk of mechanical damage caused by disordered deformation through the guiding effect of the support structure on the deformation of the air bladder.

[0081] This application further proposes that the biomimetic breathing airbag has an elastic biomimetic soft tissue layer covering the surface away from the supporting structure.

[0082] The elastic biomimetic soft tissue layer refers to a flexible covering layer with the mechanical properties of biological soft tissue. It can be made of silicone or thermoplastic elastomer materials, with a thickness ranging from 1.5 to 3 mm to simulate the combined elastic modulus of the human epidermis and subcutaneous tissue. This layer is fixed to the airbag surface through bonding or in-mold molding, serving to transmit natural tactile sensation and buffer mechanical stress during airbag deformation. The support structure refers to a rigid or semi-rigid frame with a rib-like shape, specifically made of 3D-printed lightweight alloys or engineering plastics. Its arc curvature radius can be set from 50 to 80 mm to match the anatomical features of the human thoracic cavity, constraining the deformation direction of the airbag and simulating the linkage mechanism of the ribs during breathing.

[0083] Specifically, when breathing control commands drive the airbag to periodically contract and expand, the elastic biomimetic soft tissue layer moves in sync with the airbag's deformation. The rib-like shape of the supporting structure limits the airbag's displacement in a preset direction, while the elastic layer absorbs the impact energy generated by the airbag's rapid deformation through the material's viscoelasticity. During exhalation, the elastic layer helps the airbag return to its initial shape due to the material's resilience, and its surface texture design can simulate the subtle wrinkles of human skin.

[0084] Compared to existing technologies, traditional breathing simulation devices use only a single rigid material shell, resulting in stiff tactile feedback and unnatural deformation trajectories. This solution, through the synergistic effect of an elastic layer and a supporting structure, simultaneously improves the realism of tactile sensation during breathing movements, the biosimilarity of deformation trajectories, and the durability of the device while maintaining mechanical drive efficiency.

[0085] Through the above technical solution, this application solves the problems of stiff tactile sensation and deformation trajectory that does not conform to the laws of human movement in existing robot breathing devices, enabling the robot to convey more realistic life characteristics through tactile sensation on the body surface during interaction, while reducing the device failure rate caused by mechanical fatigue.

[0086] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A biomimetic robotic breathing device, characterized in that, include: A deformation generating device, located inside the robot's torso, is used to generate periodic physical deformations under gas-driven conditions to simulate the surface undulations caused by breathing. An airflow device used to generate and regulate airflow according to breathing control commands; At least one bionic air vent is located at the position of the nasal cavity and / or oral cavity model of the robot's head; A gas passage connects the airflow device, the deformation generating device, and the biomimetic air outlet. The driving method of the biomimetic robotic breathing device includes: S1: Acquire multimodal sensor data, internal state data, and voice data to be output from the robot's external environment; the multimodal sensor data includes environmental data from the robot's external environment; the internal state data includes data on the robot's motion state. S2: Based on the multimodal sensing data and internal state data, a first breathing control parameter is generated through the physiological normal path module. The first breathing control parameter is used to characterize the basic breathing pattern that is coordinated with the robot's motion state and environmental factors. Based on the voice data to be output, the first breathing control parameter is coordinated and optimized to generate the basic breathing parameter after voice coordination. S3: Based on the multimodal sensing data, generate second breathing control parameters through an emotion decision path; S4: The basic breathing parameters after voice coordination are fused with the second breathing control parameters to generate breathing control commands for the robot; S5: Execute the drive according to the breathing control command.

2. The biomimetic robotic breathing device according to claim 1, characterized in that, The deformation generating device is a biomimetic breathing airbag.

3. The biomimetic robotic breathing device according to claim 2, characterized in that, At least one surface of the biomimetic breathing bladder is in contact with a support structure configured to mimic the shape of a human rib.

4. The biomimetic robotic breathing device according to claim 3, characterized in that, The biomimetic breathing airbag is covered with an elastic biomimetic soft tissue layer on the surface opposite to the supporting structure.

5. The biomimetic robotic breathing device according to claim 1, characterized in that, The airflow device includes an air pump.

6. The biomimetic robotic breathing device according to claim 1, characterized in that, The biomimetic air outlet is equipped with a mechanism for adjusting airflow characteristics.

7. A method for controlling the breathing of a biomimetic robot, characterized in that, Includes the following steps: S1: Acquire multimodal sensor data, internal state data, and voice data to be output from the robot's external environment; the multimodal sensor data includes environmental data from the robot's external environment; the internal state data includes data on the robot's motion state. S2: Based on the multimodal sensing data and internal state data, a first breathing control parameter is generated through the physiological normal path module. The first breathing control parameter is used to characterize the basic breathing pattern that is coordinated with the robot's motion state and environmental factors. Based on the voice data to be output, the first breathing control parameter is coordinated and optimized to generate the basic breathing parameter after voice coordination. S3: Based on the multimodal sensing data, generate second breathing control parameters through an emotion decision path; S4: The basic breathing parameters after voice coordination are fused with the second breathing control parameters to generate breathing control commands for the robot; S5: Drive the bionic robot breathing device as described in claim 1 to work according to the breathing control command.

8. The biomimetic robot breathing control method according to claim 7, characterized in that, The process of coordinating and optimizing the first respiratory control parameters includes: Based on the phrasing, pauses, and volume changes of the speech data to be output, the timing and airflow intensity of inhalation and exhalation are planned at the corresponding time points of the basic breathing pattern.

9. The biomimetic robot breathing control method according to claim 7 or 8, characterized in that, The physiological normal pathway module in step S2 also includes: Based on the robot's motion sensor data, preset breathing frequency and tidal volume models are matched for different motion intensities.

10. The biomimetic robot breathing control method according to claim 7 or 8, characterized in that, The physiological normal pathway module in step S2 is also used for: When an environmental sensor detects an abnormal gas, a protective breathing mode is triggered, which includes breath-holding, slowed breathing, or rapid breathing.

11. The biomimetic robot breathing control method according to claim 7, characterized in that, The sentiment decision path is used to: interpret the context of the multimodal sensing data based on large model semantic analysis to obtain contextual labels that include scene semantics, relationships between interactive objects, and historical interaction context; Based on the contextual tags, determine the robot's emotion type and emotion intensity; Based on the emotion type and intensity, the robot's emotion code is output through calculation using a preset multidimensional emotion model.

12. A biomimetic robotic breathing system for operating the method as described in any one of claims 7-11, characterized in that, include: A multimodal sensor array is used to collect environmental and organismal data; The processor is communicatively connected to the multimodal sensor group and is used to generate breathing control commands; A breathing device, connected to the processor, is used to receive the breathing control command and perform breathing actions.

13. A robot, characterized in that, Including the bionic robotic breathing system as described in claim 12.

Images