Robot lacrimation control method and microfluid lacrimation execution device for robot

Abstract

The invention discloses a robot lacrimation control method and a microfluid lacrimation execution device for a robot, and relates to the technical field of bionic robot emotional interaction.The robot lacrimation control method comprises the steps that multi-modal sensing data of the robot is obtained, determining a first code used for judging an emotional state and a second code used for judging a physical stimulation response of the robot according to the multi-modal sensing data; and according to the first code and / or the second code, determining an emotion code of the robot, and according to the emotion code, determining a tear instruction for controlling the operation of a microfluid tear execution device of the robot, so that the robot tears. The emotional codes are generated through an emotional state and physical stimulation double-path mechanism, the simulated lacrimation effect matched with the human emotional state can be achieved, the purpose of rich physiological feedback is achieved, and the lacrimation performance is improved through a liquid storage device, a lacrimation channel, a lacrimation outlet and a miniature pressure pump of the lacrimation execution device. And the robot can realize accurate tear control on the basis of emotion perception of physiological feedback.

Description

Technical Field

[0001] The present invention relates to the technical field of bionic robot emotional interaction, and in particular to a robot tearing control method and a microfluidic tearing execution device for the robot. Background Art

[0002] As time progresses, a wide variety of robots are appearing in production and daily life. Common types of bionic robots include humanoid robots, which can mimic human movements and behaviors, possessing a high degree of freedom and locomotion. Animal-like bionic robots, such as robot dogs and robot fish, can mimic the movements of animals and are used for exploration, rescue, and other tasks.

[0003] Current bionic robots’ emotional expressions mostly rely on facial expressions or voice modules, and lack the realism of physiological feedback such as tears and sweating. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to address the deficiencies of the existing technology, specifically the lack of physiological feedback and inaccurate tear control in bionic robots. A robot tear control method and a microfluidic tear execution device for the robot are provided, specifically as follows:

[0005] 1) In a first aspect, the present invention provides a method for controlling robot tears. The specific technical solution is as follows, including:

[0006] S1, acquiring multimodal sensor data of a robot, and determining a first code for determining an emotional state and a second code for determining a physical stimulus response of the robot based on the multimodal sensor data;

[0007] S2, determining the emotion code of the robot according to the first code and / or the second code, and determining the tearing instruction for controlling the operation of the microfluidic tearing execution device of the robot according to the emotion code, so as to make the robot shed tears.

[0008] Based on the above solution, the present invention can also be improved as follows.

[0009] Furthermore, the multimodal sensing data includes:

[0010] At least one of visual sensor data, sound sensor data, tactile sensor data, pressure sensor data, temperature sensor data, and smoke sensor data.

[0011] Furthermore, the first code used by the robot to judge the emotional state is determined based on the multimodal sensing data to be:

[0012] The multimodal sensing data is calculated based on large model semantic analysis to obtain a first code of the robot.

[0013] Furthermore, the specific process of calculating the multimodal sensor data based on the large model semantic analysis is as follows:

[0014] Determine the context labels of the multimodal sensor data according to the semantic analysis of the large model, and determine N basic emotion data of the robot and the emotion intensity corresponding to each basic emotion data according to the multimodal sensor data and the context labels through the Ekman algorithm;

[0015] Determine the robot's composite emotion data based on the context label, N basic emotion data and their corresponding emotion intensities using the Plutchik algorithm;

[0016] Based on the context label, the refined emotion label corresponding to the complex emotion data is determined according to a preset multidimensional emotion model, and the first code of the robot is determined according to the refined emotion label.

[0017] Furthermore, the second code for the robot to judge the physical stimulus response is determined based on the multimodal sensing data:

[0018] According to the physical sensing data in the multimodal sensing data, when the physical sensing data meets a preset condition, a hardware interrupt of the robot is triggered based on the physical sensing data to generate a second code of the robot.

[0019] Furthermore, it also includes:

[0020] The emotion weight of the robot is determined according to the environmental factors, relationship factors and role factors of the robot, the emotion weight is used as the weight value of the first code, and the emotion code is corrected.

[0021] Furthermore, the emotion weight of the robot is determined based on the robot's environmental factors, relationship factors, and role factors as follows:

[0022] W_adj = α×W_env + β×W_relation + γ×W_role;

[0023] Among them, W_adj is the sentiment weight, W_env is the environmental factor, α is the environmental factor coefficient, W_relation is the relationship factor, β is the relationship factor coefficient, W_role is the role factor, and γ is the role factor coefficient.

[0024] 2) In a second aspect, the present invention provides a microfluidic tearing actuator for a robot, the specific technical solution of which is as follows, including:

[0025] A fluid storage device, a tear channel, a tear outlet, and a micro pressure pump;

[0026] The liquid storage device is used to store tear fluid;

[0027] One end of the tear channel is connected to the liquid storage device, and the other end of the tear channel is connected to the tear outlet;

[0028] The tear outlet is provided at a preset position of the robot's eye for draining tears;

[0029] The micro pressure pump is arranged on the tear channel and is used to pump the tear liquid in the liquid storage device to the tear outlet through the tear channel.

[0030] Based on the above solution, the present invention can also be improved as follows.

[0031] Furthermore, the tear outlet is provided with a bionic lacrimal gland micropore array.

[0032] 3) In a third aspect, the present invention provides a robot tearing system, comprising: a microfluidic tearing execution device for a robot as in the second aspect and a processor, wherein the processor is used to execute a robot tearing control method as in the first aspect.

[0033] The robot tearing control method and the microfluidic tearing actuator for the robot provided by the present invention have the following beneficial effects:

[0034] Based on multimodal sensory data, a first code for determining emotional state and a second code for determining physical stimulus response are determined. This innovative dual-path triggering mechanism, based on both emotional state and physical stimulation, generates emotional codes. Combined decision-making using the first and / or second codes enables the robot to simulate physiological feedback from both emotional state and physical stimulation. Furthermore, by incorporating the tear flow mechanism of a microfluidic hardware device, the robot achieves precise tear control based on the emotional perception of physiological feedback.

[0035] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are specifically listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments made with reference to the following drawings:

[0037] Figure 1 This is a logical architecture diagram of an embodiment of the present invention. DETAILED DESCRIPTION

[0038] To make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0039] Figure 1 This is a logical architecture diagram of an embodiment of the present invention. As shown in the figure, a robot tear control method provided by the present invention includes:

[0040] S1, obtain the robot's multimodal sensor data, which includes at least one of: visual sensor data, sound sensor data, tactile sensor data, pressure sensor data, temperature sensor data, and smoke sensor data. Correspondingly, the structure for obtaining multimodal sensor data is called a multimodal perception layer, which includes: visual sensors, sound sensors, tactile pressure sensor matrix (density ≥ 4 / cm²), pressure sensors, temperature sensors, and smoke sensors, etc. Multiple sensor data are obtained through multiple sensors and integrated to obtain multimodal sensor data (corresponding to Figure 1 "Multimodal Sensors" in [1].

[0041] The first coding for the robot to judge the emotional state based on the multimodal sensor data is specifically as follows: the multimodal sensor data is calculated based on the semantic analysis of the large model (corresponding to Figure 1 The "large model semantic analysis + layered emotion decision engine" in the figure above is used to obtain the robot's first code. The layered emotion decision engine refers to a layered emotion calculation strategy used in the calculation of the first code. It is specifically divided into: the first layer of basic emotion calculation, the second layer of composite emotion calculation, and the third layer of refined label calculation. Large model semantic analysis is the core bridge connecting multimodal sensor data with layered emotion calculation. The logic of its integration with layered emotion calculation is as follows:

[0042] (1) The role of large-scale semantic analysis: "Interpreting the meaning" of multimodal sensor data, so that robots are no longer limited to interacting with physical data such as touch and temperature, and can perform semantic analysis of external language. Large-scale semantic analysis does not directly participate in the numerical calculation of emotional intensity, but rather interprets the multimodal sensor data at the semantic level to obtain context labels, providing "scene anchors" and "feature weight basis" for subsequent emotional layering calculations. For example:

[0043] a. Semantic extraction and analysis of text / voice data: If multimodal sensor data includes user voice input (e.g., "You just bumped into me, that really saddens me") or text interaction information, the large-scale semantic analysis model parses the sentence to identify sentiment-prone words (e.g., "sad"), event descriptions (e.g., "bumped into me"), and object relationships (e.g., "you," "me"), thereby identifying the core conflict in the interaction scenario (e.g., "unexpected collision triggers negative emotions").

[0044] b. Semantic association analysis of non-verbal data: For non-verbal sensory data such as touch and vision (such as the force of a user stroking a robot arm or an image of a frowning face), semantic analysis is combined with historical interaction data to determine the semantic intent of the action (e.g., "stroking" corresponds to "friendly comfort" and "pushing" corresponds to "resistance or anger").

[0045] c. Cross-modal semantic fusion: Fuse the semantic information of multimodal data such as voice, text, touch, and vision to generate a unified "context description label" (e.g., "the user was slightly bumped in a high-decibel quarrel environment, and the tone was tearful").

[0046] (2) Interaction logic with sentiment layered computing: The large model semantic analysis interprets the multimodal sensor data in a contextualized semantic sense, obtains context labels, and guides the "precision calibration" of each layer of sentiment layered computing based on the context labels. The impact on sentiment layered computing is as follows:

[0047] a. Impact on the first-tier basic emotion calculation: Large-scale model semantic analysis modifies the basic emotion data and the coefficients in the emotion intensity calculation formula (e.g., 0.7 and 0.3 in the formulas for "Basic emotion data is sadness" and "Calculation process for sadness emotion intensity" below). For example, if the context label determines the current scenario is "user recounting a sad past (voice with a tearful tone + text mentioning 'missing deceased loved ones')," the "voice_tremor" coefficient will be increased (e.g., from 0.3 to 0.5) when calculating the emotion intensity of "sadness," reducing the weight of other irrelevant sensor data to make the calculation of basic emotion data more relevant to the context.

[0048] b. Impact on the calculation of second-layer composite emotions: Large-model semantic analysis limits the range of composite emotions generated. For example, when the Plutchik algorithm uses basic emotion data and its intensity to determine "60% sadness + 30% anger," if the context label is "user feels aggrieved due to being misunderstood (not attacked)," the composite emotion "aggrievance" will be prioritized. If the label is "user feels angry due to being maliciously harmed," the composite emotion "resentment" may be matched (a corresponding composition formula must be pre-set), thus preventing the composite emotion from being out of context.

[0049] c. Impact on the calculation of third-layer refined labels: Semantic analysis of the large model directly determines the subdivision dimensions of the multidimensional sentiment model (hereinafter referred to as "multidimensional labels"). For example, for the same complex emotion of "grievance," if the context label indicates that "the user mentioned 'past regrets'," the refined label will be inclined towards "nostalgic grievance." If it indicates that "the user complained about 'the failure of the current plan,'" the refined label will be inclined towards "loss-type grievance." This ensures that the final refined label is consistent with the user's true emotional motivation.

[0050] Using large-model semantic analysis as the "navigation system" for emotion layering computation has the advantage of not replacing the numerical logic of emotion layering computation but rather providing scenario-adaptive guidance for emotion layering computation by interpreting the "semantic core" of multimodal sensor data. This allows robots to not only "perceive the intensity of emotion" but also "understand the causes of emotion." Ultimately, the generation of the first encoding is consistent with both the physiological characteristics of emotion (such as tearing patterns) and the semantic logic of the interaction scenario (such as "nostalgic grievance" rather than simply "grievance"). As described above, the emotion layering decision engine module completes the "data perception → semantic understanding → emotion computation" closed loop, resolving the "emotional misjudgment" that can result from relying solely on numerical computation of sensor data (for example, semantic analysis can distinguish between "tears of gratitude" and "sobs of grievance").

[0051] The above is an example to illustrate the impact of large model semantic analysis. The specific process of obtaining the first code is as follows:

[0052] The sentiment layered decision engine is mainly divided into three parts: the first layer of basic sentiment calculation, the second layer of compound sentiment calculation, and the third layer of refined label calculation.

[0053] The first layer of basic emotion calculation includes: determining the contextual labels of multimodal sensor data based on large-scale semantic analysis, and using the Ekman algorithm to determine the robot's N basic emotion data and the emotion intensity corresponding to each basic emotion data based on the multimodal sensor data. Ekman mainly refers to the emotion classification theory proposed by psychologist Paul Ekman, which classifies basic human emotions into six cross-culturally common emotions. In this embodiment, the six basic emotions (joy, sadness, anger, fear, surprise, and disgust) of the Ekman algorithm are used for calculation. Taking sadness as an example, the basic emotion data is sadness and the calculation process of the emotion intensity of sadness is as follows:

[0054] sadness_score = 0.7×eyebrow_raise + 0.3×voice_tremor;

[0055] Here, sadness_score represents the emotional intensity of sadness, eyebrow_raise and voice_tremor represent different sensor data in the multimodal sensor data, and 0.7 and 0.3 represent the corresponding coefficients (determined by the contextual labels in the large-scale model's semantic analysis) for each sensor data in the multimodal sensor data. It should be noted that this is just an example. When the emotional intensity is non-zero, it indicates that the corresponding basic emotional data exists, meaning that the robot possesses that basic emotion. Furthermore, the main structure of this formula is the sum of (different sensor data × coefficient corresponding to different sensor data). Among the six basic emotions, different basic emotions are determined by different sensor data and their coefficients. The relationship between basic emotions, sensor data, and coefficients can be preset or set empirically. For example, sadness is determined by the eyebrow_raise and voice_tremor sensor data and their coefficients, while joy is determined by the sum of the products of at least one other sensor data and its coefficient.

[0056] Therefore, after multiplying and summing the multimodal sensor data in the above manner, N basic emotion data of the robot and their corresponding emotion intensities will be obtained, which means that the robot has N basic emotions, and each basic emotion has a different emotion intensity.

[0057] The second layer of composite emotion calculation includes: using the Plutchik algorithm to determine the robot's composite emotion data based on N basic emotion data and their corresponding emotion intensities. Table 1 shows an example table of composite labels generated based on the Plutchik emotion wheel:

[0058] Table 1

[0059] Complex emotions Composition formula Tear Mode Wronged Sadness 60% + Angry 30% + Surprise 10% Intermittent sobbing tears gratitude Joy 50% + Trust 30% + Surprise 20% Continuous mild tearing

[0060] As shown in Table 1, the six basic emotions are combined (corresponding to the "Composition Formula" column in Table 1) to determine compound emotion data (corresponding to the "Compound Emotion" column in Table 1). The percentage in the composition formula can be derived from the emotion intensity calculation in the first-level basic emotion calculation. The calculation method can be: Percentage of the current basic emotion = Emotion intensity corresponding to the current basic emotion data / (Sum of the emotion intensities of all basic emotion data). The Plutchik emotion wheel is an existing technology, and the basic implementation principles of the Plutchik emotion wheel (i.e., the Plutchik algorithm) will not be discussed here. It should be noted that the basic emotional components of different compound emotions may be the same. Therefore, the context labels from the semantic analysis of the large model are combined to match compound emotions within the context label threshold. For example, if the context label is "The user feels aggrieved due to being misunderstood (not attacked)", the compound emotion "grievance" is preferentially matched. If the label is "The user feels angry due to being maliciously harmed", the compound emotion "resentment" may be matched to avoid compound emotions being out of context.

[0061] The third-level refined label calculation includes: determining the refined emotion label corresponding to the complex emotion data according to the preset multi-dimensional emotion model, and determining the first code of the robot according to the refined emotion label.

[0062] The preset multidimensional sentiment model in this embodiment is specifically a multidimensional label determined based on semantic analysis of the large model. The multidimensional label may include a time dimension, an event association dimension, a degree difference dimension, an object orientation dimension, etc. The following introduces each dimension of the multidimensional label by way of example.

[0063] (1) In terms of the time dimension, taking "nostalgic grievance" as an example, it emphasizes the grievance caused by the retrospection of past events in time. For example, the semantic analysis of the large model determines that the user recalls the grievance caused by the experience of being misunderstood by friends. It can be further subdivided into recent nostalgic grievances (such as related experiences recalled within the past year) and long-term nostalgic grievances (such as experiences recalled many years ago) based on the time of the recalled event. This subdivision based on the time of the recall can help the robot better adjust the degree of emotional resonance when responding. For recent nostalgic grievances, the robot can use more immediate comforting language and interactive methods based on the preset multi-dimensional tags, mentioning recent similar experiences or friends around to narrow the emotional distance with the user; for long-term nostalgic grievances, the robot can use a more historical and soothing tone based on the preset multi-dimensional tags to guide the user to engage in deeper emotional expression and self-reflection. In terms of tearing, after determining "nostalgic grievances", the robot can trigger "slow tears" based on the preset multi-dimensional tags.

[0064] (2) In terms of event association, taking "loss-type grievance" as an example, it can be subdivided according to the type of event that triggers the sense of loss, such as career-type loss grievance (caused by work-related events such as failed job promotion and project failure) and life-type loss grievance (caused by life events such as emotional breakdown and conflicts with family members) determined by the semantic analysis of the large model. In this way, when the robot faces loss-type grievances associated with different events, it can call on knowledge and experience in different fields to respond. When it comes to career-type loss grievances, the robot can share workplace inspirational stories and provide career development advice based on the preset multi-dimensional tags; if it is a life-type loss grievance, the robot can provide emotional communication skills, family relationship management methods, etc. based on the preset multi-dimensional tags. In terms of tearing, after determining "loss-type grievance", the robot can trigger "pulse tears" based on the preset multi-dimensional tags.

[0065] (3) In terms of degree of difference, taking "grievance" as an example, based on the intensity of the grievance emotion obtained by the semantic analysis of the large model, it is divided into mild grievance, moderate grievance and severe grievance. When it is mild grievance, the robot will tease and divert the user's attention in a light tone according to the preset multi-dimensional labels; when it is moderate grievance, the robot will give more sincere comfort and empathy according to the preset multi-dimensional labels; when it is severe grievance, the robot will provide deeper psychological support according to the preset multi-dimensional labels, such as suggesting seeking professional psychological counseling or providing some relaxation methods. In terms of tearing, the robot can control the interval time or flow rate of tears according to the preset multi-dimensional labels.

[0066] (4) In terms of the object-directed dimension, taking "grievance" as an example, based on the semantic analysis of the large model, the grievance emotion is divided according to the object it is directed at, such as directed at others (such as being aggrieved by being excluded by colleagues), directed at the user themselves (blaming themselves for their own mistakes), and directed at the environment (such as being aggrieved due to unfair social environment). For grievances directed at others, the robot helps users analyze the motivations of others' behavior based on preset multi-dimensional labels and provides interpersonal communication strategies; for grievances directed at themselves, the robot encourages self-acceptance based on preset multi-dimensional labels and guides users to correctly view their own mistakes; for grievances directed at the environment, the robot discusses social phenomena with users based on preset multi-dimensional labels and provides positive coping perspectives.

[0067] It should be noted that the above is an example of a multidimensional label. The degree of dimensional refinement of the multidimensional label is determined by the semantic analysis of the large model, including but not limited to the aforementioned time dimension, event association dimension, degree difference dimension, and object orientation dimension. By refining the preset multidimensional label (i.e., the preset multidimensional emotional model) based on the contextual label of the emotional data combined with the semantic analysis of the large model, refined emotional labels such as "nostalgic grievance" and "loss-type grievance" can be obtained. Based on the refined emotional label ("nostalgic grievance" or "loss-type grievance"), the corresponding tear effect (i.e., the "slow tears" or "pulse tears" described above) can be obtained. In this process, the instruction code corresponding to the tear effect is the first code. In other words, the refined emotional label is a further refinement of the composite emotional data based on the multidimensional label. The first code is the instruction code for achieving the corresponding tear effect. The encoding information of the first code includes the basic emotional data, the composite emotional data, and the refined emotional label. It is understandable that when the composite emotional data is "grievance", its corresponding refined emotional label may be "nostalgic grievance" or "loss-type grievance", and for the preset multi-dimensional labels, different feedback from the robot can be obtained in different dimensions.

[0068] The benefits of doing this are:

[0069] Achieving More Accurate Emotion Recognition: Traditional emotion classification is relatively general. By segmenting emotions with multi-dimensional label examples, we can more meticulously distinguish complex human emotional states, enabling robots to more accurately understand user emotions. For example, if a user is simply judged as "aggrieved," the robot's response strategy will be relatively simple. However, by segmenting the grievance into different dimensions, the robot can, like a human, understand the causes and context of the user's grievance from multiple perspectives, making a more realistic assessment.

[0070] Providing personalized emotional interaction: Different users may have different needs due to individual experiences, personalities, and other factors, even when faced with the same complex emotions. Multi-dimensional tag segmentation can address these personalized needs. For example, an outgoing user experiencing a mild career-related setback or grievance may want a robot to resolve their emotions with humor; whereas an introverted user may prefer a robot that listens quietly, offers affirmation, and encouragement.

[0071] Enriching robots' emotional expressions and response strategies: Multi-dimensional labels provide robots with more diverse emotional processing pathways. When faced with complex emotional situations, robots are no longer limited to a limited number of responses. Instead, they semantically analyze combinations of labels across different dimensions and flexibly adjust emotional expression through language, tone, and body language (if the robot has the capabilities). For example, when addressing the profound loss and grievance caused by a relationship breakup, robots not only offer gentle, empathetic comfort but even simulate hugs to enhance emotional interaction.

[0072] Expanding robot application scenarios: In diverse fields such as healthcare, education, customer service, and companionship, multi-dimensional sentiment analysis and responses are more targeted. In medical and psychotherapy scenarios, this helps doctors gain a more comprehensive understanding of their patients' mental states; in education, teachers can adjust their teaching methods based on the robot's accurate assessment of students' emotions; in customer service, this improves customer satisfaction; and in companionship, it enhances the quality of companionship and user dependence.

[0073] On the other hand, the robot's environmental factors can be determined based on environmental factors, the robot's relationship factors can be determined based on relationship factors, and the robot's role factors can be determined based on role factors. The robot's emotion weight can be determined based on the environmental factors, relationship factors, and role factors. The emotion weight is used as the weight value of the first code to modify the emotion code.

[0074] The emotional weight of the robot is determined based on environmental factors, relationship factors, and role factors as follows:

[0075] W_adj = α×W_env + β×W_relation + γ×W_role;

[0076] Among them, W_adj is the sentiment weight, W_env is the environmental factor, α is the environmental factor coefficient, W_relation is the relationship factor, β is the relationship factor coefficient, W_role is the role factor, and γ is the role factor coefficient.

[0077] The purpose of modifying the first encoding is to dynamically adjust the emotional weights of different robots in different scenarios, making the robot's tear perception more humane and accurate. For example, in family scenarios, visual perception is emphasized, so the visual weight is adjusted; in intimate relationships, excessive tears need to be reduced, so the sadness intensity is further restricted; for medical robots, the trigger threshold for fear perception needs to be increased to avoid startling patients. Table 2 shows some examples of environments, relationships, and roles in this embodiment, as well as examples of their calculation methods. It should be noted that Table 2 in this embodiment is only an example, and the coefficients can be adjusted according to actual scenarios and usage needs.

[0078] Table 2

[0079] Correction Factor Calculation method Application scenario examples Environmental factor W_env Home scene: visual weight +0.2 Home companion robot Relation factor W_relation Intimacy: Sadness intensity × 0.8 Reduce excessive tearing Role factor W_role Medical Robot: Fear trigger threshold +30% Avoid startling the patient

[0080] Determine the second code (corresponding to the second code) used by the robot to judge the physical stimulus response based on the multimodal sensing data Figure 1 The “physical stimulus decision device” in the

[0081] According to the physical sensing data in the multimodal sensing data, when the physical sensing data meets the preset conditions, the hardware interrupt of the robot is triggered based on the physical sensing data to generate the second code of the robot.

[0082] The physical sensing data in the multimodal sensing data used include the tactile pressure gradient value measured by the pressure gradient calculation circuit (tactile pressure sensor matrix) and the sound decibel value measured by the sound amplitude detection circuit (sound sensor).

[0083] When the tactile pressure gradient value and the sound decibel value meet certain preset conditions, a hardware interrupt can be triggered to generate a second code. In this embodiment, the hardware interrupt is triggered to generate the second code based on the tactile pressure gradient value > 5kPa / ms ∩ sound > 90dB.

[0084] The purpose of using physical stimulus response is to establish a rapid response mechanism for the robot to sudden strong physical stimuli by monitoring two key physical stimulus signals: tactile pressure gradient and sound decibel level. When a strong physical stimulus appears in the external environment that may affect the robot itself or the object it interacts with, the physical stimulus judgment module can determine in real time whether the stimulus intensity has reached a preset threshold. If the condition is met, it will immediately trigger a hardware interrupt and generate a second code to drive the robot to execute the emergency response, thus enabling the robot to instantly perceive and respond to sudden physical stimuli.

[0085] This approach offers the following benefits: Improved response speed and safety: Hardware interrupts are more direct and efficient than software-based logical judgment. It also addresses the lack of immediacy in the emotional decision-making system: The emotional layered decision engine focuses on the refined processing of complex emotions, resulting in a relatively complex response process; while the physical stimulus judgment module focuses on rapid response to sudden physical stimuli. These two complement each other, allowing the robot to maintain delicate emotional interaction while also exhibiting biologically-like "instinctive reactions" to extreme physical stimuli, improving overall interaction reliability.

[0086] S2, determine the emotion code of the robot according to the first code and / or the second code, determine the tear instruction for controlling the operation of the microfluidic tear execution device of the robot according to the emotion code, and make the robot cry. Particularly, the emotion code of the robot is determined according to the first code and the second code, the tear instruction of the robot is determined according to the emotion code, and the microfluidic tear execution device of the robot is driven to operate according to the tear instruction to make the robot cry. Specifically, the dual-path response mechanism of emotional state + physical stimulation provided by the present invention generates a first code and a second code, combines the first code with the weight value (calculated in the above "correction of the first code") and fuses it with the second code to obtain a dynamic emotion code, and generates microfluidic control parameters (i.e., "tear instructions") (corresponding to) through analysis by the microfluidic parameter dynamic mapping algorithm. Figure 1The "microfluidic parameter mapper" in the robot controls the microfluidic tearing execution device according to the tearing instruction, making the robot cry (corresponding to the tearing instruction). Figure 1 The "drive module controls the mechanical actuator" in the , to achieve a realistic tearing effect that matches the human emotional state.

[0087] The rules for determining the robot's emotion code based on the first code and / or the second code follow the logic of "priority + dynamic weight adjustment". The details are as follows:

[0088] (1) Priority: The second code has the highest priority and is used to show a biological-like "instinctive reaction" when facing extreme physical stimuli. When the second code is generated (that is, when a strong physical stimulus that triggers a hardware interrupt occurs), regardless of whether the first code is in the generation or execution state, the system will immediately suspend the emotional expression process corresponding to the first code and give priority to responding to the second code. In other words, when the second code exists, the robot's emotional code is determined only by the second code. For example, if the robot is executing "continuous mild tears" (such as gratitude) based on the first code, and at this time it is subjected to a strong physical stimulus such as "a slap" that causes the second code to be generated, the robot will immediately stop crying and switch to the emergency response corresponding to the second code (such as closing the eyes momentarily, tilting the head back, and other defensive postures, while pausing tears).

[0089] (2) Dynamic weight adjustment: When the physical stimulus corresponding to the second code disappears (i.e., the tactile pressure gradient value is ≤5kPa / ms and the sound is ≤90dB, and the hardware interrupt is released), the robot will resume the execution of the first code based on the weight value calculated by "correcting the first code". The weight value at this time will be dynamically adjusted in combination with the duration and intensity of the physical stimulus: if the stimulus is short and mild, the weight of the first code remains unchanged, and the tearing instruction before the interruption will continue to be executed (such as resuming mild tearing); if the stimulus is strong and lasts for a long time, the weight value will be temporarily reduced (such as from 1.0 to 0.6), so that the intensity of the emotional expression corresponding to the first code is weakened (such as the speed of tearing slows down), until it is determined that the environment has returned to stability, and the weight value will gradually return to normal level.

[0090] (3) Fusion rules for special scenarios: If the first and second codes are generated consecutively in a very short period of time (e.g., when a strong physical stimulus occurs, the robot is in a complex emotional state such as "grievance" that may be accompanied by tears), the robot will generate a fusion instruction based on the remaining first code information and the current environmental state after the second code is executed. For example, after the physical stimulus ends, the robot may first execute "rapid blinking" (a residual reaction of the second code), and then switch to "intermittent sobbing tears" corresponding to "grievance", but the intensity of the tears is slightly reduced due to the influence of the previous stimulus, achieving a smooth transition between the two codes.

[0091] The first and second codes are fused together using this fusion mechanism to create a dynamic emotional code. This mechanism ensures emergency response priority under strong physical stimuli while avoiding abrupt interruptions in emotional expression, allowing the robot to maintain the coherence and rationality of emotional interaction while responding to emergencies.

[0092] The dynamic mapping algorithm for microfluidic parameters is used to convert the emotional code formed by the fusion of the first and second codes into the microfluidic control parameters (i.e., "tear instructions") required for the robot to shed tears, thereby achieving a realistic tearing effect that matches the human emotional state. The specific steps are as follows:

[0093] (1) Establishing a library of emotion-fluid parameter mapping relationships: Through a large number of experiments and data statistical analysis, a library of corresponding relationships between different emotion codes and microfluidic control parameters (such as the flow rate in Table 3 below) is established. For example, for different intensities of "joy" emotion codes, the corresponding parameters such as the output flow rate of the liquid storage device, the pressure distribution within the microchannel group, the operating frequency of the micro-pressure pump, and the opening and closing degree of the tear outlet are recorded; the same is true for other emotion codes such as "sadness" and "grievance". These parameter combinations can accurately match the characteristics of human tears in different emotional states, such as tear speed, teardrop size, and tear duration.

[0094] (2) Normalization: Normalize the emotion codes so that the code values ​​are within a uniform standard range suitable for algorithm processing, eliminating the impact of data magnitude differences between different codes. Normalization can be done using existing processing methods. In this embodiment, all types of data in the first and second codes (such as basic emotion data, composite emotion data, and physical stimulus-related data) are converted to the [0, 1] interval to facilitate subsequent calculations.

[0095] Table 3 shows some examples of the mapping relationship between emotion types and flow rates in control parameters:

[0096] Table 3

[0097] Basic emotion types 27-dimensional label example Flow rate (μL / min) sad nostalgia 12±3 anger moral outrage 22±8

[0098] For the "pulse tears" mentioned in the third-level refined label calculation, a pulse control optimization logic is set in this embodiment. When the robot is "pulse tears", the time interval of the "pulse tears" is controlled according to the change of the emotion intensity corresponding to the basic emotion data. The stronger the emotion intensity, the shorter the time interval of the "pulse tears".

[0099] Regarding the control of microfluidic control parameters, this embodiment also provides the following control methods:

[0100] (1) Dynamic correction and feedback: During the robot's tearing action, the actual microfluidic (i.e., tearing) state is continuously monitored (e.g., real-time data is obtained through flow sensors and pressure sensors) and compared with the expected microfluidic control parameters. If there is a deviation, the microfluidic parameters are dynamically corrected based on the size and direction of the deviation using common feedback control algorithms such as the PID (proportional-integral-differential) control algorithm to ensure that the robot's tearing effect is always highly consistent with the emotional state represented by the emotional code, achieving a stable and realistic tearing performance.

[0101] (2) Weight calculation and parameter adjustment: Combined with the weight value obtained by "correcting the first code", the normalized emotional code is weighted. If in a certain situation, the weight of the physical stimulus factor (second code) is 0.6 after context correction, and the weight of the first code generated by the emotional layered decision engine is 0.4, then when calculating the microfluidic control parameters, the information contained in the two is fused according to this weight. Then, based on the weighted result, the pre-established emotion-fluid parameter mapping relationship library is queried and the parameters are adjusted. For example, if the calculation result is close to the range corresponding to the "moderate sadness" emotional code, then based on the basic microfluidic parameters of the emotion, the parameters such as the output flow of the liquid storage device are fine-tuned according to the weight, so that the tearing effect is more in line with the current fused emotional state.

[0102] In addition, this embodiment also incorporates a bionic tear duct recirculation system to recover dynamic tear fluid at the end of tear secretion, preventing residual tears from contaminating the robot's electronic components. This system utilizes a micro-negative pressure generator to generate a controllable negative pressure gradient (-15 to -25 mmHg), combined with the capillary effect of hydrophobic-coated microchannels, to achieve dynamic fluid recovery at the end of tear secretion. The operating cycle of the bionic tear duct recirculation module is synchronized with the emotional decay curve.

[0103] Because the tearing process is triggered by a dual-path response mechanism (emotional state + physical stimulation), the emotion decay curve also has a dual recovery mode: gradual recovery during the emotion decay period and emergency recovery during physical stimulation. Gradual recovery during the emotion decay period refers to when the basic emotion data is calm and the robot's tear volume is less than a preset value, the robot is considered emotionally stable and no longer crying. Emergency recovery during physical stimulation refers to the emergency recovery generated by the second code generated by the hardware interrupt triggered by the physical stimulation judgment module. The specific pseudo code example is as follows:

[0104] if (emotion code == "calm" && tear sensor > 0.1μL (default value))

[0105] Negative pressure intensity = -20×exp(-0.5×(t-t_end)); / / Exponential decay negative pressure curve

[0106] Start the hydrophobic microchannel (contact angle > 120°);

[0107] elseif (physical pass-through interrupt flag)

[0108] Forced negative pressure = -25; / / Physical path emergency stop mode

[0109] end.

[0110] Among them, exp is an exponential function operation; the negative pressure gradient (-15~-25mmHg) is the preset working range of the robot's bionic tear duct (i.e., tear channel) respiration module, which defines the value range of the negative pressure intensity; the negative pressure intensity is the execution value of the specific recovery stage and needs to be dynamically adjusted within the negative pressure gradient range (for example, in the gradual recovery during the emotional subsidence period, the calculation result of -20×exp(...) is always in the range of -15~-25mmHg); the forced negative pressure is a specific value for emergency recovery of physical stimulation, and is the upper limit of the negative pressure gradient. It is an extreme application of negative pressure intensity in emergency scenarios (in this embodiment, the forced negative pressure is set to -25, and the gradient maximum value is directly taken to achieve rapid recovery).

[0111] The present invention is further described below in conjunction with two other embodiments.

[0112] Example 2: Nostalgic Sadness and Tears (Multi-dimensional Label “Nostalgia”):

[0113] Emotion recognition is as follows: visual sensor: staring at old photos for >5 seconds + drooping mouth corners; sound sensor: sighing frequency >3 times / minute; the basic emotional data is judged to be sadness, and the corresponding emotional intensity is 72.

[0114] Emotional layering decision engine module: Plutchik algorithm composite analysis: sadness 80% + trust 20%; composite emotion data is determined to be sentimental. Combined with multi-dimensional label mapping, it is mapped to the nostalgia label.

[0115] Modify the first coding: if the environment is judged to be a home scene, set the coefficient to sadness intensity × 0.9 (reduce the amount of tears); if the role relationship is judged to be the owner, set the coefficient to pulse interval + 0.2 seconds (extend the pause).

[0116] In summary, the execution parameters of the robot in this embodiment are shown in Table 4:

[0117] Table 4

[0118] parameter value Basic flow rate 15 μL / min Number of pulses 2 times (nostalgic feature) Total duration 10 seconds

[0119] Example 3: Tears of Moral Indignation

[0120] Trigger conditions: Large-scale model semantic analysis identifies the keyword "unfairness" + sound amplitude > 85dB;

[0121] Emotional hierarchical decision engine module: Basic emotion: anger (intensity 90); Multidimensional label: moral indignation.

[0122] Modify the first code: the robot role is a public service worker, set the coefficient: the flow rate upper limit is set to 25 μL / min (to inhibit overexpression).

[0123] The robot's biological simulation effects are as follows:

[0124] [High-speed tearing 22 μL / min × 3 seconds] → Pause 0.5 seconds → [Second burst of tearing].

[0125] The present invention provides a microfluidic tearing actuator for a robot, and the specific technical solution is as follows, including:

[0126] A fluid storage device, a tear channel, a tear outlet, and a micro pressure pump;

[0127] The reservoir is used to store tear fluid;

[0128] One end of the tear channel is connected to the liquid storage device, and the other end of the tear channel is connected to the tear outlet;

[0129] The tear outlet is set at a preset position on the robot's eye to drain the tears;

[0130] The micro pressure pump is arranged on the tear channel and is used for pumping the tear liquid in the liquid storage device to the tear outlet through the tear channel.

[0131] Based on the above solution, the present invention can also be improved as follows.

[0132] Furthermore, the liquid storage device is a detachable tear capsule; the tear outlet is provided with a bionic lacrimal gland micropore array.

[0133] The fluid storage device can be a biocompatible fluid storage tank or a replaceable fluid storage bag, which can be loaded with physiological saline or artificial tears. The reserved replenishment port of the fluid storage tank supports rapid addition of tears and rapid replacement of tears, and the fluid storage bag can also be directly replaced.

[0134] A bionic lacrimal gland micropore array refers to at least one liquid drainage hole located at the tear outlet. The liquid drainage hole is located on the surface of the bionic robot's simulated eye skin. In this embodiment, the preset location can be the bionic skin surface at the lower edge of the robot's eyelid. The bionic lacrimal gland micropore array can be laser-machined into one or more liquid drainage holes on the bionic skin surface at the lower edge of the robot's eyelid. The arrangement and shape of the liquid drainage holes are not restricted; all liquid drainage holes constitute the bionic lacrimal gland micropore array.

[0135] The present invention provides a robot tearing system, comprising: a microfluidic tearing execution device for a robot as described above and a processor, wherein the processor is used to execute a robot tearing control method as described above.

[0136] The beneficial effects of the present invention are as follows:

[0137] Based on multimodal sensory data, a first code for determining emotional state and a second code for determining physical stimulus response are determined. This innovative dual-path triggering mechanism, based on both emotional state and physical stimulation, generates emotional codes. Combined decision-making using the first and / or second codes enables the robot to simulate physiological feedback from both emotional state and physical stimulation. Furthermore, by incorporating the tear flow mechanism of a microfluidic hardware device, the robot achieves precise tear control based on the emotional perception of physiological feedback.

[0138] Although the embodiments of the present invention have been shown and described above, it will be understood that the above embodiments are illustrative and are not to be construed as limitations on the present invention. A person skilled in the art may change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. A robot tearing control method, characterized in that: include: S1, acquiring multimodal sensor data of a robot, and determining a first code for determining an emotional state and a second code for determining a physical stimulus response of the robot based on the multimodal sensor data; S2, determining an emotion code of the robot according to the first code and / or the second code, and determining a tearing instruction for controlling a microfluidic tear-making execution device of the robot according to the emotion code, so as to make the robot shed tears; The first code for determining the emotional state of the robot according to the multimodal sensing data is specifically: Determine the context labels of the multimodal sensor data according to the semantic analysis of the large model, and determine N basic emotion data of the robot and the emotion intensity corresponding to each basic emotion data according to the multimodal sensor data and the context labels through the Ekman algorithm; Determine the robot's composite emotion data based on the context label, N basic emotion data and their corresponding emotion intensities using the Plutchik algorithm; Based on the context label, the refined emotion label corresponding to the complex emotion data is determined according to a preset multidimensional emotion model, and the first code of the robot is determined according to the refined emotion label.

2. A robot tearing control method according to claim 1, characterized in that: The multimodal sensing data includes: At least one of visual sensor data, sound sensor data, tactile sensor data, pressure sensor data, temperature sensor data, and smoke sensor data.

3. A robot tear control method according to claim 1, characterized in that: The second code for the robot to judge the physical stimulus response is determined according to the multimodal sensing data: According to the physical sensing data in the multimodal sensing data, when the physical sensing data meets a preset condition, a hardware interrupt of the robot is triggered based on the physical sensing data to generate a second code of the robot.

4. A robot tearing control method according to claim 1, characterized in that: Also includes: The emotion weight of the robot is determined according to the environmental factors, relationship factors and role factors of the robot, the emotion weight is used as the weight value of the first code, and the emotion code is corrected.

5. The robot tearing control method according to claim 1, characterized in that: The emotion weight of the robot is determined based on the robot's environmental factors, relationship factors, and role factors as follows: W_adj = α×W_env + β×W_relation + γ×W_role; Among them, W_adj is the sentiment weight, W_env is the environmental factor, α is the environmental factor coefficient, W_relation is the relationship factor, β is the relationship factor coefficient, W_role is the role factor, and γ is the role factor coefficient.

6. A microfluidic tearing actuator for a robot, characterized in that: include: A liquid storage device, a tear channel, a tear outlet, a micro pressure pump, and a processor, wherein the processor is used to execute a robot tear control method according to any one of claims 1 to 5; The liquid storage device is used to store tear fluid; One end of the tear channel is connected to the liquid storage device, and the other end of the tear channel is connected to the tear outlet; The tear outlet is provided at a preset position of the robot's eye for draining tears; The micro pressure pump is arranged on the tear channel and is used to pump the tear liquid in the liquid storage device to the tear outlet through the tear channel.

7. The microfluidic tearing actuator for a robot according to claim 6, characterized in that: The tear outlet is provided with a bionic lacrimal gland micropore array.

Images