An intelligent companion robot, control method, and medium

Abstract

The application discloses an intelligent accompanying robot, a control method and a medium. The robot comprises a shell (101), the shell (101) is covered with a bionic fur layer (102) and is provided with a window (103); a main control chip; an auditory perception module comprising a micro-electro-mechanical system microphone distributed in a ring shape; an infrared positioning module comprising an infrared sensor arranged in a head region; a tactile perception module comprising a tactile sensor array embedded between the bionic fur layer (102) and the shell (101); a motion module and an emotion display module. The main control chip extracts a sound source azimuth based on an audio time difference, controls the head region to turn and then performs accurate ranging and positioning through the infrared sensor, extracts a tactile feature vector from the tactile sensor array, generates a control instruction by fusing multi-channel data and multi-modalities, drives the motion module and updates the emotion display module; and a suspended force guiding framework is internally arranged. The application can realize sensor protection and stress distortion reduction, and simultaneously improve positioning accuracy and naturalness of multi-modal interaction.

Description

Technical Field

[0001] This application is mainly applied in the smart hardware industry to provide intelligent companion robots with dual-channel interaction capabilities and realistic touch sensation. Background Technology

[0002] With the increasing prevalence of artificial intelligence technology, intelligent companion robots are gradually entering home settings to provide daily interaction. In practical applications, tactile and auditory interaction capabilities are fundamental elements for improving the naturalness of multimodal interactions. However, current commercially available general-purpose technical solutions still have engineering limitations.

[0003] Existing robot shells are generally made of hard plastic or sheet metal, resulting in high surface hardness and an inability to provide realistic temperature and tactile feedback. While some solutions utilize carbon fiber frames combined with fully enclosed silicone bionic skin, this structure is complex to manufacture, difficult to assemble, and expensive, making mass production difficult. Traditional interaction methods heavily rely on visual image acquisition or single-channel voice wake-up commands, lacking a tactile perception channel. When receiving physical contact, the robot cannot sense external mechanical stress, leading to a lack of tactile response and gaps in multimodal interaction. Robots' emotional expression often relies solely on fixed-format speech output from speakers, lacking a visual facial expression feedback mechanism. Conventional listening devices are mostly single-microphone structures, preventing robots from actively perceiving the sound location of their surroundings and responding with physical turning in standby mode. This results in a passive interaction initiation process, hindering proactive posture responses. Single acoustic positioning methods have inherent errors in near-field ranging and angle accuracy, making it difficult to meet the positioning requirements for precise close-range interaction. The hardware architecture and underlying control logic of existing products are insufficient to meet multimodal interaction needs.

[0004] This application provides a robot architecture that integrates multimodal perception, is structurally stable, and has controllable cost to solve the aforementioned technical problems. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent companion robot, control method, and medium to solve the technical problems of existing companion robots, such as stiff touch, lack of tactile perception channels, limited emotional expression, insufficient positioning accuracy, and passive interactive response.

[0006] In a first aspect, the present invention provides an intelligent companion robot, comprising: a shell covered with a biomimetic fur layer, the shell including a head region, a back region, and a wing region, the head region and back region of the shell having windows; a main control chip; an auditory perception module including four microelectromechanical system (MEMS) microphones arranged in a ring within the head region; a tactile perception module including a tactile sensor array distributed at positions corresponding to the windows and embedded between the inner side of the biomimetic fur layer and the shell; an infrared positioning module including an infrared sensor disposed in the head region for precise distance measurement and positioning of a target; and a motion module including a first micromotor and a second micromotor. A micro motor is disposed in the head region, and a second micro motor is disposed in the wing region; an emotion display module, including an LCD screen; the auditory perception module, the tactile perception module, the infrared positioning module, the motion module, and the emotion display module are electrically connected to the main control chip; the main control chip is configured to extract the azimuth angle of the sound source based on the audio time difference collected by the microphones of each microelectromechanical system, control the motion module to drive the head region to turn to the direction corresponding to the azimuth angle of the sound source, and control the infrared sensor to perform precise ranging and positioning to obtain the ranging result, extract tactile feature vectors containing force and duration from the tactile sensor array, and fuse the azimuth angle of the sound source, the ranging result, and the tactile feature vectors to generate control commands; the control commands are used to drive the motion module and update the visualization interface of the emotion display module.

[0007] Furthermore, a suspended force guide frame is disposed between the biomimetic fur layer and the outer shell; the suspended force guide frame includes a polyester fiber flocking layer, a silicone substrate, an integrally molded rigid force guide column, a silicone cantilever damping layer and a rigid base frame stacked sequentially from the outside to the inside, and the touch sensor array is attached to the outer surface of the rigid base frame.

[0008] Furthermore, the touch sensor array includes a force-sensitive thin-film sensing unit; the lower end of the integrally molded rigid guide column is aligned with the sensitive center of the force-sensitive thin-film sensing unit along the same vertical normal; in the unpressurized state, a first gap distance of 0.5 mm is provided between the lower end of the integrally molded rigid guide column and the outer surface of the rigid substrate; when the suspended guide frame is subjected to an external force and the external force causes the deformation of the silicone cantilever damping layer to exceed the first gap distance, the bottom surface of the silicone substrate directly contacts the outer surface of the rigid substrate, forming a mechanically rigid support point and blocking further penetration of deformation stress into the touch sensor array.

[0009] Furthermore, the suspended force-guiding skeleton also includes a polyimide flexible electrothermal film embedded inside the silicone substrate, the polyimide flexible electrothermal film being electrically connected to the main control chip; the main control chip is configured to output a pulse width modulation heating signal to the polyimide flexible electrothermal film; the silicone substrate is configured to receive the thermal energy emitted by the polyimide flexible electrothermal film and undergo softening deformation.

[0010] Furthermore, the intelligent companion robot also includes a torso; the outer shell is provided with a neck rotator at the connection between the head region and the torso, and a hinge joint at the connection between the wing region and the torso; the bionic fur layer is inlaid with a spandex high-elasticity base strip in the peripheral areas corresponding to the neck rotator and the hinge joint, the lateral elongation of the spandex high-elasticity base strip being greater than 150%, configured to provide deformation compensation during joint operation.

[0011] Secondly, the present invention provides a control method for an intelligent companion robot, applied to an intelligent companion robot. The intelligent companion robot includes a main control chip and an auditory perception module, a tactile perception module, an infrared positioning module, a motion module, and an emotion display module electrically connected to the main control chip. The emotion display module includes a liquid crystal display screen. The control method includes: acquiring environmental audio signals through the auditory perception module and extracting the azimuth angle of the sound source based on the audio time difference; controlling the motion module to drive the head area to turn to the corresponding direction according to the azimuth angle of the sound source, and controlling the infrared positioning module to perform precise ranging and positioning to obtain the ranging result; extracting tactile feature vectors containing force and duration through the tactile perception module; fusing the azimuth angle of the sound source, the ranging result, and the tactile feature vectors to generate control commands; sending the control commands to the motion module and the emotion display module, driving the motion module and updating the visualization interface of the emotion display module through the control commands.

[0012] Furthermore, the auditory perception module includes four microelectromechanical system (MEMS) microphones arranged in a circular array; the extraction of the sound source azimuth angle based on audio time difference includes: obtaining the arrival time difference of the environmental audio signals collected by the four MEMS microphones; constructing a nonlinear positioning equation set according to the preset physical coordinates of the microphone array and the sound propagation speed constant; solving the nonlinear positioning equation set to obtain the initial three-dimensional coordinate point of the target sound source; mapping the initial three-dimensional coordinate point to a polar coordinate system centered on the intelligent companion robot, and extracting the horizontal plane deflection angle as the sound source azimuth angle; the control method further includes: controlling the infrared sensor in the infrared positioning module to accurately measure the distance of the target and obtaining the distance measurement result of the target; weighting and fusing the sound source azimuth angle with the accurate distance and the accurate angle according to weighting coefficients to generate a fused positioning result; wherein, the weighting coefficients are dynamically adjusted according to the environmental signal-to-noise ratio, and the higher the environmental signal-to-noise ratio, the greater the weight of the sound source azimuth angle.

[0013] Furthermore, the intelligent companion robot also includes a torso, and a flexible polyimide electrothermal film is embedded inside the intelligent companion robot; the motion module includes a first micro motor correspondingly disposed in the head area and a second micro motor disposed in the wing area; the control method further includes hardware cascaded stress processing, which includes: continuously monitoring the high-frequency energy value of the environmental audio signal; when the high-frequency energy value exceeds a first energy threshold, triggering abnormal high-frequency noise interruption processing logic; cutting off the power supply to the flexible polyimide electrothermal film; generating a risk avoidance drive command and sending it to the motion module, driving the first micro motor to deflect according to a first angular velocity; refreshing the display screen of the emotion display module to a preset alert interface, and driving the second micro motor to perform a contraction and closure action.

[0014] Furthermore, the first angular velocity is configured to be three times the head deflection angular velocity in the normal tracking state; the head deflection angular velocity in the normal tracking state is a following angular velocity calculated in real time based on the extracted azimuth angle of the sound source; the driving the second micro motor to perform the contraction and closure action includes generating a closure drive level and continuously sending the closure drive level to the second micro motor, driving the wings to deflect inward until they physically contact the corresponding position on the outside of the torso to form a closed defensive posture.

[0015] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the control method described in this application.

[0016] The intelligent companion robot and its control method provided in this application, on the one hand, have windows set in the back and head of the shell, and embed a tactile sensor array between the fur layer and the shell. This creates a high-density tactile receiving network while ensuring a compact physical size of the robot, allowing users to directly touch the soft fur and effectively capture pressure. On the other hand, the microelectromechanical system (MEMS) microphone array introduced into the auditory channel can extract the time difference of sound wave arrival in real time, locate the three-dimensional position of the sound source, and perform precise ranging and positioning through infrared sensors after the head region turns. This achieves a combination of far-field acoustic coarse positioning and near-field infrared fine positioning. Combined with the internally extracted contact force characteristics, multi-channel data is deeply coupled within the main control chip to generate control commands that link actions and facial expressions. This structure overcomes the interaction blind spots of traditional companion robots, achieving sound localization, precise positioning, and realistic physical touch feedback at a lower hardware cost. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the overall structure of the intelligent companion robot provided in an embodiment of the present invention.

[0018] Figure 2 This is a partial cross-sectional view of the suspended force-guiding skeleton provided in an embodiment of the present invention.

[0019] Figure 3 This is a logical structure block diagram of the intelligent companion robot provided in an embodiment of the present invention.

[0020] Figure 4 This is a flowchart of the control method for the intelligent companion robot provided in an embodiment of the present invention.

[0021] Explanation of reference numerals in the attached figures

[0022] 101 Outer shell; 102 Bionic fur layer; 103 Window; 201 Polyester fiber flocking layer; 202 Silicone substrate; 203 Polyimide flexible electrothermal film; 204 Integrated rigid force guide column; 205 Silicone cantilever damping layer; 206 Rigid base frame; 207 Force-sensitive resistor thin film sensing unit; 208 Spandex high-elasticity baseband; 301 Main control chip; 302 Auditory sensing module; 303 Tactile sensing module; 304 Motion module; 305 Emotion display module; 3021 Microelectromechanical system microphone; 3031 Tactile sensor array; 3041 First micro motor; 3042 Second micro motor. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the invention.

[0024] In one embodiment of this application, an intelligent companion robot is provided. For example... Figure 1 As shown, this intelligent companion robot adopts a miniaturized, cute pet-like design, with physical dimensions of 12cm in length, 10cm in width, and 16.5cm in height. Through optimized internal hardware stacking, the robot weighs only 350g. This size and weight distribution concentrates the robot's center of gravity at the chassis, providing excellent physical stability against tipping over when performing large-angle head turns. For external protection, the intelligent companion robot has an outer shell 101 providing internal structural support, which is entirely covered by a biomimetic fur layer 102. This biomimetic fur layer 102 is made of ultra-fine denier polyester fiber flocking material or artificial mink fur, which not only reduces surface contact hardness but also has good sound and heat transmission properties. To achieve concealed sensor placement and direct sensing, the outer shell 101 has windows 103 in its corresponding head and back areas, with rounded edges to prevent cutting the fur.

[0025] Further reading Figure 3 The logical structure diagram shown illustrates that the core control center of this intelligent companion robot is the main control chip 301 located within the inner cavity of the outer shell 101. To support system operation, the hardware circuit also includes peripheral base components, including a crystal oscillator clock circuit for providing a stable clock reference, a reset circuit, and a high-capacity lithium battery pack. The lithium battery pack is electrically connected to the main control chip 301 and other loads via a power management module. The power management module integrates a buck-boost converter circuit to output multiple working power supplies at different voltage levels. On the input side of the main control chip 301, an auditory sensing module 302 and a tactile sensing module 303 are electrically connected; on the output side of the main control chip 301, a motion module 304 and an emotion display module 305 are electrically connected.

[0026] In this embodiment, the auditory perception module 302 is responsible for spatial acquisition of sound source information. This auditory perception module 302 includes four microelectromechanical system (MEMS) microphones 3021, which are evenly distributed in a circular array along the horizontal circumference within the head region. The advantage of this circular array layout is that, regardless of whether the sound originates from the front or rear of the robot, at least three microphones can acquire direct sound waves with a high signal-to-noise ratio, providing a robust hardware sampling basis for the main control chip 301 to extract the azimuth angle of the sound source based on the time difference of arrival. Furthermore, in this embodiment, an infrared positioning module is also located at the center of the head region. This module includes an infrared sensor, which is electrically connected to the main control chip 301. It is used to accurately measure the distance and angle of the target after the head is turned and aligned with it. Its ranging accuracy can reach the centimeter level, compensating for the shortcomings of acoustic positioning in near-field ranging and angle accuracy.

[0027] The tactile sensing module 303 is responsible for converting physical deformation of the body surface into electrical signals, and it includes a tactile sensor array 3031. The tactile sensor array 3031 comprises multiple capacitive or resistive pressure sensing plates. Figure 1 and Figure 3 The corresponding relationship is that the distribution position of the touch sensor array 3031 is consistent with the position of the aforementioned window 103, that is, the sensor is embedded between the inner bottom wall of the bionic fur layer 102 and the outer shell 101. When an external object touches the fur surface, the external force acts directly on the sensing sheet, thereby outputting the change in resistance or capacitance.

[0028] At the control execution end, the motion module 304 is responsible for changing the physical posture. In this embodiment, the system is equipped with four high-precision micro servos, divided into two groups: a first micro motor 3041 and a second micro motor 3042. The first micro motor 3041 is located in the head area and consists of two independent servos, which are controlled to adjust the pitch and horizontal swing of the head area respectively. The second micro motor 3042 is located in the wing area on both sides of the torso, and also consists of two independent servos, used to drive the left and right wings to perform up and down flapping movements along the hinge axis, with a flapping mechanical angle range of 0° to 30°. The main control chip 301 sends control pulses with varying duty cycles to the motor drive circuit through a built-in pulse width modulation timer, thereby realizing continuous posture control of each micro motor.

[0029] The emotion display module 305 primarily handles visual interaction tasks and includes a high-definition LCD screen. This LCD screen 3051 is embedded in the robot's facial area, and its display driving circuit interacts with the main control chip 301 via a digital signal interface. The main control chip 301's internal storage medium contains a pre-programmed sequence of facial expression frames containing basic emotions.

[0030] During system-level data flow, the main control chip 301 concurrently receives and processes multi-channel signals. For the differential audio signals from the microphones 3021 of each microelectromechanical system (MEMS), the main control chip 301 performs analog-to-digital conversion using an audio codec and extracts the spatial azimuth angle of the external target sound source using a pre-built positioning algorithm library. For the analog level signals from the tactile sensor array 3031, the on-chip analog-to-digital converter of the main control chip 301 performs high-speed sampling and filters out glitches using a moving average filtering algorithm, extracting a tactile feature vector reflecting the interaction intent. The data structure of this tactile feature vector encapsulates the current contact force value, stroking frequency, and duration fields. Subsequently, the main control chip 301 inputs the sound source azimuth angle, the ranging results from the infrared sensor, and the tactile feature vector into a logic tree for fusion analysis, generating coordinated control commands. After the control command is issued, the motion module 304 executes a posture response and writes updated video memory to the LCD screen, switching and rendering a visual emotion interface that matches the current scenario.

[0031] It should be noted that, in order to achieve more in-depth data analysis, in an alternative embodiment, the intelligent companion robot also has a built-in wireless communication module for interacting with the cloud. The main control chip 301 uploads the packaged and compressed perception data to the cloud server, uses the more powerful emotion computing engine of the cloud to model long-term interaction behavior, and then receives high-level logic instructions issued by the cloud. For the above-mentioned sensing hardware combination, those skilled in the art can also achieve the functions of this application by using similar alternative solutions. The tactile perception channel can replace the deployment of a distributed flexible piezoresistive sensor array, or adopt a capacitive touch sensing scheme that directly detects changes in human parasitic capacitance, or adopt a piezoelectric thin film sensor that converts mechanical deformation into charge signals; the emotion visualization channel can also use a colored light-emitting diode light strip in the eye socket area in conjunction with a micro vibration motor built into the body, or use a dot matrix light-emitting diode matrix; the auditory channel can also use a dual-microphone array in conjunction with an ultrasonic positioning sensor or an infrared pyroelectric sensor for spatial detection.

[0032] By adopting the aforementioned multi-channel perception and biomimetic integrated structure, the intelligent companion robot provides a realistic tactile experience through the layering of fur and embedded sensors without increasing system redundancy modules. It enhances the robot's proactive attention and naturalness of emotional expression in the home environment by combining acoustic hearing and infrared precise ranging to control the head area response.

[0033] High-density fiber fur layers are prone to a sponge-like energy absorption effect when subjected to pressure, meaning that most of the pressure is absorbed by the fiber pores within the fur layer, leading to stress attenuation or failure of the tactile sensors at the bottom layer. Under concentrated heavy loads, the thin sensor substrate is prone to irreversible deformation. Conventional plastic laminate designs lack temperature regulation capabilities. In another embodiment of this application, the skin of the robot shell has undergone synergistic mechanical and thermodynamic improvements.

[0034] like Figure 2 As shown, in the interactive window areas corresponding to the robot's back and head, a five-layer heterogeneous stacked suspension guide frame system is introduced between the bionic fur layer and the outer shell. This suspension guide frame consists of, from the outside in, a polyester fiber flocked layer 201; a silicone substrate 202, made of food-grade silicone sheet with a thickness of 0.5mm to 1.0mm, possessing excellent deformation elongation; a frustum-shaped integrally molded rigid guide column 204, an array of columns integrally molded on the inner side of the silicone substrate, with a hardness index higher than the outer silicone layer; a silicone cantilever damping layer 205, located around the perimeter of the silicone substrate 202, composed of several Ω-shaped silicone cantilever arms, used to suspend and fix the entire silicone substrate to the bottom inner structure, maintaining a 2mm to 4mm deformation buffer gap in the vertical space; and a rigid base frame 206. The touch sensor array uses a thin-film force-sensitive resistor thin-film sensing unit 207, which is attached to the outer flat surface of the rigid base 206 by adhesive.

[0035] This embodiment constructs a physical limiting and overload limiting protection mechanism for the force-sensitive thin-film sensing unit 207. In the vertical assembly relationship, the lower end plane of each integrally molded rigid guide column 204 is on the same vertical normal as the sensitive center area of ​​the corresponding force-sensitive thin-film sensing unit 207. In the static, unpressurized state without external pressure, the lower end of the rigid guide column 204 maintains a first gap distance with the sensor surface on the outer surface of the rigid base frame 206, supported by the tension of the surrounding silicone cantilever damping layer 205. This first gap distance is preferably configured to be 0.5mm to prevent the robot from generating false trigger signals during movement and vibration.

[0036] When a normal stroking motion is applied to the back by an external object, the pressure, through the polyester fiber flocking layer 201, causes the silicone substrate 202 to undergo a vertically downward elastic displacement, providing a soft, damped feel similar to the muscle tissue of a real pet. As the substrate moves downward, the rigid guide pillar 204 crosses the 0.5mm first gap, precisely abutting and pressing against the bottom sensing unit, converting the surface load into point-to-point stress transmission and reducing distortion caused by energy absorption by the fur. When the suspended guide frame encounters an unexpected impact or other external force, and the deformation of the silicone cantilever damping layer 205 caused by this force exceeds the distance of the first gap and the sensor's tolerance limit, due to the tolerance-controlled length design of the guide pillar, the wide bottom surface of the silicone substrate 202 will sink further before the guide pillar, resulting in a large-area contact with the outer surface of the rigid base frame 206. This large-area bonding instantly forms a mechanically rigid support point, dispersing kinetic energy and deformation stress onto the high-strength rigid substrate 206. This prevents stress from penetrating and compressing the force-sensitive thin-film sensing unit 207, thus avoiding the sensor being crushed. When the external force is removed, the silicone cantilever damping layer 205 quickly releases its elastic potential energy, causing the substrate to instantly return to its original position.

[0037] The suspended force-guiding skeleton integrates a thermal structure. Inside the silicone substrate 202, a layer of flexible polyimide electrothermal film 203 and a thermistor for providing temperature feedback are tightly embedded. The power supply pin of the flexible polyimide electrothermal film 203 is electrically connected to the power control module under the main control chip 301. The main control chip 301 runs an internal proportional-integral-derivative control algorithm closed loop, continuously outputting a pulse-width modulated heating signal to the flexible polyimide electrothermal film 203. Due to the excellent thermal conductivity of the outer silicone substrate, it is configured to receive and conduct the thermal energy emitted by the flexible polyimide electrothermal film 203 to the fiber surface, regulating the surface temperature. The system performs control calculations, weighted summing the deviation between the current temperature and the target temperature using proportional, integral, and derivative operations to obtain a control signal for adjusting the output power of the heating element, as shown in the following formula:

[0038] ;

[0039] Where t represents time; u(t) represents the pulse width modulation duty cycle signal output to the heating film at time t; e(t) represents the temperature difference deviation between the target set temperature and the current surface thermistor acquisition temperature; Kp, Ki, and Kd represent the gain coefficients of the proportional, integral, and derivative elements, respectively. In the system configuration of this embodiment, the surface temperature is kept constant between 37°C and 39°C through closed-loop calculation. Continuous heating causes a slight softening deformation effect on the outer silicone substrate, improving the stiffness problem of the material in low-temperature winter environments, making the pressing deformation smoother, and achieving mechanical synergy between thermodynamic control and mechanical dynamics.

[0040] In addition to the head and wing areas, the robot also includes a torso. The outer shell 101 has a cylindrical neck revolute joint connecting the head and torso, and a hinge joint connecting the wing areas and the base of the torso. In actual operation, large-angle displacement can easily cause the outer fur to be pulled and entangled. The biomimetic fur layer 102 has a locally embedded, sewn-in, high-elasticity spandex base strip 208 within the high-frequency deformation areas corresponding to the neck revolute joint and the periphery of the hinge joint. This high-elasticity spandex base strip 208 has a lateral tensile elongation greater than 150%, configured to provide internal structural deformation compensation space when the corresponding micro-motor drives the joints to perform large-angle movements, preventing mechanical lock-up or jamming failures caused by excessive pulling of the outer fur.

[0041] The five-layer heterogeneous suspended force-guiding skeleton system and thermo-mechanical synergy mechanism provide deformation damping feedback and constant surface temperature. The spatial limiting closed loop solves the problem of thin film sensors being susceptible to overload damage, and balances data sampling sensitivity and physical reliability of operation.

[0042] In another embodiment of this application, a control method for an intelligent companion robot is provided, deployed and running on the hardware base of the intelligent companion robot disclosed in the embodiments of this application. The control method includes the following flow process:

[0043] Step S401: Acquire ambient audio signals through the auditory perception module and extract the azimuth angle of the sound source based on the time difference of arrival of the ambient audio signals. In this stage, the main control chip, in standby monitoring mode, activates four microelectromechanical system (MEMS) microphones. Due to spatial differences, there is a slight delay in the time it takes for the external sound wavefront to reach the surface of each microphone. The system timestamps and calculates the cross-correlation function of the captured acoustic analog waveforms. The arrival time difference sequence of the ambient audio signals acquired by the four MEMS microphones arranged in a circular array is obtained. Based on the preset physical spatial coordinates of the microphone array on the motherboard and the constant of sound propagation speed at room temperature, a nonlinear positioning equation set is constructed using the spherical cross-positioning principle. The system performs spherical cross-positioning calculations, using the relationship that the distance difference from each microphone to the target sound source equals the speed of sound multiplied by the signal arrival time difference to establish a nonlinear constraint on the target spatial coordinates, the formula of which is as follows:

[0044] ;

[0045] Wherein, variables x, y, and z represent the three-dimensional coordinate components of the target sound source to be solved; i is a positive integer, and its value range corresponds to the total number of microphones in the microelectromechanical system, i.e., i = 1, 2, 3, 4; x i y i z iLet represent the calibration coordinates of the i-th microphone in physical space; x1, y1, z1 represent the calibration coordinates of the first microphone used as a reference; parameter c represents the speed of sound in air; variable t i1 This represents the time difference of arrival measured between the i-th microphone and the reference microphone. By solving this set of nonlinear localization equations, the initial three-dimensional coordinates of the target sound source are obtained. This initial three-dimensional coordinate point in the Cartesian coordinate system is then mathematically mapped to a polar coordinate system with the geometric center of the intelligent companion robot as the origin. The horizontal deflection angle is calculated and extracted, and this angle is used as the sound source azimuth angle. After obtaining the sound source azimuth angle, the main control chip 301 sends a turning command to the motion module 304, driving the first micro-motor 3041 to turn the robot's head to the direction corresponding to the sound source azimuth angle. Subsequently, the infrared sensor is controlled to activate precise ranging, obtaining the precise distance and angle of the target, thus achieving the connection between acoustic coarse localization and infrared fine localization.

[0046] Step S402: The tactile sensing module acquires the pressure distribution data of the contact surface and extracts tactile feature vectors containing force and duration. The main control chip periodically scans the sensor matrix distributed in the back and head regions. When the damping frame depresses to trigger contact, the voltage division values ​​of each node are read, normalized, and then a spatial pressure distribution cloud map of the contact surface is formed. Based on this data stream, the main control logic calculates the contact peak value and uses an internal counter to record the contact duration, encapsulating and outputting the tactile feature vector.

[0047] Step S403: Perform dual-channel fusion analysis on the sound source azimuth and tactile feature vector to generate control commands. The system performs logical matching on the data streams of the two independent channels. If the sound source azimuth points to the left rear and the tactile feature vector shows that the back is currently receiving a long-lasting and gentle soothing pressure, the fusion module determines that the interaction is peaceful and synthesizes a corresponding docile following response command.

[0048] Furthermore, during the fusion positioning stage, the main control chip 301 weights and fuses the azimuth angle of the sound source with the precise distance and angle obtained by the infrared sensor according to weighting coefficients to generate a fusion positioning result. This fusion positioning result, along with the tactile feature vector, is used as input to generate control commands. The weighting coefficients are dynamically adjusted according to the environmental signal-to-noise ratio. The higher the environmental signal-to-noise ratio, the greater the weight of the azimuth angle of the sound source in the fusion positioning result. The lower the environmental signal-to-noise ratio, the greater the weight of the ranging result of the infrared sensor in the fusion positioning result. This approach combines the advantages of far-field acoustic coarse positioning and near-field infrared fine positioning.

[0049] In step S404, the control commands are sent to the motion module and the emotion display module; the motion module performs physical posture adjustment, and the emotion display module updates the facial expression status displayed on the LCD screen. The main control chip distributes the motion commands to each motor driver via serial port or bus. The first micro motor of the head adjusts its pitch and yaw angles to face the sound source; the second micro motor of the wings enters flapping mode. The memory controller writes the expression bitmap sequence to the LCD screen, completing the closed-loop control from environmental perception to hardware-based action execution.

[0050] The control method incorporates a hardware-cascaded stress processing flow to handle extreme environments. In this parallel branch, the system control chip performs a Fast Fourier Transform on the input audio signal, continuously monitoring the high-frequency energy value in the ambient audio signal. The system performs frequency band energy integration calculations, obtaining the total integral value of the acoustic energy in that high-frequency band by squaring and summing the complex spectral amplitudes at each frequency point within a specific high-frequency range. The formula is as follows:

[0051] ;

[0052] Among them, variable E high The variable X(k) represents the total integral value of acoustic energy extracted within a specific frequency band; k represents the discrete frequency index value after frequency domain transformation; the variable X(k) represents the complex spectral amplitude of the k-th frequency point after frequency domain transformation; f low with f high These represent the lower and upper frequency index values ​​that define the frequency range of sharp, abnormal noise, respectively.

[0053] When the calculated high-frequency energy value exceeds the system's preset first energy threshold, the abnormal high-frequency noise interruption logic is triggered. The main control chip pulls down the level of the corresponding power pin, cutting off the power supply to the polyimide flexible electrothermal film embedded inside the robot, simulating physiological changes under stress through physical cooling. A risk avoidance drive command is generated and sent to the motion module, driving the first micro motor to perform a defensive deflection towards the target noise direction at a first angular velocity. The value of this first angular velocity is configured in the drive register to be three times the head deflection angular velocity in normal tracking mode. The head deflection angular velocity in normal tracking mode is a following angular velocity calculated in real time based on the extracted sound source azimuth angle. An update command is sent to the display control logic, refreshing the display screen of the emotion display module to a preset alert interface with warning colors. In order to establish a physical defense posture, the control layer generates a closed drive level with suppressed pulse width and continuously sends the closed drive level to the second micro motor, driving the wings on both sides to deflect in the opposite direction to the fuselage until the inner sides of the wings physically contact and adhere to the corresponding positions on the sidewall of the torso shell, forming a closed defensive posture that tightly hugs the fuselage.

[0054] By employing control methods and cascaded stress logic, the mathematical calculation of sound wave time difference is combined with a complex hardware system to give the system an active turning response capability. Based on the environmental noise energy, it automatically triggers coherent stress contraction and interface refresh actions, enhancing the life simulation and immersion in the human-computer interaction environment.

[0055] The control method and cascaded stress logic utilize the mathematical solution of the sound wave time difference to coordinate with the hardware system to control the active steering response. Based on the environmental noise energy triggering stress contraction and interface refresh actions, the multimodal interactive response capability is improved.

[0056] It should be understood that all parts of this invention can be implemented using hardware, software, firmware, or a combination thereof. Those skilled in the art can make equivalent modifications or substitutions to the component connection structures or control signal timings in the above embodiments without departing from the core spirit of this invention; such modifications or substitutions all fall within the scope of protection of this invention.

Claims

1. An intelligent companion robot, characterized in that, include: The system comprises: an outer shell covered with a biomimetic fur layer; a head region, a back region, and a wing region; windows in the head and back regions; a main control chip; an auditory sensing module including four microelectromechanical system (MEMS) microphones arranged in a ring within the head region; a tactile sensing module including a tactile sensor array distributed at the corresponding positions of the windows and embedded between the biomimetic fur layer and the outer shell; and an infrared positioning module including an infrared sensor located in the head region for precise distance measurement and positioning of the target. The system includes a motion module comprising a first micro motor and a second micro motor, the first micro motor being disposed in the head region and the second micro motor being disposed in the wing region; an emotion display module comprising a liquid crystal display screen; an auditory perception module, a tactile perception module, an infrared positioning module, the motion module, and the emotion display module being electrically connected to the main control chip; the main control chip is configured to extract the azimuth angle of the sound source based on the audio time difference collected by the microphones of each microelectromechanical system, control the motion module to drive the head region to turn to the direction corresponding to the azimuth angle of the sound source, control the infrared sensor to perform precise ranging and positioning to obtain the ranging result, extract tactile feature vectors containing force and duration from the tactile sensor array, and fuse the azimuth angle of the sound source, the ranging result, and the tactile feature vectors to generate control commands; the control commands are used to drive the motion module and update the visualization interface of the emotion display module.

2. The intelligent companion robot as described in claim 1, characterized in that, A suspended force-guiding skeleton is disposed between the biomimetic fur layer and the outer shell; the suspended force-guiding skeleton includes a polyester fiber flocking layer, a silicone substrate, an integrally molded rigid force-guiding column, a silicone cantilever damping layer, and a rigid base frame stacked sequentially from the outside to the inside, and the touch sensor array is attached to the outer surface of the rigid base frame.

3. The intelligent companion robot as described in claim 2, characterized in that, The tactile sensor array includes a force-sensitive thin-film sensing unit; the lower end of the integrally molded rigid force guide column is aligned with the sensitive center of the force-sensitive thin-film sensing unit along the same vertical normal; in the unpressurized state, a first gap distance of 0.5 mm is provided between the lower end of the integrally molded rigid force guide column and the outer surface of the rigid substrate; when the suspended force guide frame is subjected to an external force and the external force causes the deformation of the silicone cantilever damping layer to exceed the first gap distance, the bottom surface of the silicone substrate directly contacts the outer surface of the rigid substrate, forming a mechanically rigid support point and blocking further penetration of deformation stress into the tactile sensor array.

4. The intelligent companion robot as described in claim 2, characterized in that, The suspended force-guiding skeleton also includes a polyimide flexible electrothermal film embedded inside the silicone substrate, the polyimide flexible electrothermal film being electrically connected to the main control chip; the main control chip is configured to output a pulse width modulation heating signal to the polyimide flexible electrothermal film.

5. The intelligent companion robot as described in claim 1, characterized in that, The intelligent companion robot also includes a torso; the outer shell is provided with a neck rotator at the connection between the head region and the torso, and a hinge joint at the connection between the wing region and the torso; the bionic fur layer is inlaid with a spandex high-elasticity base strip in the outer regions corresponding to the neck rotator and the hinge joint, the spandex high-elasticity base strip having a lateral elongation greater than 150%, configured to provide deformation compensation during joint operation.

6. A control method for an intelligent companion robot, characterized in that, This invention relates to an intelligent companion robot, which includes a main control chip and an auditory perception module, a tactile perception module, an infrared positioning module, a motion module, and an emotion display module electrically connected to the main control chip. The emotion display module includes an LCD screen. The control method includes: acquiring environmental audio signals through the auditory perception module and extracting the azimuth angle of the sound source based on the audio time difference; controlling the motion module to drive the head area to turn in the corresponding direction according to the azimuth angle of the sound source, and controlling the infrared positioning module to perform precise ranging and positioning to obtain the ranging result; extracting tactile feature vectors containing force and duration through the tactile perception module; fusing the azimuth angle of the sound source, the ranging result, and the tactile feature vectors to generate control commands; sending the control commands to the motion module and the emotion display module, driving the motion module and updating the visualization interface of the emotion display module through the control commands.

7. The control method as described in claim 6, characterized in that, The auditory perception module includes four microelectromechanical system (MEMS) microphones arranged in a circular array. The extraction of the sound source azimuth angle based on audio time difference includes: acquiring the arrival time difference of the environmental audio signals collected by the four MEMS microphones; constructing a nonlinear positioning equation set based on preset microphone array physical coordinates and the sound propagation speed constant; solving the nonlinear positioning equation set to obtain the initial three-dimensional coordinates of the target sound source; mapping the initial three-dimensional coordinates to a polar coordinate system centered on the intelligent companion robot, and extracting the horizontal plane deflection angle as the sound source azimuth angle. The control method further includes: controlling the infrared sensor in the infrared positioning module to accurately measure the distance to the target and obtaining the distance measurement result; weighting and fusing the sound source azimuth angle with the precise distance and the precise angle according to weighting coefficients to generate a fused positioning result; wherein the weighting coefficients are dynamically adjusted according to the environmental signal-to-noise ratio (SNR), and the higher the SNR, the greater the weight of the sound source azimuth angle.

8. The control method as described in claim 6, characterized in that, The intelligent companion robot also includes a torso, and a flexible polyimide electrothermal film is embedded inside the intelligent companion robot; the motion module includes a first micro motor corresponding to the head area and a second micro motor corresponding to the wing area; the control method also includes hardware cascaded stress processing, which includes: continuously monitoring the high-frequency energy value of the environmental audio signal; when the high-frequency energy value exceeds a first energy threshold, triggering abnormal high-frequency noise interruption processing logic; cutting off the power supply to the flexible polyimide electrothermal film; generating a risk avoidance drive command and sending it to the motion module, driving the first micro motor to deflect according to a first angular velocity; refreshing the display screen of the emotion display module to a preset alert interface, and driving the second micro motor to perform a contraction and closure action.

9. The control method as described in claim 8, characterized in that, The first angular velocity is configured to be three times the head deflection angular velocity in the normal tracking state; the head deflection angular velocity in the normal tracking state is the following angular velocity calculated in real time based on the extracted azimuth angle of the sound source; the driving the second micro motor to perform the contraction and closure action includes generating a closure drive level and continuously sending the closure drive level to the second micro motor, driving the wings to deflect inward until they physically contact the corresponding position on the outside of the torso to form a closed defensive posture.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the control method as described in any one of claims 6 to 9.

Images