Robot control method and device, robot, and storage medium

Abstract

The embodiment of the application provides a robot control method and device, a robot and a storage medium, the method comprises the following steps: acquiring first perception information; performing a first behavior; wherein the first behavior is one of a plurality of candidate behaviors corresponding to the first perception information; and / or the priority of the perception information corresponding to the first behavior is higher than or equal to the priority of the first perception information. The robot can present a more natural and richer interaction effect.

Description

Technical Field

[0001] This application relates to the field of robotics, and more particularly to a robot control method, apparatus, robot, and storage medium. Background Technology

[0002] Robots are typically able to perceive their external environment, user input, or their own state, and perform corresponding actions based on the perceived information. For example, after detecting a user's voice, gesture, or touch operation, a robot can perform tasks such as turning, moving, playing audio, or switching working modes.

[0003] In related technologies, robots typically respond to perceived information in a relatively fixed way. For example, given the same voice, gesture, or touch input, robots often consistently perform the same behavior; when given the same input multiple times, robots tend to repeat the same actions, resulting in unnatural and unvarnished behavior and a poor interactive experience. Summary of the Invention

[0004] Based on this, this application provides a robot control method, device, robot, and storage medium that enable the robot to present more natural and richer interactive effects.

[0005] In a first aspect, embodiments of this application provide a robot control method, the method comprising: Obtain first-hand sensory information; Perform the first action; Wherein, the first behavior is one of a plurality of candidate behaviors corresponding to the first perceived information; and / or If the robot is performing an action corresponding to perception information with a priority of level two, the priority of the perception information corresponding to the first action is higher than or equal to level two.

[0006] Secondly, embodiments of this application provide a control device, the control device comprising: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the aforementioned method.

[0007] Thirdly, embodiments of this application provide a robot, which includes the aforementioned control device.

[0008] Fourthly, embodiments of this application provide a computer-readable storage medium, characterized in that the computer-readable storage medium stores computer-executable instructions, which, when executed, can perform the aforementioned method.

[0009] This application provides a robot control method, device, robot, and storage medium. The method includes: acquiring first perception information; executing a first action; wherein the first action is one of a plurality of candidate actions corresponding to the first perception information; and / or if the robot executes an action corresponding to perception information with a priority of second level, the priority of the perception information corresponding to the first action is higher than or equal to the second level. This enables the robot to present more natural and richer interactive effects.

[0010] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit the disclosure of the embodiments of this application. Attached Figure Description

[0011] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a flowchart illustrating a robot control method provided in an embodiment of this application; Figure 2 This is a schematic diagram of the robot according to an embodiment of this application; Figure 3 This is a schematic diagram of a control device provided in an embodiment of this application; Figure 4 This is a schematic diagram of a robot provided in one embodiment of this application. Detailed Implementation

[0013] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0014] The flowchart shown in the attached diagram is for illustrative purposes only and does not necessarily include all content and operations / steps, nor does it necessarily have to be performed in the order described. For example, some operations / steps can be broken down, combined, or partially merged, so the actual execution order may change depending on the actual situation.

[0015] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0016] like Figure 1 The illustration shows a robot control method provided in this application embodiment. The control method can control the robot to acquire perception information and execute target behaviors corresponding to the perception information.

[0017] like Figure 1 As shown, the robot control method includes steps S110 to S120.

[0018] Step S110: Obtain first perception information; Step S120: Execute the first action.

[0019] In some implementations, the first behavior is one of multiple candidate behaviors corresponding to the first perceived information. After acquiring the first perceived information, the robot does not mechanically execute a single fixed behavior, but rather selects one of the multiple candidate behaviors corresponding to the first perceived information to execute. The same first perceived information can trigger different first behaviors, allowing the robot to repeatedly execute a single behavior in the same or similar scenarios. This can enhance the diversity of robot interaction behaviors, for example, making the robot's response closer to the non-fixed reactions of real humans or pets, enabling the robot to present more natural and richer interaction effects.

[0020] In some implementations, the priority of the sensing information corresponding to the first action is higher than or equal to the priority of the first sensing information. For example, when the robot is executing an action corresponding to sensing information with a priority of level two, if the priority of the acquired first sensing information is higher than or equal to level two, the robot can execute the first action corresponding to the first sensing information. This allows for prioritizing responses to higher-priority sensing information. If the priority of the acquired sensing information is higher than the priority of the sensing information corresponding to the currently executed action, the sensing action of the lower-priority sensing information can be interrupted. If the priority of the acquired sensing information is lower than the priority of the sensing information corresponding to the currently executed action, the lower-priority sensing information can be ignored, and the current action can continue to be executed.

[0021] It can prevent robots from continuously executing low-priority or equal-priority behaviors, enabling robots to respond promptly to first-level perceptual information with a priority higher than or equal to the second level and execute the first behavior corresponding to the first-level perceptual information. This can improve the naturalness, diversity, richness, and scene adaptability of robot interaction behaviors.

[0022] In some implementations, the robot can be any of the following: a photography robot, an inspection robot, a security robot, a transport robot, or a service robot. The robot is not limited to humanoid robots; for example, it can also be a robot dog, a wheeled robot cat, or other similar robots.

[0023] In some implementations, such as Figure 2 The diagram shown is a schematic of a robot in some embodiments of this application.

[0024] The robot may include a body 10, a head 20, and leg mechanisms 30. The head 20 is movably connected to the body 10 and includes a camera module 21 and an eye lighting module 22. The leg mechanisms 30 are connected to the body 10 and are used to drive the robot to move on the ground and / or adjust the posture of the body 10. The camera module 21 can capture images of the robot's environment to obtain environmental information, which can serve as the robot's perception information. The head 20's movable connection to the body 10 allows the head 20 to rotate, swing, or pitch relative to the body 10. The camera module 21 can adjust its shooting direction according to the head 20's posture, expanding the robot's environmental perception range and improving its perceptual flexibility towards user interaction objects. The eye lighting module 22 can express different states, for example, through light on / off states, color changes, flashing frequency, or dynamic lighting effects, enabling the robot to achieve interactive expression capabilities and intuitive interactive feedback.

[0025] In some embodiments, the leg mechanism 30 is used to adjust the height difference between the left and right sides of the robot body 10. For example, the leg mechanism 30 includes at least a first leg mechanism 31 and a second leg mechanism 32, with the first leg mechanism 31 connected to the left side of the robot body 10 and the second leg mechanism 32 connected to the right side of the robot body 10. At least one of the first leg mechanism 31 and the second leg mechanism 32 can be retracted or extended to adjust the height difference between the left and right sides of the robot body 10. For example, adjusting the height difference between the front and rear sides of the robot body 10 can adjust the posture of the robot's head 20. For instance, when the robot is traveling on uneven ground, adjusting the height difference between the front and rear sides of the robot body 10 can make the shooting direction of the shooting module 21 more consistent when the robot is traveling on flat ground.

[0026] In some embodiments, the leg mechanism 30 is used to adjust the height difference between the front and rear sides of the robot body 10. For example, the leg mechanism 30 includes at least a first leg mechanism 31 and a second leg mechanism 32, with the first leg mechanism 31 connected to the left side of the robot body 10 and the second leg mechanism 32 connected to the right side of the robot body 10. The robot body 10 can move relative to the first leg mechanism 31 and the second leg mechanism 32 to adjust the height difference between the front and rear sides of the robot body 10. For example, when the slope of the robot increases or decreases, adjusting the height difference between the front and rear sides of the robot body 10 can stabilize the shooting direction of the shooting module 21, for example, making the shooting direction of the shooting module 21 more consistent when the robot travels on different slopes.

[0027] In some implementations, the height difference between the left and right sides and / or the front and rear sides of the robot body 10 is adjusted by the leg mechanism 30, enabling the robot to actively change the roll and / or pitch posture of the robot body 10, thereby improving the naturalness of the robot's interactive behavior and the diversity of its postures.

[0028] In some implementations, the robot control method includes at least an interactive sensing step and an interactive execution step. For example... Figure 2 As shown, the interactive perception step may include step S110, and the interactive execution step may include step S120.

[0029] The interactive sensing step is used to acquire the robot's sensory information, and the interactive execution step is used to control the robot to perform corresponding sensory behaviors.

[0030] Optionally, the robot control method may also include an interactive decision-making process. This process is used to determine, based on the perceived information, whether the robot should perform a responsive action and to determine the target action to be performed.

[0031] The sensory information and the sensory behavior executed by a robot can be in a one-to-one correspondence, a one-to-many correspondence, a many-to-one correspondence, or a combination of the above.

[0032] For example, in some embodiments, the robot can perform a fixed corresponding sensing behavior based on a certain sensing information it has perceived.

[0033] In some embodiments, the robot may also execute one of a plurality of corresponding perceptual behaviors based on a certain perceptual information it has perceived.

[0034] In some embodiments, the robot may perform a single perceptual action when it perceives multiple different perceptual information.

[0035] In the process of robot interaction and decision-making, the priority of perceived information can also be considered.

[0036] For example, if a perception action with lower priority information is being executed, and a perception action with higher priority information is received, the execution of the perception action with lower priority information is stopped, and the perception action with higher priority information is executed instead.

[0037] For example, if a higher-priority perceptual action is being performed and a lower-priority perceptual action is received, then the lower-priority perceptual action will not be responded to.

[0038] For example, if a sensing action with a certain priority is currently being executed, and sensing information of the same priority is received, it may or may not respond.

[0039] In some embodiments, the robot's interactive decision-making process includes a perception cooling-off mechanism. For example, after responding to a certain perception information, the robot may not respond to that perception information again for a preset time period. The preset time period may be, for example, but is not limited to, 30 seconds. The cooling-off time for each type of perception information can be set according to the characteristics of the perception information. The cooling-off time for different types of perception information may be the same or different.

[0040] In some implementations, the robot's perception information may include at least one of visual perception information, auditory perception information, tactile perception information, posture perception information, user operation information, and system status information.

[0041] For example, visual perception information may include user gestures, human body position, human body movements, color information, or target object information. Auditory perception information may include wake words, speech content, sound type, sound intensity, or sound source direction; wherein, speech content may include, for example, praising speech, accusing speech, etc., and sound type may include, for example, laughter, crying, dog barking, loud noise, etc. Touch perception information may include the robot's head being turned, the head being pressed, the body being pushed, the body being repeatedly pushed, or the robot being picked up. Posture perception information may include the robot falling, standing up, tilting, shaking, or posture recovery. User operation information may include operations such as deleting photos from the album or viewing video highlights. System status information may include the robot being powered on, powered off, standing up, stored, low battery, or switching working modes, wherein switching working modes may include, for example, switching to shooting mode.

[0042] For example, the user gestures in the visual perception information may include at least one of the following: thumbs up, thumbs left, thumbs right, OK, V, waving, and static palm. The visual perception information may also include at least one of the following: the robot seeing red, the robot seeing a person jumping, and the robot seeing someone in front of it.

[0043] For example, the auditory perception information may include at least one of the following: the robot hears a wake word, the robot hears a loud sound, the robot hears a dog bark, the robot hears praise, the robot hears an accusation, the robot hears laughter, and the robot hears crying. For example, the wake word may include "hi," the praise voice may include "good," and the accusation voice may include "No."

[0044] For example, the touch sensing information may include at least one of the following: the robot head is rotated, the robot head is pressed, the robot body is pushed forward or backward, or the robot body is repeatedly pushed back and forth.

[0045] For example, the posture perception information may include at least one of the following: the robot leaving the ground, the robot falling and getting up, the robot tilting, the robot swaying, or the robot regaining its posture.

[0046] For example, the user operation information may include at least one of the following: the user deleted photos from the album, or the user viewed highlight moments in the album videos.

[0047] For example, the system status information may include at least one of the following: the robot is standing up, the robot is being stored away, and the robot has low battery.

[0048] In some implementations, the robot’s behavior may include one or more of the following: motion, sound effects, eye lighting effects, and mouth output.

[0049] For example, the behaviors performed by the robot at the same time may include one or more of the following: actions, sound effects, eye lighting effects, and mouth output.

[0050] For example, the behaviors performed by a robot over a period of time may include one or more of the following: actions, sound effects, eye lighting effects, and mouth outputs, performed sequentially or synchronously according to a preset time sequence. For instance, when a robot performs a greeting, the mouth output may be synchronized with a nodding action or a body raising action.

[0051] For example, a robot's actions can include at least one of the following: body movement, body turning, head yaw, body pitch, body tilt, and changes in body height. Body movement can include moving forward, backward, following a target, moving towards or away from a target; body turning can include turning left, turning right, or turning towards a sound source; head yaw can include turning the head left, turning right, or turning towards a target; body pitch can include tilting up, down, nodding, or pitch jittering; body tilt can include tilting to the left, tilting to the right, or alternating left and right swaying; changes in body height can include raising the body, lowering the body, elastic undulation, or irregular height fluctuations.

[0052] For example, a robot's sound effects can be determined by one or more sound effect parameters, including sound effect type, sound effect frequency, volume, tremolo, and amplitude modulation. Different sound effects may correspond to different sound effect parameters, allowing the robot to output sound effects matching different actions. Optionally, sound effect types may include, for example, sweep sound effects, multi-note sound effects, compound sound effects, cue sounds, or low-frequency humming. Sound effect frequencies may include, for example, an upward frequency varying from low to high frequencies, or a downward frequency varying from high to low frequencies. Optionally, tremolo sound effect parameters may include tremolo frequency parameters, and amplitude modulation sound effect parameters may include amplitude modulation depth or amplitude modulation frequency.

[0053] For example, the robot's eye lighting effects may include one or more of the following: eye opening degree, eye brightness, pupil position, pupil movement, light spot radius, blinking, eye-closing time, and brightness variation. Optionally, eye opening degree can be used to control the degree of eye opening; eye brightness can be used to control the brightness of the eye display; pupil position can be used to control the pupil's horizontal or vertical offset; pupil movement may include slow drifting, rapid scanning, circular motion, or irregular drifting; light spot radius can be used to characterize the degree of eye focus; blinking may include natural blinking, rapid blinking, or prolonged eye-closing; brightness variation may include brightness pulses, breathing-like gradual changes, or irregular fluctuations.

[0054] For example, the robot's mouth output may include one or more of the following: speech synthesis output, preset speech output, phrase output, prompt output, or onomatopoeic output. Optionally, the mouth output may be based on speech synthesis (Text-to-Speech, TTS).

[0055] In some implementations, one or more of the following can be configured for the robot's behavior: motion track, sound track, eye lighting track, and mouth track.

[0056] For example, the motion track indicates the sequence of actions that the robot needs to perform when executing a behavior. For instance, the motion sequence may include one or more actions such as body movement, body turning, head yaw, fuselage pitch, body roll, and changes in fuselage altitude. The sound effect track indicates the sequence of sound effects that the robot needs to output when executing a behavior. For instance, the sound effect sequence may include one or more sound effects. The eye lighting effect track indicates the sequence of eye lighting effects that the robot needs to output when executing a behavior. For instance, the eye lighting effect sequence may include one or more eye lighting effects. The mouth track indicates the sequence of speech that the robot needs to output when executing a behavior. For instance, the speech sequence may include one or more of synthesized speech, prompting speech, and onomatopoeic output.

[0057] The motion track, sound track, eye lighting track, and mouth track can be executed along the same timeline, allowing the robot's body movements, sound output, eye lighting effects, and mouth output to coordinate in time when performing actions. For example, when the robot is greeting, the motion track can control the robot to raise its body, face the target direction, nod, or swing; the eye lighting track can control the robot to increase its eye brightness or have its pupils look towards the target direction; the sound track can control the robot to output a greeting sound effect at the highest point of its body's bounce or at the moment of nodding; and the mouth track can control the robot to output a greeting voice.

[0058] By configuring motion tracks, sound effect tracks, eye lighting effect tracks, and mouth tracks for a robot's behavior, the robot's target behavior can be broken down into one or more execution dimensions, such as motion, sound effect, eye lighting effect, and mouth output. This allows multiple execution dimensions to be output in a coordinated manner, thereby improving the consistency and naturalness of the robot's behavior output.

[0059] In some implementations, the robot can be controlled to perform corresponding actions by using the lever position of a joystick corresponding to one or more degrees of freedom (also known as control dimensions) of the robot. The joystick corresponding to each degree of freedom may include a velocity joystick and / or a position joystick, wherein the lever position of the velocity joystick is used to control the robot to move at a corresponding velocity in the corresponding degree of freedom, and the lever position of the position joystick is used to control the robot to remain at a corresponding position in the corresponding degree of freedom.

[0060] For example, the robot's degrees of freedom may include one or more of the following: forward / backward degrees of freedom, turning degrees of freedom, head yaw degrees of freedom, body pitch degrees of freedom, body tilt degrees of freedom, and body height degrees of freedom. The robot's motion control can be converted into normalized lever or control variables across multiple degrees of freedom, facilitating parametric control of the robot's body movement, body turning, gimbal / head yaw, body pitch, body tilt, and body height.

[0061] For example, the forward / backward degree of freedom is used to control the robot's body to move forward or backward. This degree of freedom can be controlled by velocity, and its normalized lever value can be from -1.0 to +1.0. For example, the maximum absolute value of the linear velocity corresponding to this degree of freedom can be 2 m / s. When the normalized lever value is +1.0, the robot can move forward at a speed of 2 m / s; when the normalized lever value is -1.0, the robot can move backward at a speed of 2 m / s.

[0062] For example, the body turning degree of freedom is used to control the robot's body to turn left or right. This degree of freedom can be controlled by velocity, and its normalized lever value can be -1.0 to +1.0. For example, the maximum absolute value of the turning angular velocity corresponding to this degree of freedom can be 90° / s. When the normalized lever value is +1.0, the robot can turn left with an angular velocity of 90° / s; when the normalized lever value is -1.0, the robot can turn right with an angular velocity of 90° / s.

[0063] For example, the gimbal / head yaw degree of freedom is used to control the robot gimbal or head to yaw left or right. This degree of freedom can be controlled by velocity, and its normalized lever value can be -1.0 to +1.0. For example, the maximum absolute value of the yaw angular velocity corresponding to this degree of freedom can be 45° / s. When the normalized lever value is +1.0, the robot gimbal or head can yaw left at an angular velocity of 45° / s; when the normalized lever value is -1.0, the robot gimbal or head can yaw right at an angular velocity of 45° / s.

[0064] For example, the pitch degree of freedom is used to control the robot's head to tilt up, down, or nod. This degree of freedom can be controlled by position control, and its normalized lever value can be from -1.0 to +1.0. For example, the maximum absolute value of the pitch angle corresponding to this degree of freedom can be 30°. When the normalized lever value is +1.0, the robot can tilt up 30°; when the normalized lever value is -1.0, the robot can tilt down 30°.

[0065] The body tilt degree of freedom is used to control the robot's body to tilt to the left or right. This degree of freedom can be controlled by position control, and its normalized lever value can be -1.0 to +1.0. For example, the maximum absolute value of the tilt angle corresponding to this degree of freedom can be 17°. When the normalized lever value is +1.0, the robot body can tilt 17° to the left; when the normalized lever value is -1.0, the robot body can tilt 17° to the right.

[0066] The body height adjustment degree of freedom is used to control the robot's body height. This degree of freedom can be achieved through position control, and its control value can range from 0 to 31. For example, the body height adjustment range corresponding to this degree of freedom can be from 5cm to 17cm. When the control value is 0, the robot's body height is approximately 5cm; when the control value is 31, the robot's body height is approximately 17cm.

[0067] In some implementations, the correspondence between the target behavior performed by the robot and the perceived information can be shown in Table 1: Table 1 Perceive information Perceptual behavior Priority Gesture recognition: thumbs up Robot jumps in place 3 Gesture recognition: thumb pointing to the left Robot doing a left somersault 3 Gesture recognition: thumb pointing to the right Robot Right Somersault 3 Gesture recognition OK gesture The robot nodded 3 Gesture recognition V gesture Robot starts recording 3 Gesture recognition wave The robot greeted everyone happily. 3 Gesture recognition static palm The robot follows the movement of the hand. 3 The robot's head was rotated The robot was confused / tricked. 3 The robot was pressed down on the head The robot hops in place 3 The robot is pushed forward or backward. The robot made an excited sound. 3 The robot was repeatedly pushed back and forth. The robot made a painful sound. 3 The robot was picked up The robot made a painful sound. 3 The robot got up after falling down Robot feels dizzy / Shake head 3 No one interacted with the robot for a period of time. Robot resting behavior (observing in place / breathing behavior) 1 The robot hears the wake word The robot's head turned towards the direction of the sound, and its body followed suit. 3+ The robot heard a loud sound The robot was startled and stood still. 3 The robot heard the dog barking The robot was scared 3 The robot hears praise The robot felt happy / excited (and remained quite active for a period of time afterward). 3+ The robot heard the accusation The current behavior ceases (and the behavior remains relatively quiet for a period of time afterward). 3+ The robot sees red Feel happy 3 The robot saw the person in front of it jump. The robot jumps in place once 3 The robot heard laughter Robot cheers 2 The robot saw someone in front of it The robot stared at the person in front of it. 2 The robot heard the crying The robot felt sad / wronged. 2 The user deleted photos from the album. The robot felt sad / wronged. 3 Users viewed the highlight moments in the album videos. The robot is happy / cheers 3 The robot stood up robot wake-up behavior 4 Robot storage Robot ends behavior 4 Robot low battery Robot low battery behaviors: • Unsteady • Dizzy • Discomfort due to low battery 4 The priority numbers in the table above are used to represent the relative magnitude of each piece of perceived information, without limiting the priority of each piece of perceived information to the numbers themselves.

[0068] For example, the robot can perform a corresponding target behavior based on the type of gesture it recognizes. For instance, when recognizing a thumbs-up gesture, the robot's target behavior could include jumping in place; when recognizing a thumbs-left gesture, the robot's target behavior could include a left somersault; when recognizing a thumbs-right gesture, the robot's target behavior could include a right somersault; when recognizing an OK gesture, the robot's target behavior could include nodding; when recognizing a V gesture, the robot's target behavior could include starting video recording; when recognizing a waving gesture, the robot's target behavior could include greeting; and when recognizing a static hand, the robot's target behavior could include following the static hand.

[0069] For example, a robot can perform corresponding target behaviors based on detected tactile perception information. For instance, when it detects that the robot's head is turned, the robot's target behavior may include a puzzled or teased behavior; when it detects that the robot's head is pressed, the robot's target behavior may include a small hop in place; when it detects that the robot is pushed forward or backward, the robot's target behavior may include outputting an excited sound; and when it detects that the robot is repeatedly pushed back and forth, the robot's target behavior may include outputting a uncomfortable sound.

[0070] For example, a robot can perform corresponding target behaviors based on detected posture perception information. For instance, when it detects that the robot has been picked up, its target behavior could include emitting a sound indicating discomfort; when it detects that the robot has fallen and then stood up, its target behavior could include exhibiting dizziness or shaking its head.

[0071] For example, a robot can perform corresponding target behaviors based on the auditory perception information it has recognized. For instance, when a wake word is recognized, the robot's target behavior could include turning its head towards the direction of the sound and then turning its body in that direction; when a loud sound is detected, the robot's target behavior could include startled actions; when a dog barks, the robot's target behavior could include fear-related behaviors; when praise is recognized, the robot's target behavior could include happy or excited behaviors, and may also increase its activity level for a period of time afterward; when accusatory speech is recognized, the robot's target behavior could include stopping its current behavior, and may also decrease its activity level for a period of time afterward; when laughter is detected, the robot's target behavior could include cheering; and when crying is detected, the robot's target behavior could include sad or aggrieved behaviors.

[0072] For example, a robot can perform corresponding target behaviors based on the visual environment information it has recognized. For instance, when it detects a red object, the robot's target behavior could include a happy behavior; when it detects a person jumping in front of it, the robot's target behavior could include jumping once in place; when it detects someone in front of it, the robot's target behavior could include controlling its head, eyes, or body to face that person to perform a gaze-based behavior.

[0073] For example, a robot can perform corresponding target behaviors based on the user's operation information. For instance, when it detects that a user has deleted photos from their album, the robot's target behavior could include sadness or aggrieved behavior; when it detects that a user is viewing a highlight moment in a video in their album, the robot's target behavior could include happiness or cheering behavior.

[0074] For example, the robot can perform corresponding target behaviors based on the acquired system status information. For instance, when the robot is detected standing up, its target behavior may include a wake-up behavior; when the robot is detected entering a storage state, its target behavior may include an end behavior; and when the robot is detected to be in a low-battery state, its target behavior may include a low-battery behavior. The low-battery behavior may include at least one of the following: wobbly behavior, dizzy behavior, and low-battery distress behavior.

[0075] In some implementations, the robot's target behavior may include one or more of the following: happy behavior, sad behavior, resting behavior, greeting behavior, curious behavior, dizzy behavior, fearful behavior, cheering behavior, excited behavior, uncomfortable behavior, aggrieved behavior, startled behavior, following behavior, staring behavior, waking behavior, ending behavior, and low battery behavior.

[0076] For example, happy behavior can be used to induce more active behavioral output in the robot. Sad behavior can be used to induce less active behavioral output in the robot. Quiet behavior can be performed when the robot does not detect interactive sensory information within a preset time. Greeting behavior can be performed when the robot responds to interactive sensory information such as waving, waking up, or approaching the user. Curiosity behavior can be performed when the robot responds to sensory information such as newly appearing targets, sounds, movements, or touches. Dizziness behavior can be performed when the robot gets up after falling, is pushed, undergoes rapid changes in posture, or has low battery. Fear behavior can be performed when the robot responds to dog barks, loud noises, or other preset sound types. Cheerful behavior can be performed when the robot responds to sensory information such as laughter, video highlights, and positive interactions. Excited behavior can be performed when the robot responds to praise, pushing, or other positive stimuli. Discomfort behavior can be performed when the robot responds to being repeatedly pushed, picked up, or other preset touch / posture sensory information. Frustrated behavior can be performed when the robot responds to crying, a user deleting photos from the album, or other preset negative events. Startled behavior can be performed when the robot responds to sensory information such as loud noises or sudden appearances. Follow behavior can be used for robots to respond to the position of a static hand, a human body, or a target object. Gaze behavior can be used for robots to respond to detecting a person in front of them, detecting the direction of a sound source, or detecting a target object. Wake-up behavior can be used for robots to respond to standing up, powering on, a wake-up word, or other wake-up-related sensory information. Termination behavior can be used for robots to enter storage, power off, or end their working state. Low battery behavior can be used for robots to respond to a low battery state. Function execution behavior can be used for robots to respond to sensory information with a clear functional indication.

[0077] It should be noted that the terms "happy behavior," "sad behavior," "curious behavior," "fearful behavior," and "aggrieved behavior" mentioned above are used to distinguish different output types of behaviors performed by the robot and do not indicate that the robot actually possesses the corresponding emotions. For example, happy behavior can be called a type 1 interactive behavior, and sad behavior can be called a type 2 interactive behavior. Each target behavior can include one or more of the following: actions, sound effects, eye lighting effects, and mouth output. The actions, sound effects, eye lighting effects, and mouth output can be different for different target behaviors.

[0078] The "happy behavior" can be termed the first type of interactive behavior. For example, the happy behavior may include one or more of the following: the device body undulating within a first height range, the device body tilting upwards, the eyes squinting, the brightness of the eyes varying within a first brightness range, and the sound output frequency of the sound-emitting part within a first frequency range. The first height range is, for example, 10cm to 50cm, or for example, 15cm to 25cm. The first brightness range is, for example, 0.5 to 1, or for example, 0.6 to 0.9. The first frequency range is, for example, 500Hz to 5000Hz, or for example, 500Hz to 3000Hz.

[0079] For example, the robot's behavior in the "happy behavior" exercise can include one or more of the following: the robot can control its body height to maintain a high standing position of 18cm to 24cm and perform elastic fluctuations of ±2cm to ±4cm, which is the source of a sense of lightness; control the head yaw to make small left-right swings of ±0.2 to ±0.4 at a frequency of 1.5Hz to 2Hz to convey a pleasant sense of rhythm; control the body tilt to coordinate with the head yaw to increase the richness of the swaying; control the body pitch to maintain a slightly tilted head posture of +0.05 to +0.15 to express a positive and open feeling. When all movements are in place, spring easing or elastic easing can be used to produce a bouncy rebound to reduce the mechanical feeling of linear stopping. The eye lighting effects during the "Happy Behavior" activity can include: setting the eye opening to 0.65 to 0.75 so that the upper eyelid partially covers the eyes, creating a squinting smile; increasing the eye brightness to 0.7 to 0.8 to make the eyes brighter than usual; and synchronizing the eye brightness with ±5% pulses based on the body's height jump, ensuring synchronized changes between the eyes and body. The sound effects during the "Happy Behavior" activity can include: controlling the sound frequency to rise from 600Hz to 2500Hz to express a positive signal, and setting an amplitude modulation of 6Hz to 8Hz to add a lively, vibrating quality to the sound. The triggering timing of the sound effects can be aligned with the highest point of the body's jump, synchronizing the sound effects with the movement rhythm.

[0080] Sadness behaviors can be termed second-type interactive behaviors. For example, sadness behaviors may include one or more of the following: the height of the device decreases, the device droops, the eyes are half-closed or closed, the brightness of the eyes decreases, and the frequency of the vocalization decreases.

[0081] For example, the robot's behavior in expressions of sadness could include one or more of the following: the robot could control its height to slowly decrease from a default height of 16cm to 6cm to 12cm, representing a gradual collapse and loss of strength; control its pitch to gradually change to a downward head posture of -0.15 to -0.7 to express dejection; control its head to yaw slowly at a frequency of 0.2Hz to 0.8Hz with a movement of ±0.05 to ±0.4, conveying powerlessness and helplessness; control its body to tilt slightly to ±0.03 to ±0.25, expressing loss of balance and spiritlessness. The robot could also insert a pitch jerking motion in the middle of the movement, which includes a small, rapid head tilt followed by a slow descent, simulating a sighing motion. The overall movement could employ a gradual easing to create a heavy, downward feel. Eye lighting effects in sad behaviors can include: reducing eye opening to 0.2 to 0.6 to present a half-closed, listless state; reducing eye brightness to 0.1 to 0.5 to express dull, lifeless eyes; shifting the pupil position downwards, for example, making the pupil position in the vertical direction -0.2 to -0.6, forming a downward gaze; making the eye brightness gradually change slowly in a breathing pattern with a cycle of 2 to 6 seconds to express weak signs of life; reducing the blinking frequency and extending the closed eye time to 300ms to 800ms to express drowsiness and unwillingness to open the eyes. Sound effects for sad behaviors can include: using sweeping or composite sound effects, controlling the sound effect frequency from 600Hz to 1500Hz down to 100Hz to 500Hz to convey negativity and loss, using a gradual easing in the sound effect curve, making the sound effect start fast and end long, simulating the breath curve of a sigh; the sound effect duration is 0.2 seconds to 5 seconds to create a prolonged sigh feeling, the vibration frequency is 1Hz to 6Hz to increase the trembling low texture, and the volume is 0.2 to 0.6 to express that it is faint and should not be loud.

[0082] Resting behavior can be referred to as a third type of interactive behavior. For example, resting behavior may include one or more of the following: the body undulating within a first frequency range, the body pitching within the first frequency range, and the brightness of the eyes changing within the first frequency range.

[0083] For example, the robot's behavior during resting behavior can include one or more of the following: the robot can control its body height to fluctuate by ±1cm to ±10cm within a baseline of 8cm to 20cm, with a fluctuation frequency of 0.15Hz to 0.6Hz, such as a breathing cycle of 2 to 7 seconds, to simulate the natural breathing rhythm of a living organism; control the body pitch to move in tandem with body height fluctuations by ±0.02 to ±0.8, for example, slightly raising its head during the rising phase and slightly lowering its head during the lowering phase; control the head yaw to yaw slightly by ±0.03 to ±0.2 every 4 to 12 seconds, to simulate occasionally looking around and then slowly returning to center; control the body tilt to make slight center of gravity drifts by ±0.01 to ±0.2. The overall movement amplitude of the resting behavior is extremely small, inconspicuous but maintaining a sense of life. To avoid completely periodic repetition, there can be slight variations in amplitude and rhythm between multiple breathing fluctuations to create an irregular, organic feel. The eye lighting effects during the resting behavior can include: maintaining the eye opening at a natural eye state of 0.1 to 0.9 to express relaxation but basic alertness; maintaining eye brightness at 0.5 to 0.8, and making the eye brightness pulsate slightly at ±3% to ±8% in sync with breathing; controlling the robot to blink naturally every 3 to 10 seconds to express maintaining basic life signs; controlling the pupil to drift slowly in the horizontal or vertical direction at ±0.05 to ±0.4, with a transition time of 300ms to 1500ms, simulating the natural movement of the gaze. The resting behavior can remain silent without sound effects; when sound effects are needed, a low-frequency hum at a volume of 0.05 to 0.25 can be output to simulate the soft sound of breathing. The duration of the resting behavior can be 8 to 30 seconds and can be looped.

[0084] Greeting behavior can be referred to as the fourth type of interaction behavior. For example, greeting behavior may include one or more of the following: the robot jumps up, the body rises and falls, the body swings, the head faces the target direction, the eyes are open, the brightness of the eyes increases, and the frequency of the voice output increases.

[0085] For example, the robot's behavior in greeting can include one or more of the following: the robot can control its body height to quickly bounce to a height of 22cm to 31cm, such as by using gradual acceleration to achieve a rapid increase in height; then control the body to bounce at an amplitude of ±2 to ±8 and a frequency of 1.0Hz to 2.5Hz to express cheerfulness; control the body to tilt to the side and swing alternately 2 to 10 times at a range of ±0.15 to ±0.8, swaying left and right to simulate the enthusiasm of waving; control the body to tilt up first at +0.05 to +0.8 to indicate that it has spotted someone, then quickly nod to -0.15 to -0.8 to indicate a greeting, and then bounce back; control the head to yaw at ±0.15 to ±0.6 and face the target direction (such as the direction of the person); control the forward / backward degrees of freedom at +0.03 to +0.15 to form a slightly forward-facing action to express the willingness to approach and enthusiasm. The aforementioned actions can employ gradual acceleration and easing, combined with elastic or rebound easing, to create an enthusiastic burst of energy and a bouncy, swaying effect. Eye lighting effects during greetings can include: setting the eye opening to 0.5 to 1.0 (wide open) to express a state of surprise and delight; increasing eye brightness to 0.5 to 1.0 (high brightness) for an excited, bright look; controlling the pupils to quickly look in the target direction (e.g., the direction of the person approaching) to indicate that the gaze has locked onto the target; and performing a rapid blink at the start of the greeting to simulate the instantaneous reaction of attention being awakened. Sound effects during greetings can include: using sweeping or multi-note sound effects, controlling the sound frequency from 400Hz to 1200Hz and then rising to 1200Hz to 2500Hz to create a proactive greeting sound; the sound effect duration should be short and clear, ranging from 0.1 to 5 seconds. The sound effects can also include a double-note structure varying from low to high to simulate the tone of "Hi!", which is highly recognizable. The timing of the sound effect can be aligned with the moment the aircraft bounces up to its highest point or nods, so that the sound effect and the rhythm of the action are synchronized.

[0086] Curiosity behavior can be termed the fifth type of interaction behavior. For example, curiosity behavior may include one or more of the following: the height of the device first increases and then decreases, the head turns, the device and the head turn to the same side, the device tilts upward, the eyes are open, and the frequency of the sound emitted by the vocalization part rises.

[0087] For example, a robot's behavior in the context of curiosity can include one or more of the following: The robot can control its height to initially rise slightly to 16cm-24cm to express alertness and standing upright, then slowly descend to 12cm-16cm to express leaning forward for closer observation; control its head yaw to tilt to one side by ±0.2 to ±0.7 and hold for 0.8 to 5 seconds to display a head tilt to express curiosity; control its body tilt to coordinate with its head yaw in the same direction by ±0.08 to ±0.8, enhancing the overall effect of the head tilt; control its body pitch to initially tilt its head slightly upward by +0.05 to +0.6, then turn to tilt its head downward by -0.05 to -0.6, creating a look-up and down effect; control its forward / backward freedom to +0.02 to +0.3 to create a slow forward leaning motion, expressing the intention to get closer for observation. The head tilt can be slow and deliberate to create a playful feel, while the forward leaning motion can be slow to create a sense of caution. The robot can also be programmed to tilt its head to one side and then to the other, creating a multi-angle, repeated observation effect. Eye lighting effects for curiosity behaviors can include: setting the eye opening to 0.4 to 1.0 to express focused observation; setting the eye brightness to 0.4 to 1.0 to express the brightness brought by interest; controlling the pupil to perform rapid, skipping scans, with the pupil moving to its designated position in 80ms to 300ms; controlling the pupil to move towards the target direction 80ms to 250ms before the head yaw, creating a natural sequence of "looking at the target first, then turning the head"; and reducing the pupil's light spot radius to 0.2 to 0.7 to simulate pupil focusing and contraction, expressing high concentration. Sound effects for curiosity behaviors can include: using an upward sweeping sound effect, controlling the sound frequency to rise from 400Hz to 1000Hz to 1000Hz to 2500Hz, creating a short "Hmm?" "Questioning tone; the sound effect is extremely short, lasting from 0.08 seconds to 5 seconds, with the sound effect curve rising and accelerating at the end to enhance the questioning tone. The triggering time of the sound effect can be aligned with the start of the head tilting action, so that the sound effect appears synchronously with the discovery action."

[0088] Dizziness behavior can be referred to as the sixth type of interactive behavior. For example, dizziness behavior may include one or more of the following: large head shaking, body tilting and swaying, irregular pitching and swaying of the aircraft, irregular changes in aircraft height, slight body turning in place, irregular fluctuations in eye brightness, circular or irregular pupil movement, outputting downlink sound effects or outputting vibrating sound effects, or one or more of these.

[0089] For example, the robot's behavior in the dizziness scenario could include one or more of the following: The robot could control its head yaw to make large-amplitude swaying movements of ±0.2 to ±0.9, with a frequency of 0.4 Hz to 5 Hz and linear decay; control its body tilt to make unbalanced swaying movements of ±0.2 to ±0.8, with the body tilt and head yaw having different frequencies, so that the frequency difference superimposed to create a non-periodic sense of disorder; for example, the head yaw frequency is 0.8 Hz and the body tilt frequency is 1.1 Hz; control the body pitch to make irregular back-and-forth swaying movements of ±0.15 to ±0.8, expressing staggering; control the body height to fluctuate irregularly between 8 cm and 31 cm, to express fluctuating center of gravity, indicating instability and weak knees; control the body turning to make small, stationary turns of ±0.03 to ±0.5 to express a feeling of disorientation. The amplitude of each motion channel can decay linearly over time, gradually recovering from large-amplitude swaying to stability, simulating the process of dizziness subsiding; the movements can use elastic easing to produce a large-amplitude oscillation that gradually converges. The eye lighting effects in dizziness can include: setting the eye opening to 0.1 to 0.7, i.e., a semi-closed state, to express dazedness and an inability to open the eyes, and controlling the eye brightness to fluctuate irregularly between 0.1 and 1.0 to express blurred vision and unstable brightness; controlling the pupil to perform circular or irregular movements to express the classic dizzying visual manifestation of eye rotation; wherein, the circular movement of the pupil can satisfy the horizontal position x=sin(t), the vertical position y=cos(t), and the frequency 0.5Hz to 2.0Hz. The sound effects in dizziness can include: using a downward sweep sound effect or a composite sound effect, controlling the sound effect frequency from 800Hz to 1800Hz down to 200Hz to 600Hz to convey a sense of loss of control and sinking; the sound effect can be a moderate groan with a duration of 0.2 seconds to 5 seconds; the tremor frequency can be 3Hz to 10Hz to express a trembling and unstable timbre; the amplitude modulation depth can be 0.2 to 0.6 to make the volume fluctuate and create a sense of fluctuation. The sound effect can be triggered at the moment of greatest shaking, so that the sound effect and the rhythm of the movement are synchronized, expressing that the sound is emitted at the moment of greatest dizziness.

[0090] In some embodiments, the robot includes a body and a head. When the first perceptual information acquired by the robot includes a preset audio, the first action includes first rotating the head towards the direction of the preset audio, and then rotating the body towards the direction of the audio. For example, if the robot's first perceptual information includes a preset audio, performing a greeting or curiosity action involves first rotating the head towards the direction of the preset audio, and then rotating the body towards the direction of the audio. This can improve the robot's response speed and intuitiveness to preset audio, enhance the naturalness, smoothness, and stability of robot movements, and strengthen the robot's anthropomorphic or biomimetic interactive effects.

[0091] In some implementations, the target behavior corresponding to the robot's current perception information can be determined based on a pre-defined correspondence between perception information and target behavior, and the robot can be controlled to execute the target behavior. However, this is not the only option. For example, in other implementations, the robot's current perception information can be input into a behavior decision model, and the target behavior can be determined based on the output of the behavior decision model, and the robot can be controlled to execute the target behavior.

[0092] In some implementations, the robot control method may include an interactive sensing step, an interactive decision-making step, and an interactive execution step. For example, the interactive sensing step is used to acquire the robot's perception information, the interactive decision-making step is used to determine the target behavior corresponding to the perception information, and the interactive execution step is used to control the robot to perform the target behavior determined by the interactive decision-making step.

[0093] In some implementations, the interactive decision-making step may also determine whether the robot responds to the perceived information based on the perceived information, and determine the corresponding target behavior when it is determined that the robot will respond to the perceived information.

[0094] In some implementations, please refer to Figure 1 The robot control method may include: acquiring first perception information; executing a first action; wherein the first action is one of a plurality of candidate actions corresponding to the first perception information.

[0095] The same perceptual information can correspond to multiple candidate behaviors. When a robot acquires the same perceptual information multiple times, it can determine different target behaviors from multiple candidate behaviors, thereby reducing the sense of repetition and mechanicalness caused by the robot performing the same behavior every time it acquires the same perceptual information.

[0096] For example, the multiple candidate behaviors corresponding to the first perception information include the first behavior and the second behavior. When the first perception information is obtained, the first probability of executing the first behavior is different from the second probability of executing the second behavior.

[0097] For example, different candidate behaviors have different execution probabilities, which means that when the robot obtains the same perceptual information multiple times, it does not have to execute the same behavior in a fixed way. Instead, it can call one of the candidate behaviors as the target behavior according to the preset probability distribution of multiple candidate behaviors, which can improve the variability, rhythm rationality and naturalness of the robot's behavior output.

[0098] For example, each piece of sensory information can be mapped to a corresponding behavior pool. The behavior pool corresponding to the sensory information can include multiple candidate behaviors, and each of the multiple candidate behaviors can have a corresponding execution probability. After the robot obtains the corresponding sensory information, it can determine a candidate behavior as the target behavior from the behavior pool according to the execution probability of each candidate behavior in the behavior pool to which the sensory information is mapped.

[0099] For example, the sum of the execution probabilities of all candidate behaviors in the behavior pool can be 100%, wherein the execution probabilities of at least two candidate behaviors can be different, so that when the first perception information is obtained, the first probability of the robot executing the first behavior and the second probability of executing the second behavior can be different.

[0100] For example, multiple candidate behaviors corresponding to the same perceptual information can be of the same type. For instance, multiple candidate behaviors might all be a greeting action corresponding to the perceptual information of waving. The behavioral parameters of these multiple candidate behaviors corresponding to the same perceptual information can be different. The robot can execute behaviors of the same type with different specific manifestations when it receives the same perceptual information multiple times; this improves the stability and richness of the robot's interactive feedback.

[0101] For example, waving can serve as a form of interactive sensing information, and the interactive sensing identifier corresponding to the wave could be, for example, 101. Interactive sensing identifier 101 can be mapped to a greeting behavior pool, which can include multiple candidate greeting behaviors, such as 10 greeting behaviors. Each of the 10 greeting behaviors can have a corresponding execution probability, and the sum of the execution probabilities of all greeting behaviors in the greeting behavior pool can be 100%. Optionally, at least two of the 10 greeting behaviors may have different execution probabilities. After recognizing a wave, the robot can, based on the execution probabilities corresponding to each of the 10 greeting behaviors, select a greeting behavior from the greeting behavior pool as the target behavior.

[0102] In some implementations, the execution probability of a candidate behavior can be determined based on its behavioral performance, intensity, and / or applicable scenario. For example, a candidate behavior with greater noise or movement may have a lower execution probability, while a candidate behavior with less noise or movement may have a higher execution probability. For instance, behaviors with strong performance, large movements, or noticeable sound effects can be assigned a lower execution probability, while behaviors with milder performance, smaller movements, or smoother movements have a higher execution probability. This can prevent the robot from frequently executing high-intensity behaviors, which could lead to abrupt interaction rhythms, and can also improve the naturalness and richness of the robot's behavioral output.

[0103] In some implementations, the first probability of performing the first behavior is related to a behavior parameter of the first behavior, which indicates the magnitude of the first behavior.

[0104] For example, the execution probability of a candidate behavior can be related to the magnitude of the candidate behavior's activity level.

[0105] For example, the execution probability of a candidate behavior can be related to the behavior parameters of the candidate behavior, which can indicate the magnitude of the dynamics of the candidate behavior.

[0106] For example, the behavioral parameters may include at least one of the following: movement amplitude, movement speed, movement acceleration, movement duration, volume, brightness, light flashing frequency, light effect duration, and posture deflection angle. For instance, candidate behaviors with larger movement amplitude, faster movement speed, louder volume, higher brightness, or larger posture deflection angles exhibit greater noise; conversely, candidate behaviors with smaller movement amplitude, slower movement speed, lower volume, lower brightness, or smaller posture deflection angles exhibit less noise. The corresponding execution probability can be determined based on the behavioral parameters of the candidate behaviors.

[0107] For example, the first probability is negatively correlated with the magnitude of the first action. Specifically, the first probability of performing the first action is negatively correlated with the action parameters of the first action.

[0108] For example, the execution probability of a candidate behavior can be negatively correlated with the magnitude of its noise level. The greater the noise level of a candidate behavior, the lower its execution probability; conversely, the smaller the noise level, the higher its execution probability. This reduces the probability of frequently executing candidate behaviors with high noise levels, enabling the robot to generate different responses when repeatedly acquiring the same perceptual information, while also preventing abrupt changes in the interaction rhythm caused by frequent occurrences of high-intensity behaviors.

[0109] In some implementations, please refer to Figure 1 If the robot is performing an action corresponding to perception information with a priority of level two, the priority of the perception information corresponding to the first action is higher than or equal to level two.

[0110] For example, the robot control method includes: acquiring first perception information; and, when the robot is performing a second action corresponding to second perception information, controlling the robot to perform a first action corresponding to the first perception information if the priority of the first perception information is higher than or equal to the priority of the second perception information.

[0111] For example, when the robot is executing behavior A corresponding to perception information A with priority level 2, if the priority of the first perception information is level 3, and the first behavior is the behavior corresponding to the first perception information, the priority of the level 3 perception information is higher than that of the level 2.

[0112] For example, if a robot is executing a second action corresponding to second perception information with priority 2, and then acquires first perception information with priority 3, since the priority of the first perception information is higher than that of the second perception information, the robot can stop executing the second action and can execute the first action corresponding to the first perception information. Similarly, if a robot is executing a second action corresponding to second perception information with priority 3, and then acquires first perception information with priority 3, since the priority of the first perception information is equal to that of the second perception information, the robot can stop executing the second action and can execute the first action corresponding to the first perception information. Enabling perception information with a priority higher than or equal to the priority of the current action to promptly trigger the corresponding target action improves the timeliness of the robot's response to higher-priority current perception information and the continuity of the interactive response.

[0113] For example, when the robot is executing a second action corresponding to second perception information, if the priority of the first perception information is lower than the priority of the second perception information, the robot is controlled to continue executing the second action. For instance, when the robot is executing action A corresponding to perception information A with a priority of level two, if the priority of the first perception information is level one, and the first action is action A, the priority of the level one perception information is lower than that of the level two. For instance, when the robot is executing a second action corresponding to second perception information with a priority of 2 or 3, if the robot obtains first perception information with a priority of 1, since the priority of the first perception information is lower than that of the second perception information, the robot can continue executing the second action corresponding to the second perception information. This prevents low-priority perception information from interrupting the robot's execution of actions corresponding to higher-priority perception information.

[0114] For example, the perception information with priority at the second level includes interactive information about the environment perceived by the robot; and / or the perception information with priority at the second level includes at least one of unconscious user interaction information recognized by the robot and conscious user interaction information recognized by the robot; and / or the perception information with priority at the second level includes interactive information about another robot recognized by the robot.

[0115] For example, the interactive information of the environment perceived by the robot can be information related to the environment that the robot detects. This interactive information may include environmental changes perceived by the robot through one or more of the following: a camera module, a microphone, a distance sensor, a touch sensor, an inertial sensor, and a ground detection sensor. For example, the interactive information perceived by the robot may include one or more of the following: obstacle information, moving object information, sound source information, sound source direction information, lighting changes, ground surface information, ground height changes, wall or furniture information, target object information, touch information, vibration information, collision information, spatial area information, restricted area information, and charging dock location information.

[0116] For example, unconscious user interaction information can be information that the user does not actively express a clear control intention to the robot, but which can be recognized by the robot and used to trigger interactive behavior. For example, unconscious user interaction information can be weak human-computer interaction perception information, and further, it can include at least one of the following: the robot detects someone in front of it, the robot detects someone passing by, the robot hears someone speaking, the robot detects a target object in the environment, or the robot detects a non-command sound or action.

[0117] For example, a user's conscious interaction information can be information that the user actively sends to the robot with a clear interactive intention. For instance, the user's conscious interaction information can be strong human-computer interaction perception information, and further, it can include at least one of the following: the user makes a preset gesture, the user waves their hand, the user extends a static palm, the user touches the robot, the user turns the robot's head, the user presses the robot's head, the user pushes the robot, the user repeatedly pushes the robot, the user picks up the robot, the robot hears a wake word, the robot hears a praising voice, the robot hears a blaming voice, the robot hears laughter, the robot hears crying, the robot hears a dog barking, the robot hears a loud noise, the robot sees a red object, the robot sees a person jumping in front of it, the user deletes photos from the album, or the user views a highlight moment in a video in the album.

[0118] For example, the interaction information of the other robot includes one or more of the following: the other robot's identity information, location information, distance information, movement direction information, movement state information, posture information, lighting effect information, sound information, task status information, interaction request information, collaborative task information, avoidance request information, follow request information, formation request information, and charging queue information.

[0119] For example, the perception information with priority at level two includes unconscious user interaction information recognized by the robot, while the perception information with priority at level three includes conscious user interaction information recognized by the robot. The third level has a higher priority than the second level. By prioritizing conscious user interaction information over unconscious user interaction information, the robot can prioritize responding to explicit user commands when simultaneously recognizing both natural user behavior and explicit user instructions. This prevents unconsciously triggered interactions from occupying or interfering with the execution of interactions corresponding to explicit commands, thereby improving the accuracy, timeliness, and matching degree of the robot's interaction response with the user's intent.

[0120] For example, the perception information with a priority of first level includes the robot's internal resting perception information. The priority of the first level is lower than that of the second level and / or the third level. When the robot only acquires its internal resting perception information, it can perform a resting behavior. When the robot performs a resting behavior, if it recognizes at least one of the following: environmental interaction information, unconscious user interaction information, conscious user interaction information, or interaction information from another robot, it can respond promptly to prevent the resting behavior from interfering with the execution of the behaviors corresponding to the environmental interaction information, unconscious user interaction information, conscious user interaction information, or interaction information from another robot, thereby improving the timeliness and controllability of the robot's interactive response.

[0121] In some implementations, the first sensing information has a system-level priority, and the first action is a system-level action corresponding to the first sensing information. The system-level action includes adjusting the robot's working state or mode. System-level sensing information may include system state sensing information, which can be sensing information related to changes in the robot's own system state, working state, or working mode. For example, as long as the first sensing information with a system-level priority is acquired, the system-level action corresponding to the first sensing information can be executed. This prevents lower-priority sensing information from interfering with the robot's response to the system-level first sensing information and its adjustment of its working state or mode, thereby improving the timeliness of the robot's response in abnormal or critical states.

[0122] In some implementations, the sensory information that the robot can acquire may include one or more of the following: resting sensory information, weak human-computer interaction sensory information, strong human-computer interaction sensory information, and system state sensory information.

[0123] For example, resting perception information can be internal state-related perception information generated by the robot when it does not detect explicit user interaction intent or external unexpected events. For instance, resting perception information can include at least one of the following: the robot has not detected human-robot interaction for a preset period of time, the robot is in standby mode, the robot is idle, or the robot is in a low-activity operating state. Resting perception information can be used to trigger resting behaviors in the robot, such as observing its surroundings, breathing-like movements, slight head tilting, natural blinking, or slow pupil drift.

[0124] For example, weak human-machine interaction (HMI) perception information can be information that the robot detects related to the user or environment, but this information usually does not have a clear user-initiated interaction purpose, or does not require the robot to make an immediate strong response. Optionally, weak HMI perception information may include at least one of the following: the robot detects someone in front of it, the robot detects someone passing by, the robot hears someone speaking, the robot detects a target object in the environment, or the robot detects a non-command sound or movement. Weak HMI perception information can be used to trigger weak interactive behaviors of the robot, such as staring at a person in front of it, turning its head or eyes towards a target, or briefly gazing at a target object.

[0125] For example, strong human-computer interaction (HCI) perception information can be user-triggered, with a clear interaction purpose, or requiring a clear response from the robot. Optionally, HCI perception information can include at least one of the following: user making a preset gesture, user waving, user extending a static palm, user touching the robot, user turning the robot's head, user pressing the robot's head, user pushing the robot, user repeatedly pushing the robot, user picking up the robot, robot hearing a wake-up word, robot hearing praise, robot hearing blame, robot hearing laughter, robot hearing crying, robot hearing a dog barking, robot hearing a loud noise, robot seeing a red object, robot seeing a person jumping in front of it, user deleting photos from an album, or user viewing highlights in a video album. HCI perception information can be used to trigger relatively specific target behaviors of the robot, such as greeting, following, happiness, sadness, curiosity, dizziness, fear, cheering, startled, aggrieved, or functional execution.

[0126] For example, system status awareness information can be awareness information related to changes in the robot's own system status, working status, or working mode. Optionally, system status awareness information can include at least one of the following: robot power-on, robot power-off, robot standing up, robot storage, robot low battery, robot switching to shooting mode, robot exiting shooting mode, robot entering charging state, robot completing a preset task, robot malfunctioning, or robot entering a protection state. System status awareness information can be used to trigger system-level behaviors of the robot, such as power-on behavior, power-off behavior, wake-up behavior, termination behavior, low battery behavior, mode switching behavior, fault indication behavior, or protection behavior.

[0127] For example, the priority of the resting perception information is lower than that of the weak human-computer interaction perception information; and / or the priority of the weak human-computer interaction perception information is lower than that of the strong human-computer interaction perception information; and / or the priority of the strong human-computer interaction perception information is lower than that of the system status perception information. For instance, the resting perception information may correspond to a first-level priority (e.g., priority 1), which can be the lowest priority; the weak human-computer interaction perception information may correspond to a second-level priority (e.g., priority 2); the strong human-computer interaction perception information may correspond to a third-level priority (e.g., priority 3); and the system status perception information may correspond to a fourth-level priority (e.g., priority 4), which can be the highest priority.

[0128] During the continuous acquisition of sensory information, the robot can determine the order of behavioral responses based on the category and priority of the sensory information. For example, the robot can prioritize responding to more important sensory information such as changes in system state and user-initiated interactions in order to execute the corresponding target behavior. It can also prevent low-priority sensory information, such as resting sensory information, from interrupting the robot's execution of behaviors corresponding to higher-priority sensory information.

[0129] In some implementations, the robot control method includes: acquiring first perception information; and, if the first perception information and the second perception information are pre-set to be interruptible, controlling the robot to execute a first behavior corresponding to the first perception information when the robot is executing a second behavior corresponding to the second perception information. For example, when the robot is continuously acquiring perception information and executing an existing behavior, it can determine whether to switch to the first behavior corresponding to the first perception information based on the interruption condition between the newly acquired first perception information and the currently responding second perception information; this can prevent the robot from switching behaviors in an unordered manner when continuously acquiring multiple perception information, and improve the rationality of the robot's response order to perception information under multiple perception information.

[0130] For example, the interruption condition may include one or more of the following: the priority of the first sensing information is higher than the priority of the second sensing information; the priority of the first sensing information is equal to the priority of the second sensing information; the first sensing information and the second sensing information are pre-set to be interruptible.

[0131] For example, when the robot is performing the second action corresponding to the second perception information, if the first perception information and the second perception information are pre-set to be interruptible, the robot is controlled to stop performing the second action and perform the first action corresponding to the first perception information.

[0132] Optionally, the interruptibility relationship between the first perception information and the second perception information can be a unidirectional interruptibility relationship. For example, if the robot is performing a second action corresponding to the second perception information, and the first perception information is acquired, then the first perception information can interrupt the second action corresponding to the second perception information; if the robot is performing a first action corresponding to the first perception information, and the second perception information is acquired, then the second perception information cannot interrupt the first action corresponding to the first perception information.

[0133] For example, a one-way interruptible relationship can be established between the perceptual information of accusatory speech and the perceptual information of happiness. For instance, if a robot recognizes an accusatory speech while performing a happy or excited behavior, it can stop the current happy or excited behavior and execute the target behavior corresponding to the accusatory speech, such as stopping the current behavior and reducing its activity level for a period of time afterward. Conversely, if a robot acquires perceptual information of happiness while performing a low-activity or silent behavior corresponding to an accusatory speech, it can choose not to immediately interrupt the current low-activity behavior. Perceptual information with stronger control or corrective significance can be prioritized for response.

[0134] Optionally, the interruptibility relationship between the first perception information and the second perception information can be bidirectional. For example, if the robot is performing a second action corresponding to the second perception information, and the first perception information is acquired, the first perception information can interrupt the second action corresponding to the second perception information; and if the robot is performing a first action corresponding to the first perception information, and the second perception information is acquired, the second perception information can also interrupt the first action corresponding to the first perception information.

[0135] For example, the perception information of praise and criticism can be set as a bidirectional interruptible relationship. For instance, if the robot is performing a happy or excited behavior corresponding to a praise voice and then recognizes a criticism voice, it can stop the happy or excited behavior and perform the target behavior corresponding to the criticism voice; similarly, if the robot is performing the target behavior corresponding to a criticism voice and then recognizes a praise voice, it can also stop the current behavior and perform the happy or excited behavior corresponding to the praise voice. This allows the robot to respond promptly to both consecutive positive and negative voices.

[0136] In some implementations, when the robot is executing behavior A corresponding to perception information A with a priority of level two, if the priority of the first perception information is level two, the first behavior is either behavior A or behavior B, where behavior B is the behavior corresponding to the first perception information. The interruption condition may include: the priority of the first perception information is equal to the priority of the second perception information. This can prevent behaviors corresponding to current perception information of the same priority from occupying the robot's execution time, improving the robot's timeliness in responding to behaviors corresponding to new perception information.

[0137] For example, a robot control method includes: acquiring first perception information; and, when the robot is executing a second behavior corresponding to second perception information, if the priority of the first perception information is equal to the priority of the second perception information, and the first and second perception information are pre-set to be interruptible, controlling the robot to execute a first behavior corresponding to the first perception information. For example, when the robot is executing a second behavior corresponding to second perception information, if the priority of the first perception information is equal to the priority of the second perception information, and the first and second perception information are pre-set to be interruptible, controlling the robot to stop executing the second behavior and start executing the first behavior corresponding to the first perception information. For instance, a robot control method may include: when the robot is executing behavior A corresponding to perception information A with a priority of level two, if the acquired first perception information has a priority of level two, and perception information A and the first perception information are pre-set to be interruptible, controlling the robot to execute the target behavior corresponding to the first perception information. More refined behavior switching rules can be established between perception information of the same priority, improving the rationality of decision-making when perception information of the same priority is triggered concurrently or continuously. For example, when a robot is executing a behavior corresponding to ordinary Level 3 interactive sensory information, if it receives sensory information of the same Level 3 but pre-set to be interruptible, such as a praising voice, a blaming voice, or a wake word, the robot can stop its current behavior and switch to execute the behavior corresponding to the new sensory information. This prevents behaviors corresponding to sensory information of the same priority from occupying the robot's execution time, improving the robot's responsiveness to behaviors corresponding to new sensory information. For example, even if the first and second sensory information have the same priority, if the first sensory information is not set to interrupt the second sensory information, the robot can continue to execute the second behavior, preventing sensory information of the same priority from unconditionally interrupting each other and causing the robot to frequently switch behaviors.

[0138] For example, one or more priorities can be divided into multiple sub-priorities. For instance, a third-level priority might include a first sub-priority and a second sub-priority. The first sub-priority could be called a third-level ordinary priority (e.g., priority 3), and the second sub-priority could be called a third-level enhanced priority (e.g., priority 3+). Understandably, second-level, fourth-level, or other priorities can also be divided into multiple sub-priorities according to actual needs. By setting sub-priorities, the response order and interruption relationship of different sensing information can be further refined within the same priority level.

[0139] For example, perception information of a second sub-priority of the same priority can interrupt perception information of a first sub-priority of the same priority, such as perception information of a third-level enhanced priority interrupting perception information of a third-level normal priority, or vice versa; or perception information of a first sub-priority of the same priority cannot interrupt perception information of a second sub-priority of the same priority, such as perception information of a third-level normal priority not interrupting perception information of a third-level enhanced priority. For example, if a robot is executing the behavior corresponding to the perception information of a third-level normal priority, and then acquires perception information of a third-level enhanced priority, the robot can be controlled to stop executing the behavior corresponding to the perception information of a third-level normal priority and execute the behavior corresponding to the perception information of a third-level enhanced priority. For example, if a robot is executing the behavior corresponding to the perception information of a third-level enhanced priority, and then acquires perception information of a third-level normal priority, the robot can continue executing the behavior corresponding to the perception information of a third-level enhanced priority.

[0140] For example, multiple second-level sub-priority perceptual information of the same priority can interrupt each other. For instance, any perceptual information of the third-level enhanced priority can interrupt any other perceptual information of the third-level enhanced priority. This allows the robot to promptly switch to the behavior corresponding to the most recently acquired enhanced interactive perceptual information when continuously acquiring wake words, praise voices, or accusatory voices, thereby improving the timeliness of response to key interactive events.

[0141] For example, any type of perception information of the first sub-priority of the same priority cannot interrupt any other type of perception information of the first sub-priority of the same priority. This can prevent robots from frequently switching behaviors when performing ordinary strong interactive behaviors due to the continuous appearance of ordinary perception information of the same level, thus preventing and reducing problems such as action jumps, incomplete behavior execution, or chaotic interaction rhythm, and improving the continuity and stability of robot behavior execution.

[0142] In some implementations, if the first sensing information is obtained again within a first time period after it was first obtained, the robot will not respond to the first sensing information. The robot control method can limit the repeated triggering of corresponding behaviors by the robot with the same sensing information within a short period of time.

[0143] For example, if the same perceptual information is acquired again within a first time period after the robot executes the behavior corresponding to that perceptual information, the robot may not respond to the acquired same perceptual information again, at least within that first time period; that is, the robot does not need to restart the execution of the behavior corresponding to the acquired same perceptual information within that first time period. When the same perceptual information appears repeatedly in a short period of time, by limiting the robot to repeatedly execute the same behavior, the repetitive movements, execution jitter, frequent switching, twitching, or mechanical feeling caused by the continuous triggering of the same behavior can be reduced.

[0144] For example, after acquiring certain perceptual information and controlling the robot to execute the behavior corresponding to that perceptual information for a first time period, if the robot acquires the same perceptual information again, it can respond to that perceptual information again and control the robot to execute the behavior corresponding to that perceptual information again. That is, after the first time period ends, the robot can still respond to that perceptual information again, which can achieve a balance between suppressing short-term repeated triggering and preserving normal interactive response capabilities, thereby improving the naturalness of the robot's response rhythm and the stability of its behavioral output.

[0145] For example, the first duration range corresponding to at least two different types of perceived information can be different. For instance, the first duration range corresponding to perceived information can be determined based on one or more of the following: the type of perceived information, the trigger frequency, the intensity of the response behavior, the duration of the response behavior, and the importance of the interaction.

[0146] This application provides a robot control method, including: acquiring first perception information; executing a first behavior; wherein the first behavior is one of a plurality of candidate behaviors corresponding to the first perception information; and / or if the robot executes a behavior corresponding to perception information with a priority of second level, the priority of the perception information corresponding to the first behavior is higher than or equal to the second level. This enables the robot to present more natural and richer interactive effects.

[0147] Please refer to the above embodiments. Figure 3 , Figure 3 This is a schematic block diagram of a control device 60 provided in an embodiment of this application. The control device 60 includes at least one processor 61; and a memory 62 communicatively connected to the at least one processor 61; wherein the memory 62 stores instructions executable by the at least one processor 61, which, when executed by the at least one processor 61, enable the at least one processor 61 to perform the method described in the embodiment of this application.

[0148] In some embodiments, the control device 60 can be a control device mounted on the robot, such as a control device located on the robot's head, body, chassis, or other mounting location. In other embodiments, the control device 60 can be a control device separate from the robot, capable of communicating with the robot; for example, the control device 60 can be a server, cloud device, mobile terminal, remote control device, edge computing device, or other external control device. In still other embodiments, the control device 60 can include both a control device on the robot and a control device separate from the robot, both of which can jointly implement the control method of the embodiments of this application.

[0149] The specific principles and implementation methods of the control device provided in this application are similar to those of the methods in the foregoing embodiments, and will not be repeated here. Furthermore, any parts not mentioned in this application's embodiments can be referred to the relevant descriptions in the foregoing method embodiments, and will not be repeated here.

[0150] Please refer to the above embodiments. Figure 4 , Figure 4 This is a schematic block diagram of a robot 70 provided in an embodiment of this application. The robot 70 includes the aforementioned control device 60.

[0151] The specific principles and implementation methods of the robot provided in this application are similar to those in the foregoing embodiments, and will not be repeated here. Furthermore, any parts not mentioned in this application can be referred to the relevant descriptions in the foregoing method embodiments, and will not be repeated here.

[0152] This application also provides a computer-readable storage medium storing computer-executable instructions that, when executed, can perform the aforementioned method.

[0153] The computer-readable storage medium can be an internal storage unit of the control device or robot described in any of the foregoing embodiments, such as the hard drive or memory of the control device or robot. The computer-readable storage medium can also be an external storage device of the control device or robot, such as a plug-in hard drive, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the control device or robot.

[0154] It should be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of the application.

[0155] It should also be understood that the term “and / or” as used in this application and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0156] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for controlling a robot, characterized in that, The method includes: Obtain first-hand sensory information; Perform the first action; Wherein, the first behavior is one of a plurality of candidate behaviors corresponding to the first perceived information; and / or The priority of the sensing information corresponding to the first behavior is higher than or equal to the priority of the first sensing information.

2. The method according to claim 1, characterized in that, The first behavior is one of a plurality of candidate behaviors corresponding to the first perceived information. The plurality of candidate behaviors include the first behavior and the second behavior. When the first perceived information is obtained, the first probability of executing the first behavior is different from the second probability of executing the second behavior.

3. The method according to claim 1 or 2, characterized in that, The first behavior is one of a plurality of candidate behaviors corresponding to the first perceived information. The first probability of executing the first behavior is related to the behavior parameters of the first behavior, which can indicate the magnitude of the first behavior.

4. The method according to claim 3, characterized in that, The first probability is negatively correlated with the magnitude of the movement or stillness of the first action; and / or The behavioral parameters include one or more of the following: amplitude of movement, speed of movement, acceleration of movement, duration of movement, volume, brightness, frequency of light flashing, duration of light effect, and at least one of the following: attitude deflection angle.

5. The method according to any one of claims 1-4, characterized in that, When the robot performs action A corresponding to perceived information A, If the priority of the first sensing information is higher than the priority of the sensing information A, the first behavior is the behavior corresponding to the first sensing information; and / or If the priority of the first perceived information is lower than the priority of the perceived information A, the first behavior is behavior A; and / or If the priority of the first sensing information is the same as the priority of the sensing information A, the first behavior is either behavior A or behavior B, where behavior B is the behavior corresponding to the first sensing information.

6. The method according to any one of claims 1-5, characterized in that, The method further includes: If the first sensing information is obtained again within a first time period after the first sensing information is obtained, the first sensing information will not be responded to.

7. The method according to any one of claims 1-6, characterized in that, The sensory information that the robot can obtain includes one or more of the following: Resting perception information, weak human-computer interaction perception information, strong human-computer interaction perception information, and system status perception information; Wherein, the priority of the resting perception information is lower than the priority of the weak human-computer interaction perception information; and / or, The priority of the weak human-computer interaction perception information is lower than the priority of the strong human-computer interaction perception information; and / or The priority of the human-computer interaction perception information is lower than the priority of the system state perception information.

8. The method according to any one of claims 1-7, characterized in that, The robot includes a body, head, eyes, and a voice-emitting part; The behaviors corresponding to different perceptual information include one or more of the following categories: The first type of interactive behavior includes one or more of the following: the body of the device undulates within a first height range, the body tilts upward, the eyes squint, the brightness of the eyes changes within a first brightness range, and the sound output frequency of the sound output part is within a first frequency range. The second type of interactive behavior includes one or more of the following: the height of the body decreases, the body lowers, the eyes are half-closed or closed, the brightness of the eyes decreases, and the sound frequency of the sound-emitting part decreases. The third type of interactive behavior includes one or more of the following: the body of the device undulates within a first frequency range, the body of the device tilts within a first frequency range, and the brightness of the eyes changes within a first frequency range. The fourth type of interactive behavior includes one or more of the following: the robot jumps up, the body undulates, the body swings, the head faces the target direction, the eyes are open, the brightness of the eyes increases, and the frequency of the voice output increases. The fifth type of interactive behavior: The third type of interactive behavior includes one or more of the following: the height of the body first rises and then falls, the head turns, the body and the head turn to the same side, the body tilts upward, the eyes are open, and the sound frequency of the sound-emitting part rises.

9. The method according to any one of claims 1-8, characterized in that, The robot includes a body and a head. When the first perception information acquired by the robot includes a preset audio, the first action includes the head first turning towards the direction of the preset audio, and then the body turning towards the direction of the audio.

10. A control device, characterized in that, The control device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1-9.

11. A robot, characterized in that, The robot includes the control device as described in claim 10.

12. The robot according to claim 11, characterized in that, The robot includes body; A head, which is movably connected to the body, includes a shooting module; A leg mechanism connected to the body, the leg mechanism being used to drive the robot to move on the ground and / or adjust the posture of the body.

13. The robot according to claim 12, characterized in that, The leg mechanism is used to adjust the height difference between the left and right sides of the fuselage; and / or, The leg mechanism is used to adjust the height difference between the front and rear sides of the fuselage.

14. The robot according to claim 13, characterized in that, The leg mechanism includes at least a first leg mechanism and a second leg mechanism, with the first leg mechanism connected to the left side of the fuselage and the second leg mechanism connected to the right side of the fuselage; Wherein, at least one of the first leg mechanism and the second leg mechanism can be retracted or extended to adjust the height difference of the fuselage on the left and right sides; and / or, the fuselage can move relative to the first leg mechanism and the second leg mechanism to adjust the height difference of the fuselage on the front and rear sides.

15. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions that, when executed, enable the execution of the method described in any one of claims 1-9.

Images