CN-121981156-A - Robot psychology characteristic staged constraint control method

CN121981156ACN 121981156 ACN121981156 ACN 121981156ACN-121981156-A

Abstract

The application discloses a robot psychological characteristic staged constraint control method which comprises the steps of dividing a psychological growth stage of a robot, collecting man-machine interaction data and environment data in real time in the running process of the robot, comparing the collected data with preset judgment data, matching a working scene of the robot and determining the psychological growth stage where the robot is currently located, calling psychological characteristic parameters and constraint rules configured in the current psychological growth stage to carry out constraint control on the robot, obtaining interaction feedback data of the robot, and carrying out dynamic adaptation optimization on constraint parameters corresponding to the current psychological growth stage. The method has the beneficial effects that the psychological growth stage of the robot is dynamically judged by collecting the interactive data and the environment data in real time, and the dynamic adaptation of the psychological characteristics of the robot is realized according to the continuous optimization constraint parameters of the user feedback, so that the interactive requirements of different users and different scenes are met.

Inventors

YANG CHAO
YING GUOGANG
YANG YUCHEN
ZHANG YIMING
SHI KUN

Assignees

宁波朗达科技有限公司

Dates

Publication Date: 20260505
Application Date: 20260331

Claims (10)

1. The method for controlling the physical characteristics of the robot by restraint in stages is characterized by comprising the following steps of: S100, dividing psychological growth stages of the robot, configuring corresponding psychological characteristic parameters, constraint rules and dynamic adjustment thresholds for each psychological growth stage, and constructing a staged constraint control model; s200, collecting man-machine interaction data and environment data in real time in the running process of the robot, comparing the collected data with preset judgment data, matching the working scene of the robot and determining the current psychological growth stage; S300, invoking psychological characteristic parameters and constraint rules configured in the current psychological growth stage, and performing constraint control on at least one of emotion expression, behavioral response and decision logic of the robot; S400, acquiring interactive feedback data generated by the robot in the constraint control process, and dynamically adapting and optimizing constraint parameters corresponding to the current psychological growth stage based on the interactive feedback data.
2. The method for controlling the mental feature phasing constraint of a robot according to claim 1, wherein in step S100, the mental growth phase division of the robot is performed based on the growth psychology, including infancy, childhood, school age, puberty, and adulthood; The infant stage takes basic emotion response and passive interaction as cores, and psychological characteristics are focused on dependency and simple emotion expression; The active exploration and imitation learning is used as a core in childhood, and the psychological characteristics are focused on curiosity and imitation emotion and behavior; the learning age takes rule cognition and logic preliminary establishment as a core, and psychological characteristics focus on rule-type and rational preliminary expression; puberty takes the enrichment of emotion and autonomous decision germination as a core, and psychological characteristics focus on autonomous and multiple emotion expression; in adulthood, rational decision and emotion stabilization are taken as cores, and psychological characteristics focus on maturity, and emotion and decision logic stabilization.
3. The method of claim 1, wherein in step S100, the psychological characteristic parameters include an emotion intensity parameter, a behavioral response parameter, a decision hesitation parameter, and an interaction initiative parameter; The emotion intensity parameter is used for representing the emotion expression intensity range of the robot in different concentric growth phases; The behavior response parameters are used for representing the response speed of the robot to the user instruction in different concentric theory long stages; the interaction initiative parameter is used for representing the frequency of actively initiating interaction of the robot in different concentric theory growth stages; the decision hesitation parameter is used for representing the hesitation degree of the robot when making decisions in different theoretical growth stages, and the concrete expression of the decision hesitation parameter is as follows: ; Where H decision (S) represents the decision hesitation in the S-th mental growth stage, H base (S) represents the hesitation reference value in the S-th mental growth stage, δ represents the uncertainty coefficient, and S uncertainty represents the decision uncertainty score.
4. The method for controlling the constraint of the physical characteristics of the robot in stages according to claim 1, wherein in the step S100, the constraint rules include emotion expression threshold constraint, behavior response boundary constraint and decision logic priority constraint; the emotion expression threshold constraint is used for limiting emotion expression of the robot in the current psychological growth stage not to exceed a preset intensity range; the behavior response boundary constraint is used for limiting the behavior response speed of the robot; the decision logic priority constraint is used for representing the priority of the robot when making decisions in each psychological growth stage, and the specific expression is as follows: ; Where W decision (s) represents the decision-making comprehensive weight of the s-th mental growth stage, ω i (s) represents the weight coefficient of the i-th decision-making factor in the s-th mental growth stage, P i represents the score of the i-th decision-making factor, and n represents the total number of decision-making factors.
5. The method of claim 1-4, wherein in step S200, the human-computer interaction data includes at least one of user speech emotion, facial expression, interaction instruction, and interaction duration, and the environmental data includes at least one of environmental noise, illumination intensity, and interaction scene type; And extracting interaction scene features and user demand features from the man-machine interaction data and the environment data through the deep learning model, performing similarity calculation on the extracted features and scene features contained in preset judgment data, and judging robot work scene matching according to a similarity calculation result.
6. The method for controlling the mental feature phasing constraint of the robot according to claim 5, wherein in step S200, the mental growth phase of the robot is determined by a multi-dimensional weighted scoring method, which specifically comprises the following steps: Setting scoring dimension comprises at least one of interaction complexity, user demand level and accumulated interaction experience value of the robot; Distributing corresponding weights for each scoring dimension, and calculating comprehensive scores according to the distributed weights; and determining the current mental growth stage of the robot based on a scoring range defined for each mental growth stage in preset judgment data.
7. The method for controlling the physical characteristics of a robot in a phased constraint manner according to claim 6, wherein an initial mental growth phase is preset based on a current working scene of the robot; Triggering the robot to upgrade from the initial psychological growth stage to the next psychological growth stage if the calculated comprehensive score exceeds a stage upgrade threshold defined for the initial psychological growth stage in preset judgment data; And triggering the robot to degrade from the initial psychological growth stage to the previous psychological growth stage if the calculated comprehensive score is lower than a stage degradation threshold defined for the initial psychological growth stage in the preset judgment data.
8. The method for controlling the mental feature phasing constraint of the robot according to claim 7, wherein the matching degree of each mental growth stage corresponding to the robot and the current working scene is calculated based on scene factors of the current working scene, and the mental growth stage with the highest matching degree is selected as the initial mental growth stage of the robot in the current working scene; The calculation expression of the matching degree M init is as follows: ; Where t represents the total number of scene factors, γ j represents the weight of the j-th type of scene factors, and M j,scene represents the matching score of the j-th type of scene factors and each psychological growth stage.
9. The method of claim 1, wherein dynamically adjusting the threshold comprises feeding back an adaptation threshold and a scene adaptation threshold, and automatically adjusting the constraint parameters when the interactive feedback data reaches the set feeding back adaptation threshold and/or the scene change reaches the set scene adaptation threshold.
10. The method for controlling the mental feature phasing constraint of the robot according to claim 9, wherein in step S400, constraint parameters of the current mental growth phase are adjusted by using a gradient descent algorithm based on the interactive feedback data, and a specific adjustment formula of the constraint parameters is as follows: ; ; Wherein θ new (s) represents constraint parameters of the s-th psychological growth stage after adjustment, θ old (s) represents constraint parameters of the s-th psychological growth stage before adjustment, η represents learning rate, Representing the gradient of the parameter loss function J (θ (s)), m representing the number of samples fed back, F k,feedback representing the user actual feedback score for the kth sample, and F k,predict (θ (s)) representing the feedback prediction score for the kth sample based on the current constraint parameter θ(s).

Description

Robot psychology characteristic staged constraint control method Technical Field The application relates to the technical field of robots, in particular to a method for controlling the physical characteristics of a robot by restraint in stages. Background Along with the rapid development of intelligent robot technology, robots gradually enter a plurality of fields such as families, offices, education and the like, the depth and frequency of interaction with human-computer are continuously improved, and higher requirements are provided for emotion expression, naturalness of behavioral response and suitability of the robots. The psychological characteristic control of the existing robot mostly adopts a fixed parameter control mode, namely the emotion expression and behavior response logic of the robot are fixed, and cannot be dynamically adjusted according to interaction scenes, user demands and self interaction experience, so that the robot is seriously disjointed with the dynamic rule of human psychological growth. The psychological characteristics of human beings gradually develop along with the change of the growth stage, from dependent emotion and simple interaction in infancy to mature emotion and rational decision in adulthood, the psychological characteristics of different stages have obvious differences, and the existing robot lacks the constraint control of the psychological characteristics in stages, so that the problems of hard emotion response, non-conforming behaviors to human cognition habits, incapability of realizing emotion resonance and the like occur in the human-computer interaction process, and the human-computer interaction experience is seriously influenced. In addition, although emotion control mechanisms are introduced into the existing part of robots, the existing part of robots are not combined into a long psychology theory, and lack of scientific stage division and constraint rules, so that dynamic adaptation and personalized adjustment of psychological characteristics cannot be realized, and interaction requirements of different users and different scenes are difficult to meet. Therefore, how to realize the phased constraint control of the psychological characteristics of the robot based on the growth psychology rule, and improve the naturalness and the suitability of man-machine interaction becomes a technical problem to be solved urgently in the current robot control technical field. Disclosure of Invention One of the objects of the present application is to provide a method for controlling the mental feature of a robot in a phased manner, which solves at least one of the above-mentioned drawbacks of the prior art. In order to achieve at least one of the above purposes, the technical scheme adopted by the application is that the robot psychological characteristic staged constraint control method comprises the following steps: S100, dividing psychological growth stages of the robot, configuring corresponding psychological characteristic parameters, constraint rules and dynamic adjustment thresholds for each psychological growth stage, and constructing a staged constraint control model; s200, collecting man-machine interaction data and environment data in real time in the running process of the robot, comparing the collected data with preset judgment data, matching the working scene of the robot and determining the current psychological growth stage; S300, invoking psychological characteristic parameters and constraint rules configured in the current psychological growth stage, and performing constraint control on at least one of emotion expression, behavioral response and decision logic of the robot; S400, acquiring interactive feedback data generated by the robot in the constraint control process, and dynamically adapting and optimizing constraint parameters corresponding to the current psychological growth stage based on the interactive feedback data. Preferably, in step S100, the mental growth stages of the robot are divided based on growth psychology, including infancy, childhood, school age, puberty and adulthood, wherein the infancy uses basic emotion response and passive interaction as cores, the psychological characteristic is emphasized and simple emotion expression, the childhood uses active exploration and imitation learning as cores, the psychological characteristic is emphasized and curiosity and imitation emotion and behavior, the school age uses regular cognition and logic preliminary establishment as cores, the psychological characteristic is emphasized and regular and rational preliminary expression, the puberty uses emotion enrichment and autonomous decision germination as cores, the psychological characteristic is emphasized and self-dominant and multi-emotion expression, the adulthood uses rational decision and emotion stabilization as cores, and the psychological characteristic is emphasized and mature and stable emotion and decision logic. Preferably, in st