Search

CN-121999663-A - Multi-virtual-person simulation teaching method and related device

CN121999663ACN 121999663 ACN121999663 ACN 121999663ACN-121999663-A

Abstract

The application discloses a multi-virtual person simulation teaching method and a related device, which relate to the field of virtual teaching, and are used for acquiring multi-modal interaction data of a target object in a virtual classroom, determining a target dimension influenced by the multi-modal interaction data, updating a state value of a virtual student in the target dimension according to the multi-modal interaction data, driving the virtual student to make related behavior feedback based on the updated state value, acquiring the multi-modal interaction data again until the teaching is finished, and generating teaching evaluation data according to the multi-modal interaction data and the behavior feedback acquired in the teaching process. According to the application, only the state value of the affected target dimension is updated, the efficiency of the class-receiving state jump is improved, the training scene is more consistent with the real teaching situation, the teaching is safer and more controllable in the virtual training scene, the assessment is performed based on objective data in the teaching process, and the objectivity and the credibility of teaching assessment data are improved.

Inventors

  • YAN CHENXI
  • LIU YANG
  • JIANG MANMAN
  • Lv xing
  • LIU CHANG

Assignees

  • 科大讯飞股份有限公司

Dates

Publication Date
20260508
Application Date
20260325

Claims (17)

  1. 1. The multi-virtual-person simulation teaching method is characterized by comprising the following steps of: acquiring multi-mode interaction data of a target object in a virtual classroom, wherein initial class-listening state vectors of a plurality of virtual students in the virtual classroom are determined by preset class portraits, and the initial class-listening state vectors comprise state values of a plurality of dimensions; determining a target dimension influenced by the multi-modal interaction data, wherein the target dimension is part or all of the multiple dimensions; Updating the state value of the virtual student under the target dimension according to the multi-modal interaction data to drive the virtual student to make relevant behavior feedback based on the updated state value, and returning to the step of acquiring the multi-modal interaction data until the teaching is finished; And generating teaching evaluation data according to the multi-mode interaction data and the behavior feedback obtained in the teaching process.
  2. 2. The multi-virtual human simulated lecture method according to claim 1, wherein the updating the status value of the virtual student in the target dimension according to the multi-modal interaction data includes: Acquiring individual parameters of the virtual students, wherein the individual parameters are parameters describing individual difference characteristics of the virtual students, and the individual difference characteristics at least comprise attribute characteristics for distinguishing classroom behavior responses of the students; And updating the state value of the virtual student in the target dimension according to the multi-mode interaction data and the personality parameters of the virtual student.
  3. 3. The multi-virtual simulated lecture method according to claim 2, further comprising: Determining a first virtual student affected by the multi-modal interaction data, the first virtual student being part or all of the plurality of virtual students; the updating the status value of the virtual student in the target dimension according to the multi-modal interaction data and the personality parameters of the virtual student to drive the virtual student to make relevant behavior feedback based on the updated status value comprises: And updating the state value of the first virtual student under the target dimension according to the multi-mode interaction data and the personality parameters of the first virtual student so as to drive the first virtual student to make relevant behavior feedback based on the updated state value.
  4. 4. A multi-virtual human simulated lecture method according to claim 3, wherein said updating the status value of the first virtual student in the target dimension according to the multi-modal interaction data and the personality parameters of the first virtual student to drive the first virtual student to make relevant behavioral feedback based on the updated status value comprises: Inputting the multi-modal interaction data, the personality parameters of the first virtual students and the state values of the first virtual students in the target dimension into the sub-model corresponding to the target dimension to obtain updated state values in the target dimension; Generating a next class-listening state vector of the first virtual student according to the updated state value in the target dimension by using state transition logic in a state machine corresponding to the first virtual student, and determining a behavior instruction corresponding to the first virtual student; And driving the first virtual student to make relevant behavior feedback according to the behavior instruction.
  5. 5. A multi-virtual human simulated lecture method according to claim 3, wherein the determining process of the target dimension and the first virtual student includes: Determining teaching behaviors corresponding to the multi-mode interaction data as target teaching behaviors; Determining a dimension with a corresponding relation with the target teaching behavior from the plurality of dimensions based on a pre-established corresponding relation between the teaching behavior and the influenced dimension, wherein the dimension is used as the target dimension; And determining a virtual student with a corresponding relation with the target teaching behavior from the plurality of virtual students based on the corresponding relation between the pre-established teaching behavior and the affected virtual student, and taking the virtual student as the first virtual student.
  6. 6. The multi-virtual human simulated lecture method according to claim 5, wherein when the target teaching behavior is a behavior of continuously gazing at a first area, the first virtual student includes a virtual student in a second area, the second area being an area other than the first area in the virtual classroom, the target dimension including an attention dimension; The updating the status value of the first virtual student in the target dimension according to the multi-modal interaction data and the personality parameters of the first virtual student to drive the first virtual student to make relevant behavior feedback based on the updated status value includes: According to the multi-mode interaction data and the personality parameters of the virtual students in the second area, reducing the attention value of the virtual students in the second area; and when the attention value reaches a preset attention threshold value, driving the virtual students in the second area to make preset target behavior feedback.
  7. 7. The multi-virtual human simulated lecture method according to claim 1, wherein said driving the virtual student to make relevant behavioral feedback based on the updated state value comprises: Acquiring individual parameters of the virtual students, wherein the individual parameters are parameters describing individual difference characteristics of the virtual students, and the individual difference characteristics at least comprise attribute characteristics for distinguishing classroom behavior responses of the students; And generating behavior instructions matched with the personality parameters according to the updated state values, and driving the virtual students to make relevant behavior feedback.
  8. 8. The multi-virtual human simulated lecture method according to any one of claims 2-7, wherein the parameters describing the individual difference characteristics include at least one of: Character characteristic parameters for representing the emotion response mode and social behavior tendency of students; Cognitive style parameters that characterize student information processing habits; and the behavior trend parameter is used for representing the interaction habit of students in class.
  9. 9. The multi-virtual simulated lecture method according to claim 3, further comprising: In the teaching process, when a preset emergency triggering condition is reached, driving a second virtual student to make an emergency behavior corresponding to the emergency triggering condition with a preset probability, and returning to the step of acquiring the multi-mode interaction data, wherein the second virtual student is used as a first virtual student when the first virtual student related to the multi-mode interaction data is determined.
  10. 10. The multi-virtual human simulated lecture method according to claim 9, wherein the second virtual student is a virtual student whose personality parameter satisfies a preset requirement.
  11. 11. The multi-virtual simulated lecture method according to claim 9, further comprising: Starting timing when the second virtual student makes the sudden behavior, and acquiring the response time of the target object for generating the multi-mode interaction data corresponding to the sudden behavior; According to the multi-mode interaction data and the behavior feedback obtained in the teaching process, generating teaching evaluation data comprises the following steps: and generating the teaching evaluation data according to the multi-mode interaction data, the behavior feedback and the reaction time obtained in the teaching process.
  12. 12. The multi-virtual human simulated lecture method according to claim 1, wherein the plurality of dimensions includes at least two of an attention dimension, a mood dimension, and a school emotion dimension.
  13. 13. The multi-virtual simulation teaching method according to claim 1, wherein the teaching evaluation data includes one or more of basic information of the target object teaching, classroom overall evaluation data, target event playback and analysis data, and comprehensive evaluation and improvement advice.
  14. 14. The multi-virtual human simulated lecture method according to claim 1, wherein the multi-modal interaction data includes at least two of three-dimensional skeleton data, facial expression data, line-of-sight drop data, voice recognition text, and acoustic feature data of the target object.
  15. 15. A computer program product comprising computer readable instructions which, when run on an electronic device, cause the electronic device to implement the multi-virtual human simulated lecture method of any one of claims 1 to 14.
  16. 16. An electronic device comprising at least one processor and a memory coupled to the processor, wherein: the memory is used for storing a computer program; The processor is configured to execute the computer program to enable the electronic device to implement the multi-virtual person simulation lecture method according to any one of claims 1 to 14.
  17. 17. A computer storage medium carrying one or more computer programs which, when executed by an electronic device, enable the electronic device to implement a multi-virtual human simulation teaching method as claimed in any one of claims 1 to 14.

Description

Multi-virtual-person simulation teaching method and related device Technical Field The application relates to the technical field of virtual teaching, in particular to a multi-virtual-person simulation teaching method and a related device. Background Teaching skills are the core competitiveness of the career of the teacher, and the cultivation quality of the teaching skills is directly related to the level of basic education. The traditional teaching skill training mode mainly comprises micro-grid teaching, teaching practice and virtual simulation system teaching. The micro-grid teaching is to conduct segment teaching on students played by students or instructors in a small range, but the reactions of the students played by the students or instructors are often preset and lack of spontaneity, individual differences of student groups in a real classroom are difficult to simulate, and therefore a training scene is disjointed from the real teaching situation, the teaching practice is to conduct teaching on the real students, but the students are possibly affected due to insufficient experience of a trainee and the like, the assessment of a teacher depends on personal experience, unified, objective and quantized data support is lacking, the teaching behaviors of the trainee are difficult to conduct fine analysis and guidance, and the virtual simulation system teaching is to conduct simulated teaching on a plurality of virtual students built in a virtual classroom scene, but the reactions of all virtual students in the virtual classroom are uniform, and complex ecology of the real classroom cannot be reflected, so that the trainee cannot effectively train the complex class, manage the rules and conduct differential teaching in the virtual classroom. In summary, there is an urgent need for a safe and controllable teaching and training solution that can highly simulate real classroom ecology and provide objective and quantitative feedback. Disclosure of Invention In view of the above problems, the present application provides a multi-virtual-person simulation teaching method and related apparatus, so as to achieve the objective and quantitative evaluation of the teaching process of the target object by simulating the high-simulation classroom ecology. The specific scheme is as follows: the first aspect of the application provides a multi-virtual-person simulation teaching method, which comprises the following steps: acquiring multi-mode interaction data of a target object in a virtual classroom, wherein initial class-listening state vectors of a plurality of virtual students in the virtual classroom are determined by preset class portraits, and the initial class-listening state vectors comprise state values of a plurality of dimensions; determining a target dimension influenced by the multi-modal interaction data, wherein the target dimension is part or all of the multiple dimensions; Updating the state value of the virtual student under the target dimension according to the multi-modal interaction data to drive the virtual student to make relevant behavior feedback based on the updated state value, and returning to the step of acquiring the multi-modal interaction data until the teaching is finished; And generating teaching evaluation data according to the multi-mode interaction data and the behavior feedback obtained in the teaching process. In a possible implementation, the updating the status value of the virtual student in the target dimension according to the multi-modal interaction data includes: Acquiring individual parameters of the virtual students, wherein the individual parameters are parameters describing individual difference characteristics of the virtual students, and the individual difference characteristics at least comprise attribute characteristics for distinguishing classroom behavior responses of the students; And updating the state value of the virtual student in the target dimension according to the multi-mode interaction data and the personality parameters of the virtual student. In one possible implementation, the method further includes: Determining a first virtual student affected by the multi-modal interaction data, the first virtual student being part or all of the plurality of virtual students; the updating the status value of the virtual student in the target dimension according to the multi-modal interaction data and the personality parameters of the virtual student to drive the virtual student to make relevant behavior feedback based on the updated status value comprises: And updating the state value of the first virtual student under the target dimension according to the multi-mode interaction data and the personality parameters of the first virtual student so as to drive the first virtual student to make relevant behavior feedback based on the updated state value. In one possible implementation, the updating the status value of the first virtual student in the target dimension a