Robot system and method for teaching and training

Abstract

The invention discloses a robot system and method for teaching and training, and relates to the technical field of artificial intelligence education, and the system comprises a data initialization information collection module which is used for enabling a student to contact a robot for the first time, inputting a learning ability test result, hobbies and interests, and basic information of knowledge mastering degree self-evaluation by means of a multi-mode interaction interface of the robot, the basic information of the students is collected through the initialized information acquisition module, and the learning ability of the students is scientifically evaluated by means of the personalized learning planning module and the comprehensive characteristic index construction formula of the students, so that personalized learning paths are customized for the students, the pertinence and efficiency of learning are improved, and the learning ability of the students is improved. Meanwhile, the project type learning guiding module evaluates according to the project difficulty adaptation degree, it is ensured that the project difficulty is matched with the ability of the student, and the learning effect is further improved.

Description

Technical Field

[0001] The present invention relates to the technical field of artificial intelligence education, and specifically to a robot system and method for teaching and training. Background Art

[0002] With the continuous development and innovation of educational technology, modern teaching methods are gradually shifting towards the direction of intelligence, personalization, and interaction. Against this background, robot systems for teaching and training have emerged as a new force in the education field. These robot systems, by leveraging advanced artificial intelligence technologies, can provide students with more efficient, accurate, and personalized learning support. They collect students' learning data, analyze their learning habits and abilities, and then formulate learning plans that meet the students' needs, thus greatly improving the teaching effect and learning experience. It is precisely under such a technical background that the present invention proposes a new type of robot system for teaching and training, aiming to further promote the innovation and development of educational technology.

[0003] However, despite the certain achievements made by existing robot systems for teaching and training, there are still many deficiencies. Traditional robot systems often overly focus on a single teaching form, ignoring the individual differences among students and the diversity of learning needs. They usually adopt standardized learning resources and activities, lacking the planning and guidance for students' personalized learning paths. In addition, during the interaction process with students, traditional systems often lack the real-time recognition and dynamic adjustment of students' emotional states, resulting in unsatisfactory teaching effects. These problems not only limit the learning potential and interest of students but also affect the further popularization and application of educational technology.

[0004] Therefore, developing a robot system and method for teaching and training is expected to solve the problems existing in traditional robot systems for teaching and training and promote the development of educational technology to a higher level. Summary of the Invention

[0005] The purpose of the present invention is to make up for the deficiencies of the existing technology and provide a robot system and method for teaching and training. By integrating information collection, personalized learning planning, project-based learning guidance, multi-modal interaction, and real-time feedback adjustment, through a multi-modal interaction interface, the system can comprehensively understand the learning abilities, interests, and knowledge mastery levels of students, and then formulate personalized learning paths and project plans. During the execution of the project, the robot provides full support, real-time recognizes students' emotions, and dynamically adjusts the interaction strategy. At the same time, it collects various data during the learning process, conducts in-depth analysis, and continuously optimizes the learning path and plan, which will greatly improve the intelligence and personalization levels of teaching and training and promote the improvement of learning effects.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: On the one hand, a robot system for teaching and training, the system includes: Initial information collection module: When students first come into contact with the robot, with the help of the multi-modal interaction interface equipped with the robot, they enter basic information such as learning ability test results, hobbies, and self-evaluation of knowledge mastery, and encrypt and transmit the collected information to the personalized path planning module; Personalized learning planning module: Receives the information collected by the initial information collection module, cleans and standardizes the data, comprehensively considers the performance of students in various aspects to evaluate their learning ability, selects learning directions in combination with interests, and matches learning resources and activities from the resource library to formulate a personalized learning path for students; Project-based learning guidance module: According to the teaching objectives and personalized learning path, formulates a project plan for students. During the implementation of the project, the robot provides knowledge and technical support, answers questions, records the performance of students in all aspects, generates a visual report, and optimizes the project plan according to the actual situation; Multi-modal interaction execution module: With the help of microphone arrays, cameras, depth sensors, touch sensors, and pressure sensor devices, receives information such as students' voices, gestures, expressions, and touches, executes learning resource retrieval and function control, identifies the emotional state of students by constructing an emotion recognition model, and dynamically adjusts the interaction strategy according to the emotional state of students; Real-time feedback adjustment module: Collects the usage of learning resources, project progress, test scores, interaction data, and emotional changes during the learning process of students, deeply analyzes these data using the learning dynamics comprehensive evaluation formula, and adjusts the personalized learning path, optimizes the project-based learning plan, and improves the multi-modal interaction strategy according to the analysis results.

[0007] Furthermore, in the personalized learning planning module, a formula for constructing the student comprehensive characteristic index is used to evaluate the learning ability of students. The calculation formula is: , where is the learning ability evaluation value, which is determined by the speed at which students learn new knowledge and whether they can flexibly apply the learned knowledge to new scenarios. is the degree of knowledge mastery, which is determined based on the test scores of students, the accuracy and completion efficiency of homework. is the interest preference, which is determined by the frequency and degree of participation of students in various interest activities. , , are weight coefficients, which are optimized through a supervised learning model according to historical student data, and are dynamically adjusted according to different teaching scenarios and student groups of different ages and different learning bases.

[0008] Furthermore, in the personalized learning planning module, a personalized learning path is formulated for students through the personalized learning path matching degree formula, and the formula is: , where is the cosine similarity function, and the calculation formula is: , is the comprehensive feature index vector of the student, is the th feature vector of the learning resource or activity, is the dot product of vectors, and are the norms of the vectors respectively, is the importance weight of the th learning resource or activity, which is set through data analysis according to the importance and scarcity factors of the learning resource or activity in the teaching objective. The higher the value, the more suitable the learning path is for the student. Based on this, the robot recommends the most suitable learning resources and activities for the student.

[0009] Furthermore, the specific steps for formulating a project plan in the project-based learning guidance module according to the teaching objective and the personalized learning path: Analyze the teaching objective, clarify the knowledge and skill requirements and the expected results, and at the same time sort out the student's learning progress, weak points and interest preferences, conceive a suitable project theme and framework, determine the stage tasks and expected results, adapt resources and teaching methods according to the project requirements, arrange the schedule, plan the time, evaluate the matching degree between the project difficulty and the student's ability in project-based learning through the matching degree algorithm, and adjust the project according to the evaluation results.

[0010] Furthermore, in the project-based learning guidance module, the matching degree between the project difficulty and the student's ability in project-based learning is evaluated through the matching degree algorithm, and the formula is: , where is the project difficulty value, which is determined by the complexity of the knowledge and skills required by the project, the amount of tasks to be completed for the project, and the expected time to be spent. is the comprehensive feature index of the student , will approach 1, indicating that the project is relatively simple for the student; approaches 0, indicating that the project is too difficult for the student.

[0011] Furthermore, in the multi-modal interaction execution module, the student's emotion is recognized by constructing an emotion recognition model. Let be the finally calculated emotion index, and the calculation formula is: , where is the voice intonation feature value, F is the facial expression feature value, T is the touch force feature value, , , is a weight coefficient, and , the optimal value is determined through multi-modal data training and cross-validation. If E > Epositive, it is determined that the student is in a positive emotional state; if E < Enegative, it is determined that the student is in a negative emotional state; if Enegative ≤ E ≤ Epositive, it is determined that the student is in a neutral emotional state. Epositive and Enegative are preset thresholds determined based on historical data.

[0012] Furthermore, the pitch contour feature value in the multi-modal interaction execution module is calculated as follows: from the pitch change rate , the volume change rate and the speech rate change rate combined. The calculation formula is: , where , , are weight coefficients, and ; the calculation of the facial expression feature value F: obtained by weighted summation of various basic expression features . The formula is: , where is the weight coefficient of each expression feature, ; the calculation of the touch force feature value T: the touch force feature value T is related to the touch force t and the touch duration l. The calculation formula is: , where and are weight coefficients.

[0013] Furthermore, in the multi-modal interaction execution module, the interaction strategy is dynamically adjusted according to the student's emotional state. When the student is in a positive emotion: increase the difficulty and challenge of the learning content; when the student shows negative emotion: communicate in a more gentle and encouraging language and provide more support and help; when the student is in a neutral emotional state: maintain the existing teaching rhythm and method, continuously observe and guide appropriately, and provide extended learning resources.

[0014] Furthermore, in the multi-modal interaction execution module, the learning dynamic comprehensive evaluation formula is used to deeply analyze relevant data. The formula is: , where Adjust is the comprehensive adjustment coefficient, Us represents the learning resource usage situation, Pr is the project progress situation, Sc is the test score, In represents the interaction data, and Em represents the student's emotional change. , , , , is a weight coefficient, which is dynamically adjusted according to different teaching scenarios, learning stages, and individual differences of students.

[0015] On the other hand, a robot method for teaching and training, the specific steps of which are as follows: S100, Information collection and processing: When a student first comes into contact with the teaching and training robot, enter the basic information of the learning ability test results, hobbies, and self-evaluation of knowledge mastery through the multimodal interaction interface; S200, Personalized learning path planning: Receive the relevant data entered by the student, clean and standardize the data, screen the learning direction in combination with the student's interests, select learning resources and activities from the resource library, and formulate a personalized learning path exclusive to the student; S300, Project-based learning guidance: Develop a project plan according to the teaching objectives and personalized learning path. During the project execution, provide the student with knowledge and technical support, answer questions, record the student's performance to form a project learning record, and optimize the project plan according to the actual situation; S400, Multimodal interaction and teaching strategy adjustment: Use multiple sensors to receive the information of the student's voice, gestures, expressions, and touches, perform the operation of retrieving learning resources, identify the student's emotions with the constructed emotion recognition model, and adjust the interaction strategy according to the emotions; S500, Real-time feedback and adjustment: Collect the usage of learning resources, project progress, test scores, interaction data, and emotional changes during the student's learning process, analyze the student's learning progress and status using the learning dynamics comprehensive evaluation formula, and adjust the personalized learning path, optimize the project-based learning plan, and improve the multimodal interaction strategy according to the analysis results.

[0016] Compared with the prior art, the robot system and method for teaching and training have the following beneficial effects: First, the present invention collects the basic information of students through the initialization information collection module, and with the help of the personalized learning planning module, uses the formula for constructing the student comprehensive characteristic index to scientifically evaluate the learning ability of students, so as to customize a personalized learning path for them. This not only improves the pertinence and efficiency of learning, but also greatly stimulates the learning interest and enthusiasm of students. At the same time, the project-based learning guidance module evaluates according to the project difficulty adaptability to ensure that the project difficulty matches the student's ability, further improving the learning effect. In addition, the multimodal interaction execution module can identify the student's emotions in real time and dynamically adjust the interaction strategy, making the learning process smoother and more enjoyable.

[0017] II. The present invention introduces project-based learning and a real-time feedback adjustment mechanism. Project-based learning not only helps students apply the knowledge they have learned to practical situations, but also cultivates their teamwork, problem-solving, and innovation abilities. The real-time feedback adjustment mechanism can timely reflect the learning progress and status of students. By deeply analyzing this data using the learning dynamics comprehensive evaluation formula, the personalized learning path can be adjusted, the project-based learning plan can be optimized, and the multimodal interaction strategy can be improved. This dynamically adjusted teaching method makes teaching more flexible and efficient, and can better adapt to the learning needs and ability levels of different students, not only improving the teaching quality, but also providing students with a richer, more diverse, and personalized learning experience.

[0018] Other advantages, objectives, and features of the present invention will be described to some extent in the subsequent specification, and to some extent, will be obvious to those skilled in the art based on the study of the following text, or can be taught from the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0020] Figure 1 It is a framework diagram of a robot system for teaching and training; Figure 2 It is a flowchart of a robot method for teaching and training. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0021] To further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention objective, the following will, in combination with the drawings and preferred embodiments, detail the specific embodiments, structures, features, and their effects of the present invention as follows.

[0022] Embodiment 1:

[0023] Teaching of primary school science courses.

[0024] Initial information collection: At the beginning of the new semester, third-grade students in a primary school start learning science courses using a robot system for teaching training. In the science classroom, the students sit around the robot and, through the robot's multi-modal interaction interface, successively enter the results of their learning speed of scientific knowledge and knowledge application ability in the previous learning ability test, fill in their situations regarding scientific-related interests and hobbies, such as liking to observe animals and plants and doing simple experiments, and at the same time conduct self-evaluation on their current mastery of scientific knowledge. This information is encrypted and transmitted to the personalized learning planning module.

[0025] Personalized learning planning: After receiving the students' information, the personalized learning planning module uses the formula for constructing the comprehensive student characteristic index to evaluate the students' learning ability. The calculation formula is: , where is the learning ability evaluation value, which is determined by the speed at which students learn new knowledge and the indicator of whether they can flexibly apply the learned knowledge to new scenarios. is the degree of knowledge mastery, which is determined based on the students' test scores, the accuracy rate and completion efficiency of their homework. is the interest preference, which is determined by the frequency and degree of participation of students in various interest activities. , , are the weight coefficients, which are optimized through a supervised learning model based on historical student data and are dynamically adjusted according to different teaching scenarios and student groups of different ages and different learning bases. Through analysis, it is found that some students learn new knowledge relatively fast but are slightly weak in knowledge application, and some students have a good degree of knowledge mastery but have an average speed of learning new knowledge. Combining the students' interest preferences, suitable learning resources and activities are matched from the science curriculum resource library. For example, for students who like experiments, videos of simple circuit experiment operations and learning resources for the introduction of experimental equipment are matched, and for students who like to observe, animal and plant observation manuals and natural documentaries are recommended. According to the matching results, a personalized learning path is formulated for the students through the personalized learning path matching degree formula. The formula is: Historical student data is optimized and dynamically adjusted according to different teaching scenarios and student groups of different ages and different learning bases. Through analysis, it is found that some students learn new knowledge relatively fast but are slightly weak in knowledge application, and some students have a good degree of knowledge mastery but have an average speed of learning new knowledge. Combining the students' interest preferences, suitable learning resources and activities are matched from the science curriculum resource library. For example, for students who like experiments, videos of simple circuit experiment operations and learning resources for the introduction of experimental equipment are matched, and for students who like to observe, animal and plant observation manuals and natural documentaries are recommended. According to the matching results, a personalized learning path is formulated for the students through the personalized learning path matching degree formula. The formula is: , where is the cosine similarity function, and the calculation formula is: , is the comprehensive characteristic index vector of the student, is the th characteristic vector of the learning resource or activity, is the dot product of vectors, and are the norms of the vectors respectively, is the importance weight of the th learning resource or activity, which is set through data analysis according to the importance and scarcity factors of the learning resource or activity in the teaching objective. The higher the value, the more suitable the learning path is for the student. Based on this, the robot recommends the most suitable learning resources and activities for the student.

[0026] Project-based learning guidance: In the teaching unit of "Circuit Exploration", the teaching objective is to enable students to understand the composition and connection methods of simple circuits and be able to actually operate and build circuits. The project-based learning guidance module formulates a project plan based on the teaching objective and the personalized learning path. After analyzing the teaching objective, it is clear that students need to master circuit element knowledge and learn the skills of connecting circuits. The expected outcome is that students can independently build a simple circuit and make the light bulb emit light. Sort out the students' learning progress, weak points and interest preferences, conceive the project theme of "making a simple flashlight", determine the stage tasks. The first stage is to learn circuit element knowledge, the second stage is to design the flashlight circuit, and the third stage is to actually make it. According to the project requirements, adapt the circuit element models, experimental equipment resources and group cooperation teaching methods, arrange the schedule, plan the time for each stage, and use the project-based learning difficulty adaptation degree evaluation formula, the formula is: , where is the project difficulty value, which is determined by the complexity of the knowledge and skills required by the project, the amount of tasks to be completed for the project, and the expected time spent. is the comprehensive characteristic index of the student , will approach 1, indicating that the project is relatively simple for the student; approaches 0, indicating that the project is too difficult for the student. If it is found that the project is too difficult, appropriately reduce the task requirements, such as simplifying the circuit design link. During the project implementation process, the robot provides students with circuit knowledge explanations, technical support, and answers questions encountered by students when building circuits, records students' performance in all aspects, such as participation and operation accuracy, generates a visual report, and optimizes the project plan according to the actual situation.

[0027] Multimodal Interaction Execution: During the project implementation, the multimodal interaction execution module receives students' voice, gesture, expression, and touch information with the help of microphone arrays, cameras, depth sensors, touch sensors, and pressure sensor devices. For example, when a student asks the robot a question about the connection of the positive and negative poles of a battery through voice, the robot accurately identifies and answers. The camera captures the student's expression when observing the circuit connection. If it is found that the student frowns or has a confused look in their eyes, the emotion index is calculated through the constructed emotion recognition model. Let be the finally calculated emotion index, and the calculation formula is: , where is the voice intonation feature value, and the voice intonation feature value is obtained by synthesizing the pitch change rate , the volume change rate , and the speech rate change rate . The calculation formula is: , , , are the weight coefficients, and , F is the facial expression feature value, and the facial expression feature value F is obtained by weighted summation of various basic expression features . The formula is: , is the weight coefficient of each expression feature, , T is the touch force feature value, and the touch force feature value T is related to the touch force t and the touch duration l. The calculation formula is: , where and are the weight coefficients, , , are the weight coefficients, and . After calculation, E < Enegative. At this time, it is judged that the student is in a negative emotional state. The robot uses a more gentle and encouraging language for communication, such as "Don't worry. Let's take a look together. You've done a great job", and provides more support to simplify the learning task, such as first asking the student to build a simpler series circuit. At the same time, the student's touch operation content is saved for subsequent analysis.

[0028] Real-time Feedback and Adjustment: The real-time feedback and adjustment module collects information on the use of learning resources during the student's learning process, such as the duration and frequency of watching experiment videos, the progress of the project, such as whether the stage tasks are completed on time, the test scores, such as the scores in the circuit knowledge quizzes, the interaction data, such as the number of questions asked and the duration of voice interaction, and the data on emotional changes. It deeply analyzes the relevant data using the learning dynamic comprehensive evaluation formula. The formula is: , where Adjust is the comprehensive adjustment coefficient, Us represents the learning resource usage, Pr is the project progress, Sc is the test score, In represents the interaction data, and Em represents the emotional changes of students. 、 、 、 、 are the weight coefficients, which are dynamically adjusted according to different teaching scenarios, learning stages, and individual differences of students. Through analysis, it is found that some students have difficulties in the actual operation of circuit connection. According to the analysis results, the personalized learning path is adjusted, and more circuit connection practice resources are added; the project-based learning plan is optimized, and the circuit building practice time is extended; the multi-modal interaction strategy is improved, and more detailed voice guidance and animation demonstrations are provided for the difficult points of operation.

[0029] In summary, in the teaching scenario of primary school science courses, the teaching and training robot system has played a significant role. The initialization information collection module collects various information of students, providing a basis for subsequent planning. The personalized learning planning module accurately matches learning resources to meet the needs of students. The project-based learning guidance module designs projects around teaching objectives to help students master knowledge and skills. The multi-modal interaction execution module senses the emotions of students in real time, dynamically adjusts the interaction strategy. The real-time feedback adjustment module optimizes the learning path and plan based on multi-dimensional data. Through the collaborative work of each module, the system effectively improves the teaching effect of primary school science courses, stimulates students' learning interest, and enhances students' practical and thinking abilities.

[0030] Example Two:

[0031] Vocational skill training (automobile repair training).

[0032] Initialization information collection: In a vocational skill training institution for automobile repair, a new batch of students enroll. The students use the teaching and training robot system in the training classroom and enter the learning ability results related to automobile repair in their previous learning ability tests through the multi-modal interaction interface, such as the learning speed of automobile structure knowledge and the ability to apply the learned repair theory to actual fault judgment. They fill in their hobbies in the field of automobile repair, such as liking to study automobile engines and automobile electrical systems, and self-evaluate their mastery of current automobile repair knowledge. These information are encrypted and transmitted to the personalized learning planning module.

[0033] Personalized learning planning: After receiving the students' information, the personalized learning planning module uses the formula for constructing the comprehensive student characteristic index to evaluate the students' learning ability. The calculation formula is: , after analysis, it is found that some students absorb new knowledge quickly but lack practical operation experience, while some students have a good grasp of theoretical knowledge but lack enthusiasm for learning new knowledge. Combining the students' interest preferences, learning resources and activities are matched from the automotive repair training resource library. For example, for students who like to study automotive engines, video tutorials on engine disassembly and assembly and learning resources on case analysis of engine fault diagnosis are matched; for students who like automotive electrical systems, interpretations of automotive electrical circuit diagrams and user manuals for electrical fault detection equipment are recommended. According to the matching results, personalized learning paths are developed for different student groups, and the formula is: , The higher the value, the more suitable the learning path is for the student. Based on this, the robot recommends the most suitable learning resources and activities for the student.

[0034] Project-based learning guidance: In the training course of "Automotive Engine Fault Diagnosis and Repair", the teaching objective is to enable students to master the diagnostic methods and repair skills of common engine faults and be able to independently complete engine fault troubleshooting and repair work. The project-based learning guidance module formulates a project plan based on the teaching objective and the personalized learning path, analyzes the teaching objective, and clarifies that students need to master the knowledge and skills of the structure, working principle, and fault diagnosis process of each engine component. The expected outcome is that students can accurately diagnose engine faults and perform effective repairs. The learning progress, weak points, and interest preferences of students are sorted out, and the project theme of "solving the problem of difficult engine starting" is conceived. The stage tasks are determined. In the first stage, students learn the knowledge of the engine starting system; in the second stage, they conduct practical fault troubleshooting; in the third stage, they repair the faults and verify the effects. According to the project requirements, physical engine models, fault diagnosis equipment resources, and on-site practical teaching methods are adapted, and a schedule is arranged to plan the time for each stage. The adaptation degree between the comprehensive characteristic index SCI of the student and the project difficulty value PD is calculated using the project-based learning difficulty adaptation degree evaluation formula, and the formula is: , if it is found that the project difficulty is too high for some students with weak foundations, the project difficulty is adjusted. For example, start with the troubleshooting of simple electrical circuit faults in the engine starting system. During the project implementation, the robot provides students with explanations of engine knowledge and technical support for fault diagnosis, answers the questions encountered by students during fault troubleshooting, records the students' performance in all aspects, such as the fault troubleshooting ideas and the standard degree of repair operations, generates a visual report, and optimizes the project plan according to the actual situation.

[0035] Multimodal Interaction Execution: During the project implementation, the multimodal interaction execution module uses devices such as microphone arrays, cameras, depth sensors, touch sensors, and pressure sensors to receive the voice, gestures, expressions, and touch information of the trainees. For example, the trainee asks the robot about the working principle of a certain component of the engine through voice, and the robot accurately identifies and answers. The camera captures the expressions of the trainee when observing the engine fault phenomenon. If it is found that the trainee has a concentrated expression and nods, the emotion index is calculated through the constructed emotion recognition model. The calculation formula is: , after calculation, when E > Epositive, it is judged that the trainee is in a positive emotional state. The robot appropriately increases the difficulty and challenge of the learning content, such as presenting more complex engine fault scenarios for the trainee to analyze and diagnose. At the same time, the trainee's touch operation content is saved, such as the touch record when operating the fault diagnosis device.

[0036] Real-time Feedback and Adjustment: The real-time feedback and adjustment module collects the usage of learning resources during the trainee's learning process, such as the duration of watching maintenance videos and the frequency of consulting materials, the progress of the project, such as whether the fault troubleshooting task is completed on time, the test scores, such as the scores in the theoretical test and practical assessment of engine fault diagnosis, the interaction data, such as the number of times of communicating with the robot and the content of the questions, and the data of emotional changes. The relevant data is deeply analyzed using the learning dynamic comprehensive evaluation formula. The formula is: , through analysis, it is found that some trainees have problems with non-standard operations in the practical operation link of engine fault diagnosis. According to the analysis results, the personalized learning path is adjusted to increase more practical training resources; the project-based learning plan is optimized to increase the time and frequency of practical exercises; the multimodal interaction strategy is improved to provide timely corrective prompts and demonstration videos for non-standard operation actions.

[0037] In summary, in the scenario of automotive repair vocational skills training, this robot system has prominent advantages. The initialization information collection enables the system to deeply understand the trainees' basis and interests. The personalized learning plan customizes the learning path for the trainees, improving the learning efficiency. The project-based learning guidance module conducts teaching around actual repair projects, enhancing the trainees' ability to solve practical problems. The multimodal interaction execution module adjusts the teaching according to the trainees' emotions, enhancing the learning experience. The real-time feedback and adjustment module optimizes the teaching strategy based on learning data. Each module cooperates closely to help trainees efficiently master automotive repair knowledge and skills, shorten the training cycle, improve the training quality, and deliver professional talents to the automotive repair industry.

[0038] Example 3:

[0039] Implemented in junior high school mathematics curriculum teaching.

[0040] In the mathematics teaching of a junior high school, the teaching training robot system conducts teaching activities according to the teaching training robot method.

[0041] Information collection and processing (S100): At the beginning of the new semester, students in multiple classes of the first grade of junior high school start to use this robot system to assist in mathematics learning. In the mathematics classroom, through the multi-modal interaction interface equipped on the robot, students input their mathematics learning ability test scores from the previous semester, their interest tendencies in different sections of algebra and geometry during the mathematics learning process, and their subjective evaluations of their own mathematics knowledge mastery. These pieces of information are quickly collected.

[0042] Personalized learning path planning (S200): After the system receives the data input by students, it first cleans and standardizes the data. For example, it removes abnormal test score data and unifies the criteria for self-evaluation of knowledge mastery. Then, it selects learning directions based on students' interests. For students who like algebra, it selects extended exercise questions related to algebraic equations and functions, as well as interesting mathematics story learning resources from the resource library; for students who like geometry, it matches content on exploring the properties of geometric figures and analyzing the ideas of geometric proofs. According to the selected resources, it formulates personalized learning paths for different student groups. For example, for students with a good foundation and an interest in algebra, it plans for them to first deeply study the comprehensive application of functions and then participate in a mathematics modeling project; for students with a weak foundation in geometry but an interest, it arranges for them to start from the recognition of basic geometric figures and gradually increase the difficulty.

[0043] Project-based learning guidance (S300): When teaching the unit of "Pythagorean theorem", the teaching goal is to enable students to understand the concept and proof method of the Pythagorean theorem and be able to use it to solve practical problems. The system formulates a project plan based on the teaching goal and the personalized learning path. The project theme is set as "Measuring the side length relationship of right-angled triangle objects on campus". The stage tasks are determined as follows: In the first stage, students independently study the theoretical knowledge of the Pythagorean theorem; in the second stage, they work in groups to use measuring tools to measure the side lengths of right-angled triangle objects on campus such as stair steps and flagpole bases; in the third stage, they analyze the measurement data, verify the Pythagorean theorem, and write a report. During the project implementation process, the robot explains the proof ideas of the Pythagorean theorem to students and answers the questions they encounter in measurement and data processing, such as the processing method of measurement errors. At the same time, it records the performance of each student in terms of their participation in group discussions, the accuracy of measurement operations, and the rationality of data analysis, forming a project learning record.

[0044] Multimodal Interaction and Teaching Strategy Adjustment (S400): During the project, the robot uses various devices such as microphone arrays, cameras, and touch sensors to receive multimodal information from students. For example, when a student asks about the application of the Pythagorean theorem in an irregular right triangle through voice, the robot quickly retrieves relevant learning resources for answering. When the camera captures the student's frowning and shaking head expressions during group discussions, the emotion index is calculated through an emotion recognition model. Assuming that the emotion index calculated comprehensively from voice intonation feature values, facial expression feature values, and touch force feature values shows that the student is in a negative emotion, the robot adjusts the interaction strategy, guides the student with encouraging words such as "Don't worry, this problem is a bit difficult, let's take it step by step", and provides more detailed problem-solving ideas and examples. At the same time, the student's touch operation content is saved, such as the touch record when operating the measurement tool, for subsequent analysis of the student's familiarity with the tool.

[0045] Real-time Feedback and Adjustment (S500): The system collects various data of students during the learning process, including the usage frequency and duration of learning resources, such as the number of times and duration of watching the video of the proof of the Pythagorean theorem; the progress of the project, such as whether each stage task is completed on time; test scores, such as the scores of unit quizzes; interaction data, such as the content and number of questions; and emotion changes. Analyzed using a comprehensive learning dynamics assessment formula, if it is found that some students make more mistakes in the practical application questions of the Pythagorean theorem, according to the analysis results, the personalized learning path is adjusted, and more practice resources for practical application cases are added; the project-based learning plan is optimized, and more time is arranged for practical measurement and case analysis; the multimodal interaction strategy is improved, and for the knowledge points where students are prone to make mistakes, explanatory videos and practice questions are actively pushed to improve the teaching effect.

[0046] In summary, in the teaching of the "Pythagorean theorem" in junior high school mathematics courses, the teaching and training method of this robot has achieved remarkable results. The information collection and processing link comprehensively collects students' data, providing strong support for subsequent teaching planning. The personalized learning path planning accurately matches learning resources based on students' interests and foundations, meeting the needs of different students. The project-based learning guidance drives students to master knowledge with actual projects, cultivating practical and cooperative abilities. The multimodal interaction promptly senses students' emotions and flexibly adjusts teaching strategies. The real-time feedback and adjustment optimize teaching based on multi-dimensional data, solving students' learning difficulties. Through the close cooperation of each link, it effectively improves students' understanding and application abilities of the Pythagorean theorem, and enhances students' enthusiasm and initiative in mathematics learning.

[0047] The above are only the preferred embodiments of the present invention and do not impose any formal restrictions on the present invention. Although the present invention has been disclosed above with the preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some changes or modifications to equivalent embodiments with equivalent changes within the scope of the technical solution of the present invention. However, as long as it does not depart from the content of the technical solution of the present invention, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. A robot system for teaching and training, characterized in that, The system includes: Initial information collection module: When students first come into contact with the robot, with the help of the multi-modal interaction interface equipped on the robot, they input basic information such as the test results of learning ability, hobbies, and self-evaluation of knowledge mastery, and encrypt and transmit the collected information to the personalized path planning module; Personalized learning planning module: Receives the information collected by the initial information collection module, cleans and standardizes the data, comprehensively evaluates the learning ability of students considering various aspects of their performance, selects learning directions in combination with interests, and matches learning resources and activities from the resource library to formulate a personalized learning path for students; Project-based learning guidance module: According to the teaching objectives and personalized learning path, formulates a project plan for students. During the implementation of the project, the robot provides knowledge and technical support, answers questions, records the performance of students comprehensively, generates a visual report, and optimizes the project plan according to the actual situation; Multi-modal interaction execution module: With the help of devices such as microphone arrays, cameras, depth sensors, touch sensors, and pressure sensors, receives information of students' voices, gestures, expressions, and touches, executes the retrieval of learning resources and function control, identifies the emotional state of students by building an emotion recognition model, and dynamically adjusts the interaction strategy according to the emotional state of students; Real-time feedback and adjustment module: Collects the usage of learning resources, project progress, test scores, interaction data, and emotional changes during the learning process of students, deeply analyzes these data using the learning dynamics comprehensive evaluation formula, and adjusts the personalized learning path, optimizes the project-based learning plan, and improves the multi-modal interaction strategy according to the analysis results.

2. The robot system for teaching and training according to claim 1, characterized in that, In the personalized learning planning module, a formula for constructing a comprehensive student characteristic index is used to evaluate the learning ability of students. The calculation formula is as follows: , where is the learning ability evaluation value, which is determined by the speed at which students learn new knowledge and the indicator of whether they can flexibly apply the learned knowledge to new scenarios. is the degree of knowledge mastery, which is determined based on the test scores of students, the accuracy rate and completion efficiency of homework. is the interest preference, which is determined by the frequency and degree of participation of students in various interest activities. , , are weight coefficients, which are optimized according to historical student data through a supervised learning model and are dynamically adjusted according to different teaching scenarios and student groups of different ages and different learning bases.

3. The robot system for teaching and training according to claim 2, wherein In the personalized learning planning module, a personalized learning path is formulated for students through the personalized learning path matching degree formula, and the formula is: , where is the cosine similarity function, and the calculation formula is: , is the comprehensive feature index vector of the student, is the th feature vector of the learning resource or activity, is the dot product of vectors, and are the norms of the vectors respectively, is the th importance weight of the learning resource or activity, which is set through data analysis according to the importance and scarcity factors of the learning resource or activity in the teaching objective, The higher the value, the more suitable the learning path is for the student, and the robot recommends the most suitable learning resources and activities for the student accordingly.

4. A robot system for teaching and training according to claim 1, characterized in that, The specific steps for formulating a project plan in the project-based learning guidance module according to the teaching objectives and personalized learning path: Analyze the teaching objectives, clarify the knowledge and skill requirements and expected outcomes, while sorting out the learning progress, weak points, and interest preferences of students, conceive a suitable project theme and framework, determine the stage tasks and expected outcomes, adapt resources and teaching methods according to project requirements, arrange a schedule, plan the time, evaluate the suitability of the project difficulty and students' abilities in project-based learning through a suitability algorithm, and adjust the project according to the evaluation results.

5. The robot system for teaching and training according to claim 4, wherein, In the project-based learning guidance module, the adaptation degree between the project difficulty and the student ability in project-based learning is evaluated through an adaptation degree algorithm, and the formula is: , where is the project difficulty value, which is determined by the complexity of the knowledge and skills required for the project, the amount of tasks to be completed for the project, and the estimated time to be spent. is the student comprehensive characteristic index . will approach 1, indicating that the project is relatively simple for the student; approaches 0, indicating that the project is too difficult for the student.

6. The robot system for teaching and training according to claim 1, characterized in that, In the multi-modal interaction execution module, a sentiment recognition model is constructed to recognize the emotions of students. Let be the finally calculated sentiment index, and the calculation formula is: , where is the voice intonation feature value, F is the facial expression feature value, T is the touch force feature value, , , are the weight coefficients, and . The optimal values are determined through multi-modal data training and cross-validation. If E > Epositive, it is judged that the student is in a positive emotional state; if E < Enegative, it is judged that the student is in a negative emotional state; if Enegative ≤ E ≤ Epositive, it is judged that the student is in a neutral emotional state. Epositive and Enegative are preset thresholds determined according to historical data.

7. A robot system for teaching and training according to claim 5, characterized in that, Calculation of the voice intonation feature value in the multi-modal interaction execution module : It is comprehensively obtained from the pitch change rate , the volume change rate and the speech rate change rate . The calculation formula is: , where , , are weight coefficients, and ; Calculation of the facial expression feature value F: It is obtained by weighted summation of various basic expression features . The formula is: , where is the weight coefficient of each expression feature, ; Calculation of the touch force eigenvalue T: The touch force eigenvalue T is related to the touch force t and the touch duration l, and the calculation formula is: , where and are weighting coefficients.

8. The robot system for teaching and training according to claim 1, wherein In the multi-modal interaction execution module, the interaction strategy is dynamically adjusted according to the emotional state of students. When students are in a positive mood: increase the difficulty and challenge of learning content; when students show negative emotions: communicate in a more gentle and encouraging language and provide more support and help; when students are in a neutral emotional state: maintain the existing teaching rhythm and method, continuously observe and guide appropriately, and provide extended learning resources.

9. A robot system for teaching and training according to claim 1, characterized in that, In the multi-modal interaction execution module, a learning dynamic comprehensive evaluation formula is used to deeply analyze relevant data. The formula is: , where Adjust represents the comprehensive adjustment coefficient, Us represents the learning resource usage, Pr represents the project progress, Sc represents the test scores, In represents the interaction data, and Em represents the emotional changes of students. , , , , are weight coefficients, which are dynamically adjusted according to different teaching scenarios, learning stages, and individual differences of students.

10. A robot method for teaching and training, characterized in that, This method is applicable to a teaching and training robot system described in any one of claims 1-9. The specific steps of this method are as follows: S100, Information collection and processing: When students first come into contact with the teaching and training robot, they input basic information such as the test results of learning ability, hobbies, and self-evaluation of knowledge mastery through the multi-modal interaction interface; S200, Personalized Learning Path Planning: Receive relevant data input by students, clean and standardize the data, screen learning directions in combination with students' interests, select learning resources and activities from the resource library, and formulate a personalized learning path exclusive to students; S300, Project-based Learning Guidance: Formulate a project plan according to teaching objectives and the personalized learning path. During project execution, provide students with knowledge and technical support, answer questions, record students' performance to form project learning records, and optimize the project plan according to the actual situation; S400, Multimodal Interaction and Teaching Strategy Adjustment: Use multiple sensors to receive students' voice, gesture, expression, and touch information, execute learning resource retrieval operations, and the constructed emotion recognition model recognizes students' emotions and adjusts interaction strategies according to the emotions; S500, Real-time Feedback and Adjustment: Collect students' learning resource usage, project progress, test scores, interaction data, and emotional changes during the learning process, analyze students' learning progress and status using the learning dynamics comprehensive evaluation formula, and adjust the personalized learning path, optimize the project-based learning plan, and improve the multimodal interaction strategy based on the analysis results.

Images