AI-based teaching system for the elderly

Abstract

This invention relates to the field of elderly education technology and provides an artificial intelligence-based teaching system for the elderly. The system tests the user's cognitive level to obtain cognitive data, matches the user with a corresponding teaching mode, and ensures the user can understand the teaching content well. It acquires the user's preference characteristics, determines the breadth of teaching based on these characteristics, and determines the depth of teaching based on the cognitive data. Teaching content is generated based on the breadth and depth of teaching to achieve personalized teaching and expand the audience. The system divides the teaching content into learned content, skipped content, and pending content. A memory curve is generated based on the learned content, and the pending content is dynamically adjusted based on the memory curve and skipped content to ensure the teaching of the pending content is targeted. It provides multiple age-friendly interactive methods to improve the user's learning experience. The system monitors the user's physiological health indicators in real time. If abnormal physiological health indicators are detected, the teaching progress is adjusted and / or a preset emergency contact is contacted to ensure the user's learning status.

Description

Technical Field

[0001] This invention relates to the field of elderly education technology, and in particular to an artificial intelligence-based teaching system for the elderly. Background Technology

[0002] Senior education refers to educational activities aimed at enabling older adults to continue learning. Within the overall education system, it is as important as basic education, vocational and technical education, general higher education, and adult education. As my country's society continues to age, the number of elderly people is gradually increasing, and senior education is gaining popularity among this demographic.

[0003] Current elderly education is usually conducted through offline interest groups or senior universities, which prevents some elderly people who cannot go out from learning. Moreover, the teaching is mostly focused on health preservation and art, resulting in a lack of variety in teaching content. Since the cognitive abilities, learning needs, and economic capabilities of different elderly people vary, the limited teaching content and methods cannot meet the needs of most elderly people, thus leading to a relatively narrow audience for current elderly education.

[0004] Therefore, it is necessary to propose an elderly education system to address the above problems. Summary of the Invention

[0005] Given that the existing teaching content and methods of elderly education are relatively monotonous, and considering the differences in cognitive abilities, learning capabilities, learning needs, and economic resources among different elderly people, the monotonous teaching content and methods are insufficient to meet the needs of most elderly people, resulting in a relatively narrow audience for existing elderly education. One of the purposes of this application is to provide an artificial intelligence-based elderly education system that expands the audience of elderly education among the elderly population by providing diversified teaching content and personalized teaching tailored to the actual situation of the elderly.

[0006] To achieve the above objectives, this application adopts the following technical solution:

[0007] An AI-based teaching system for the elderly includes:

[0008] The user feature analysis module is used to test and analyze the user's cognitive level to obtain cognitive data; based on the cognitive data, a corresponding teaching mode is matched for the user.

[0009] The teaching content generation module is used to acquire users' preference characteristics, determine the breadth of teaching based on the preference characteristics, and determine the depth of teaching based on the cognitive data; and generate teaching content based on the teaching breadth and depth.

[0010] The learning status detection module is used to divide the teaching content into learned content, skipped content, and pending content; and to generate a memory curve based on the learned content, and to dynamically adjust the pending content based on the memory curve and the skipped content;

[0011] An age-friendly interaction module is provided to offer a variety of interaction methods;

[0012] The health monitoring module is used to monitor the user's physiological health indicators in real time; determine whether the physiological health indicators are abnormal, and if so, adjust the teaching progress and / or contact the preset emergency contact person.

[0013] In one embodiment disclosed in this application, the user feature analysis module is used to test and analyze the user's cognitive level to obtain cognitive data; based on the cognitive data, it matches a corresponding teaching mode for the user, including:

[0014] Collect game interaction data related to cognitive level, and calculate the user's logical reasoning index, attention stability index and memory capacity index based on the game interaction data, and use them as cognitive data;

[0015] The teaching modes, from easy to difficult, include an interactive question-and-answer video mode, an interactive scenario simulation mode, and a self-exploration mode. Each mode has pre-set initial teaching information, including index threshold ranges and teaching data parameters. The index threshold ranges include logical reasoning index threshold ranges, attention stability index threshold ranges, and memory capacity index threshold ranges. The teaching data parameters include cue explicitness, question interval density, time pressure, environmental openness, failure tolerance, and the number of hidden clues. Based on the index threshold ranges that the logical reasoning index, attention stability index, and memory capacity index fall into, the corresponding teaching mode and teaching data parameters are matched to the user.

[0016] In one embodiment disclosed in this application, it further includes: an adjustment module, configured to evaluate the user's actual performance score under a matched teaching mode, compare the actual performance score with an expected performance threshold range, and dynamically adjust the teaching mode and / or teaching data parameters, including:

[0017] Within a preset time period, the system monitors user operation data in real time under a matched teaching mode. Based on this data, it calculates the user's accuracy score, efficiency score, fluency score, and pattern specificity score. The pattern specificity score includes question-answer interval acceptability, decision consistency, and exploration coverage. The accuracy score, efficiency score, fluency score, and pattern specificity score are then linearly weighted and summed to obtain the actual performance score P. real :

[0018] ,

[0019] in, , , and All are weighted coefficients, where C is the accuracy score, E is the efficiency score, F is the fluency score, and S is the pattern specificity score.

[0020] The actual performance score is compared with the expected performance threshold range. If the actual performance score exceeds the expected performance threshold range, the matching teaching mode is switched to a more difficult teaching mode. If the actual performance score is lower than the expected performance threshold range, the matching teaching mode is switched to a less difficult teaching mode. If the actual performance score is within the expected performance threshold range, the teaching mode is not switched, and the teaching data parameters are adjusted.

[0021] In one embodiment of this application, the teaching content generation module is further configured to pre-set teaching feature tags for all the teaching subjects according to the teaching characteristics and content characteristics of the teaching subjects, and to classify the difficulty level of each teaching subject according to the difficulty level to obtain the difficulty level of the teaching subject.

[0022] In one embodiment disclosed in this application, the teaching content generation module is used to acquire user preference features, determine teaching breadth based on the preference features, and determine teaching depth based on the cognitive data; generating teaching content based on the teaching breadth and teaching depth includes:

[0023] The user's preference features are obtained, including the user's needs and interests. The needs and interests are compared with the teaching feature tags to filter out target teaching subjects that match the needs and interests, thereby determining the breadth of teaching.

[0024] The logical reasoning index, attention stability index, and memory capacity index are linearly weighted to obtain a cognitive level. The difficulty level of the target teaching subject is matched with the cognitive level to determine the teaching depth. Teaching content is generated based on the target teaching subject and its difficulty level.

[0025] In one embodiment disclosed in this application, the learning state detection module is used to divide the teaching content into learned content, skipped content, and pending content; generate a memory curve based on the learned content; and dynamically adjust the pending content based on the memory curve and the skipped content, including:

[0026] Based on learning progress and user actions, the teaching content is divided into learned content, skipped content, and pending content. Review content for different time periods is generated according to the learning time of the learned content. Memory curves are generated based on the user's learning of the review content for different time periods. The pending content is dynamically adjusted based on the memory curves and the skipped content.

[0027] In one embodiment disclosed in this application, the age-friendly interaction module includes a dialect voice database, and the interaction methods include physical button interaction, voice interaction, image interaction, remote assistance interaction, touch interaction and gesture interaction that can be performed in a coordinated manner, and the dialect voice database is associated with the voice interaction.

[0028] In one embodiment disclosed in this application, the age-friendly interactive module further includes an intelligent correction function and an automatic recovery function for erroneous operations.

[0029] In one embodiment disclosed in this application, the health monitoring module is used to monitor the user's physiological health indicators in real time; determine whether the physiological health indicators are abnormal, and if so, adjust the teaching schedule and / or contact a preset emergency contact person, including:

[0030] The system uses embedded biosensors to monitor users' physiological health indicators in real time; it pre-sets the abnormality levels of the physiological health indicators, determines whether the physiological health indicators are in an abnormal state, and if so, determines the abnormal state level based on the abnormality level, and adjusts the teaching progress and / or contacts the preset emergency contact person based on the abnormal state level.

[0031] In one embodiment disclosed in this application, an anti-addiction module is also included, which is used to record the duration of a user's single learning session. When the duration of a single learning session reaches a preset duration, the teaching mode is paused and a relaxation mode is entered. After a preset time, the user can re-enter the learning mode.

[0032] Compared with existing technologies, the beneficial effects of this invention are as follows: By testing and analyzing the user's cognitive level, cognitive data is obtained, and a corresponding teaching mode is matched based on the user to ensure that the user can understand the teaching content well; the user's preference characteristics are obtained, the teaching breadth is determined according to the preference characteristics, and the teaching depth is determined according to the cognitive data. Teaching content is generated based on the teaching breadth and depth to achieve personalized teaching and expand the audience; the teaching content is divided into learned content, skipped content, and pending content, a memory curve is generated based on the learned content, and the pending content is dynamically adjusted based on the memory curve and skipped content to ensure the teaching relevance of the pending content; multiple age-friendly interactive methods are provided to improve the user's learning experience; and the user's physiological health indicators are monitored in real time to determine whether the physiological health indicators are abnormal. If so, the teaching progress is adjusted and / or a preset emergency contact is contacted to ensure the user's learning status and physical and mental health. Attached Figure Description

[0033] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0034] Figure 1 A schematic diagram of the framework of the AI-based elderly teaching system provided in this application;

[0035] Figure 2 A schematic diagram of the workflow of the health monitoring module provided in this application. Detailed Implementation

[0036] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of the invention. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0037] The terms “comprising” and “having”, and any variations thereof, used in this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the steps or units listed, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to such process, method, product, or apparatus.

[0038] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0039] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0040] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0041] Figure 1 A schematic diagram illustrating the framework of the AI-based elderly education system provided in this application. This AI-based elderly education system includes:

[0042] The user feature analysis module is used to test and analyze users' cognitive levels to obtain cognitive data; based on this cognitive data, a corresponding teaching mode is matched for the user.

[0043] The teaching content generation module is used to obtain users' preference characteristics, determine the breadth of teaching based on these preference characteristics, and determine the depth of teaching based on the cognitive data; and generate teaching content based on the breadth and depth of teaching.

[0044] The learning status detection module is used to divide the teaching content into learned content, skipped content, and pending content; and to generate a memory curve based on the learned content, and to dynamically adjust the pending content based on the memory curve and the skipped content.

[0045] An age-friendly interaction module is provided to offer a variety of interaction methods;

[0046] The health monitoring module is used to monitor the user's physiological health indicators in real time; determine whether the physiological health indicators are abnormal, and if so, adjust the teaching progress and / or contact the preset emergency contact person.

[0047] This AI-based elderly education system tests and analyzes users' cognitive levels to obtain cognitive data. Based on this data, it matches users with corresponding teaching models to ensure they can understand the content effectively. It also acquires users' preference characteristics, determines the breadth of instruction based on these characteristics, and determines the depth of instruction based on the cognitive data. Based on this breadth and depth, it generates personalized teaching content, expanding the audience reach. The system categorizes teaching content into learned content, skipped content, and pending content. It generates a memory curve based on the learned content and dynamically adjusts the pending content based on the memory curve and skipped content, ensuring the targeted nature of the teaching. Furthermore, it provides various age-friendly interactive methods to enhance the user's learning experience. Finally, it monitors users' physiological health indicators in real time, determining if any abnormalities occur. If so, it adjusts the teaching progress and / or contacts a pre-set emergency contact, thereby ensuring the user's learning status and physical and mental well-being.

[0048] Preferably, the user feature analysis module is used to test and analyze the user's cognitive level to obtain cognitive data; based on the cognitive data, it matches the user with a corresponding teaching mode, including: collecting game interaction data related to cognitive level, and calculating the user's logical reasoning index, attention stability index and memory capacity index based on the game interaction data, which are used as cognitive data;

[0049] The teaching modes, arranged from easy to difficult, include an interactive question-and-answer video mode, an interactive scenario simulation mode, and a self-exploration mode. Each mode has pre-set initial teaching information, including index threshold ranges and teaching data parameters. The index threshold ranges include those for logical reasoning, attention stability, and memory capacity. The teaching data parameters include cue explicitness, question interval density, decision-making time pressure, environmental openness, failure tolerance, and the number of hidden clues. Based on the index threshold ranges that the logical reasoning, attention stability, and memory capacity indices fall into, the corresponding teaching mode and teaching data parameters are matched to the user.

[0050] In the above technical solution, users perform game tasks related to their cognitive level through a human-computer interaction interface. After completing the game test, a game log is generated. Based on the game log, game interaction data related to the user's cognitive level is collected. This game interaction data includes, but is not limited to, the number of correct / incorrect times, reaction time, reaction coefficient of variation, operation sequence (including the operation steps and number of steps performed by the user), backtracking steps (the number of times the user returns to previously visited states / steps or the average number of backtracking steps), and the number of task interruptions (the number of interruptions per unit time, such as switching windows or no operation timeout). The game interaction data is preprocessed, which includes calculating the accuracy rate, the mean and standard deviation of the reaction time (outliers need to be removed), the reaction coefficient of variation = standard deviation / mean, the path optimization ratio = optimal number of steps / operation steps (the optimal number of steps is the number of steps used in the preset optimal path), and the interruption rate = number of task interruptions / total task time. The longest correct operation step count (the number of consecutive error-free operation steps) is taken according to the operation sequence. The longest correct operation step count and backtracking steps are normalized, and the reaction coefficient of variation and interruption rate are mapped to [0,1]. The user's logical reasoning index is obtained by linearly weighting and summing the accuracy rate and path optimization ratio; the user's attention stability index is obtained by linearly weighting and summing the response variation coefficient and interruption rate; the user's memory capacity index is obtained by linearly weighting and summing the normalized backtracking steps and the backtracking steps with positive scores, thus obtaining the user's cognitive data.

[0051] The teaching modes are divided into three categories: interactive question-and-answer video mode, interactive scenario simulation mode, and self-exploration mode. The interactive question-and-answer video mode achieves teaching through interactive question-and-answer sessions with video content. The interactive scenario simulation mode allows users to learn through practice by simulating familiar daily situations for the elderly. The self-exploration mode allows users to learn at their own pace in an open learning environment. The difficulty of these three modes increases progressively. Preset index threshold ranges and teaching data parameters are provided for each mode. The teaching data parameters for the interactive question-and-answer video mode include, but are not limited to, the explicitness of prompts and the density of question intervals; the teaching data parameters for the interactive scenario simulation mode include, but are not limited to, scenario complexity and decision-making time pressure; and the teaching data parameters for the self-exploration mode include, but are not limited to, environmental openness, failure tolerance (whether to reset), and the number of hidden clues.

[0052] When the logical reasoning index, attention stability index, and memory capacity index all fall within the threshold range corresponding to the same teaching mode, that teaching mode is matched to the user. When these indices fall within the threshold ranges of two different teaching modes, the teaching mode with the higher index value is matched. When these indices fall within the threshold ranges of three different teaching modes, the teaching mode is matched according to the priority order of memory capacity index > attention stability index > logical reasoning index. This allows for flexible selection of the appropriate teaching mode based on the user's cognitive level, helping them better understand the teaching content. It should be noted that all teaching data parameters are divided into three levels: easy, moderate, and difficult. After matching a teaching mode, the moderate level of teaching data parameters is set. These parameters can be adjusted based on the user's real-time performance to achieve targeted and flexible teaching.

[0053] Preferably, it further includes: an adjustment module for evaluating the user's actual performance score in the matched teaching mode, comparing the actual performance score with an expected performance threshold range, and dynamically adjusting the teaching mode and / or teaching data parameters, including:

[0054] Within a preset time period, the system monitors user operation data in real time under the matched teaching mode. Based on this data, it calculates the user's accuracy score, efficiency score, fluency score, and pattern specificity score. The pattern specificity score includes question-answer interval acceptability, decision consistency, and exploration coverage. The accuracy score, efficiency score, fluency score, and pattern specificity score are then linearly weighted and summed to obtain the actual performance score P. real :

[0055] ,

[0056] in, , , and All are weighted coefficients, where C is the accuracy score, E is the efficiency score, F is the fluency score, and S is the pattern specificity score.

[0057] The actual performance score is compared with the expected performance threshold range. If the actual performance score exceeds the expected performance threshold range, the matching teaching mode is switched to a more difficult teaching mode; if the actual performance score is below the expected performance threshold range, the matching teaching mode is switched to a less difficult teaching mode; if the actual performance score is within the expected performance threshold range, the teaching mode is not switched, and the teaching data parameters are adjusted.

[0058] In the above technical solution, the operational data includes, but is not limited to, accuracy rate, reaction time, time from task start to submission, number of new scene nodes accessed by the user, and number of requests for help. The operational data is normalized to obtain accuracy score, efficiency score, fluency score, and pattern specificity score. Among them, the pattern specificity score corresponding to the interactive question-and-answer video mode is the question-and-answer interval acceptability (characterizing whether the user responds quickly after asking a question, expressed as the proportion of each answer response time < expected duration), the pattern specificity score corresponding to the interactive scenario simulation mode is the decision consistency (characterizing the degree of logical consistency between consecutive decisions, compared through semantic similarity), and the pattern specificity score corresponding to the autonomous exploration mode is the exploration coverage (measuring the user's exploration initiative, expressed as the proportion of the number of new scene nodes accessed to the total number of nodes). The accuracy score, efficiency score, fluency score, and pattern specificity score are linearly weighted and summed to obtain the actual performance score. This score is then used to adjust the user's teaching mode and / or teaching data parameters to ensure targeted teaching throughout the entire process. Specifically, when the actual performance score falls within the expected performance threshold range, the teaching data parameters are adjusted based on the score's position within this range (upstream, midstream, and downstream). For example, if the actual performance score is upstream of the expected performance threshold range, the teaching data parameters are adjusted to easy; if the actual performance score is midstream, no adjustment is made; and if the actual performance score is downstream, the teaching data parameters are adjusted to hard. It should be noted that when the difficulty of the teaching mode or teaching data parameters is already at the easiest or hardest level, the teaching depth needs to be adjusted.

[0059] Preferably, the teaching content generation module is also used to pre-set teaching feature tags for all teaching subjects according to their teaching characteristics and content characteristics, and to classify each teaching subject into difficulty levels according to their difficulty to obtain the difficulty level of the teaching subject.

[0060] In the aforementioned technical solutions, with the development of science and technology and the progress of society, new knowledge is constantly emerging, leading to a widespread phenomenon of the elderly becoming disconnected from society. To address this, this application adopts a multi-faceted approach to teaching subjects, pre-setting teaching characteristic tags for all subjects based on their teaching characteristics and content. These tags include, but are not limited to, basic skills enhancement, smart technology, psychology, health and wellness, fraud prevention, leisure and entertainment, language, and culture. Basic skills enhancement includes, but is not limited to, practical teaching related to daily life, such as how to use household appliances and how to ride public transportation. Smart technology includes, but is not limited to, smartphone applications, tablet operation, and cybersecurity knowledge. By diversifying teaching subjects, the learning needs of more elderly people in contemporary society are met, bridging the gap between the elderly and modern society and enhancing their sense of accomplishment. Furthermore, each teaching subject is divided into three difficulty levels: basic, intermediate, and exploratory, thus matching different levels of difficulty to users with different cognitive levels and achieving differentiated teaching.

[0061] Preferably, the teaching content generation module is used to acquire user preference characteristics, determine the breadth of instruction based on these preference characteristics, and determine the depth of instruction based on the cognitive data; and generate teaching content based on the breadth and depth of instruction, including:

[0062] The method obtains user preference characteristics, which include user needs and interests. These needs and interests are compared with the teaching feature tags to filter out target teaching subjects that match the needs and interests, thereby determining the breadth of teaching.

[0063] The logical reasoning index, attention stability index, and memory capacity index are linearly weighted to obtain the cognitive level. The difficulty level of the target teaching subject is then matched with the cognitive level to determine the teaching depth. Teaching content is generated based on the target teaching subject and its difficulty level.

[0064] In the above technical solution, demand characteristics and interest characteristics are compared with teaching feature tags through semantic similarity. When the semantic similarity reaches a preset similarity threshold, the corresponding teaching subject is determined as the target teaching subject, thus determining the breadth of teaching. Logical reasoning index, attention stability index, and memory capacity index are linearly weighted and summed to obtain cognitive levels, which include beginner, intermediate, and advanced levels. Basic difficulty levels of target teaching subjects are matched to users at the beginner cognitive level, intermediate difficulty levels to users at the intermediate cognitive level, and exploratory difficulty levels to users at the advanced cognitive level, thus determining the depth of teaching. Teaching content is generated based on the target teaching subject and its difficulty level, achieving personalized teaching.

[0065] Preferably, the learning status detection module is used to divide the teaching content into learned content, skipped content, and pending content; generate a memory curve based on the learned content; and dynamically adjust the pending content based on the memory curve and the skipped content, including:

[0066] Based on learning progress and user actions, the teaching content is divided into learned content, skipped content, and pending content. Review content for different time periods is generated based on the learning time of the learned content. Memory curves are generated based on the user's learning of the review content for different time periods. The pending content is dynamically adjusted based on the memory curves and the skipped content.

[0067] In the above technical solution, teaching content is divided into learned content, skipped content, and pending content based on learning progress. Teaching content with a learning progress rate above 85% is classified as learned content, while content with a learning progress rate below 85% is classified as pending content. Teaching content that the user triggers the skip button is classified as skipped content (i.e., content the user is not interested in). Review content for different time periods is generated based on the learning time of the learned content. These review periods may include, but are not limited to, 15 minutes, 1 hour, 1 day, or 1 week after learning. The user's memory retention rate for the review content in each time period is measured and recorded, and the curve formed by all memory retention rates is used as a memory curve. Based on the memory curve and skipped content, the pending content is dynamically adjusted, including adjusting the frequency and number of times learned knowledge points appear in the pending content and deleting teaching content related to the skipped content, thereby ensuring that the pending content is tailored to different users.

[0068] Preferably, the age-friendly interactive module includes a dialect voice database, and the interaction methods include physical button interaction, voice interaction, image interaction, remote assistance interaction, touch interaction and gesture interaction that can be performed in a coordinated manner, and the dialect voice database is associated with the voice interaction.

[0069] In the aforementioned technical solutions, given the varying cognitive levels among the elderly, this application integrates multiple interaction methods to reduce barriers to participation in the digital age and enhance their learning experience. Furthermore, due to regional differences and educational limitations, most elderly individuals do not speak standard Mandarin. This application utilizes a dialect speech database to support dialect-based speech interaction for the elderly, thereby further improving the user's learning experience and ensuring teaching efficiency.

[0070] Preferably, the age-friendly interactive module also includes an intelligent correction function and an automatic recovery function for erroneous operations.

[0071] In the above technical solutions, most elderly people have difficulties in digital participation. The accuracy and efficiency of interaction can be guaranteed through intelligent correction and automatic recovery of misoperation functions.

[0072] Preferably, the health monitoring module is used to monitor the user's physiological health indicators in real time; determine whether the physiological health indicators are abnormal, and if so, adjust the teaching schedule and / or contact a preset emergency contact person, including:

[0073] The system uses embedded biosensors to monitor users' physiological health indicators in real time; it pre-sets the abnormality level of the physiological health indicator, determines whether the physiological health indicator is in an abnormal state, and if so, determines the abnormal state level based on the abnormality level, and adjusts the teaching progress and / or contacts the preset emergency contact person based on the abnormal state level.

[0074] In the above technical solution, the physiological health indicators include, but are not limited to, multiple indicators such as heart rate, blood oxygen, and heart rate variability. An abnormality level is pre-set for each indicator, which includes an abnormality threshold range and a corresponding abnormality level. The abnormality level can be, but is not limited to, level one, level two, and level three. When any one or more physiological health indicators exceed the corresponding abnormality threshold range, it can be determined that one or more of those physiological health indicators are in an abnormal state. When only one indicator is in an abnormal state, the abnormality level is determined based on the abnormality level of the corresponding indicator. When multiple indicators are in an abnormal state, the abnormality level corresponding to the indicator with the highest abnormality level is taken as the current abnormality level. This abnormality level includes low-level, medium-level, and high-level abnormal states, and these low-level, medium-level, and high-level abnormal states correspond to levels one, two, and three, respectively. Wherein, as Figure 2 As shown, a low-level abnormal state will pause teaching with a countdown and switch the teaching mode to relaxation mode. Teaching mode will resume after a preset time, and relaxation mode includes, but is not limited to, playing soothing music. A medium-level abnormal state will forcibly pause teaching. Teaching mode can resume after the health monitoring module checks all physiological health indicators and confirms they are normal. A high-level abnormal state will forcibly pause teaching and send GPS location and physiological health indicator data to a preset emergency contact. Teaching mode can only resume after confirmation from the preset emergency contact. By monitoring the user's physiological health indicators in real time, the system ensures the user's learning progress and physical health.

[0075] Preferably, it also includes an anti-addiction module, which records the duration of a user's single learning session. When the duration of a single learning session reaches a preset duration, the countdown pauses the teaching mode and enters the relaxation mode. After a preset time, the user can re-enter the learning mode.

[0076] In the above technical solution, in order to balance learning effectiveness and health and safety, intelligent intervention is used to prevent physical and mental damage to the elderly due to overuse, and they can only resume learning after a preset time of mandatory rest.

[0077] As can be seen from the above embodiments, this AI-based elderly education system tests and analyzes users' cognitive levels to obtain cognitive data. Based on this cognitive data, it matches corresponding teaching modes to users to ensure that users can understand the teaching content well. It obtains users' preference characteristics, determines the breadth of teaching based on these characteristics, and determines the depth of teaching based on the cognitive data. Based on the breadth and depth of teaching, it generates teaching content to achieve personalized teaching and expand the audience. It divides the teaching content into learned content, skipped content, and pending content. Based on the learned content, it generates a memory curve, and based on the memory curve and skipped content, it dynamically adjusts the pending content to ensure the teaching is targeted. It also provides a variety of age-friendly interactive methods to improve the user's learning experience. Furthermore, it monitors the user's physiological health indicators in real time to determine whether the physiological health indicators are abnormal. If so, it adjusts the teaching progress and / or contacts a preset emergency contact person to ensure the user's learning status and physical and mental health.

[0078] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. An artificial intelligence-based teaching system for the elderly, characterized in that: include: The user feature analysis module is used to test and analyze users' cognitive levels to obtain cognitive data; Based on the cognitive data, a corresponding teaching mode is matched for the user; The teaching content generation module is used to acquire users' preference characteristics, determine the breadth of teaching based on the preference characteristics, and determine the depth of teaching based on the cognitive data; and generate teaching content based on the teaching breadth and depth. The learning status detection module is used to divide the teaching content into learned content, skipped content, and pending content; and to generate a memory curve based on the learned content, and to dynamically adjust the pending content based on the memory curve and the skipped content; An age-friendly interaction module is provided to offer a variety of interaction methods; The health monitoring module is used to monitor the user's physiological health indicators in real time; determine whether the physiological health indicators are abnormal, and if so, adjust the teaching progress and / or contact the preset emergency contact person.

2. The artificial intelligence-based elderly education system according to claim 1, characterized in that, The user feature analysis module is used to test and analyze the user's cognitive level to obtain cognitive data; Based on the cognitive data, a corresponding teaching mode is matched for the user, including: Collect game interaction data related to cognitive level, and calculate the user's logical reasoning index, attention stability index and memory capacity index based on the game interaction data, and use them as cognitive data; The teaching modes, from easy to difficult, include an interactive question-and-answer video mode, an interactive scenario simulation mode, and a self-exploration mode. Each mode has pre-set initial teaching information, including index threshold ranges and teaching data parameters. The index threshold ranges include logical reasoning index threshold ranges, attention stability index threshold ranges, and memory capacity index threshold ranges. The teaching data parameters include cue explicitness, question interval density, decision-making time pressure, environmental openness, failure tolerance, and the number of hidden clues. Based on the index threshold ranges that the logical reasoning index, attention stability index, and memory capacity index fall into, the corresponding teaching mode and teaching data parameters are matched to the user.

3. The artificial intelligence-based elderly education system according to claim 2, characterized in that, Also includes: The adjustment module is used to evaluate the user's actual performance score under the matched teaching mode, compare the actual performance score with the expected performance threshold range, and dynamically adjust the teaching mode and / or teaching data parameters, including: Within a preset time period, the system monitors user operation data in real time under a matched teaching mode. Based on this data, it calculates the user's accuracy score, efficiency score, fluency score, and pattern specificity score. The pattern specificity score includes question-answer interval acceptability, decision consistency, and exploration coverage. The accuracy score, efficiency score, fluency score, and pattern specificity score are then linearly weighted and summed to obtain the actual performance score P. real : ， in, , , and All are weighted coefficients, where C is the accuracy score, E is the efficiency score, F is the fluency score, and S is the pattern specificity score. The actual performance score is compared with the expected performance threshold range. If the actual performance score exceeds the expected performance threshold range, the matching teaching mode is switched to a more difficult teaching mode. If the actual performance score is lower than the expected performance threshold range, the matching teaching mode is switched to a less difficult teaching mode. If the actual performance score is within the expected performance threshold range, the teaching mode is not switched, and the teaching data parameters are adjusted.

4. The artificial intelligence-based elderly teaching system according to claim 2, characterized in that: The teaching content generation module is also used to pre-set teaching feature tags for all the teaching subjects according to their teaching characteristics and content characteristics, and to classify each teaching subject into difficulty levels according to their difficulty to obtain the difficulty level of the teaching subject.

5. The artificial intelligence-based elderly education system according to claim 4, characterized in that, The teaching content generation module is used to acquire user preference characteristics, determine teaching breadth based on the preference characteristics, and determine teaching depth based on the cognitive data; and generate teaching content based on the teaching breadth and depth, including: The user's preference features are obtained, including the user's needs and interests. The needs and interests are compared with the teaching feature tags to filter out target teaching subjects that match the needs and interests, thereby determining the breadth of teaching. The logical reasoning index, attention stability index, and memory capacity index are linearly weighted to obtain a cognitive level. The difficulty level of the target teaching subject is matched with the cognitive level to determine the teaching depth. Teaching content is generated based on the target teaching subject and its difficulty level.

6. The artificial intelligence-based elderly education system according to claim 1, characterized in that, The learning status detection module is used to divide the teaching content into learned content, skipped content, and pending content; A memory curve is generated based on the learned content, and the content to be learned is dynamically adjusted based on the memory curve and the skipped content, including: Based on learning progress and user actions, the teaching content is divided into learned content, skipped content, and content to be learned; Based on the learning time of the learned content, review content for different time periods is generated. Based on the user's learning of the review content for different time periods, a memory curve is generated. Based on the memory curve and the skipped content, the content to be learned is dynamically adjusted.

7. The artificial intelligence-based elderly education system according to claim 1, characterized in that: The age-friendly interactive module includes a dialect voice database, and the interaction methods include physical button interaction, voice interaction, image interaction, remote assistance interaction, touch interaction, and gesture interaction that can be performed in a coordinated manner. The dialect voice database is associated with the voice interaction.

8. The artificial intelligence-based elderly education system according to claim 1, characterized in that: The age-friendly interactive module also includes intelligent correction functions and automatic recovery functions for erroneous operations.

9. The artificial intelligence-based elderly education system according to claim 1, characterized in that, The health monitoring module is used to monitor the user's physiological health indicators in real time; Determine whether the physiological health indicators are abnormal. If so, adjust the teaching schedule and / or contact the pre-designated emergency contact person, including: Real-time monitoring of users' physiological health indicators through embedded biosensors; The abnormality level of the physiological health indicator is preset, and it is determined whether the physiological health indicator is in an abnormal state. If so, the abnormal state level is determined based on the abnormality level, and the teaching progress is adjusted and / or the preset emergency contact person is contacted based on the abnormal state level.

10. The artificial intelligence-based elderly education system according to claim 1, characterized in that, It also includes an anti-addiction module, which records the duration of a user's single study session. When the duration of a single study session reaches the preset time, the countdown pauses the teaching mode and enters the relaxation mode. After the preset time, the user can re-enter the learning mode.

Images