Virtual teaching system and method based on dynamic knowledge graph

Abstract

The invention relates to the technical field of computer education, and particularly discloses a virtual teaching system and method based on a dynamic knowledge graph, the system comprises a knowledge graph database, a camera and an AR teaching instrument, the knowledge graph database is constructed based on knowledge resources involved in professional courses; the camera is used for capturing a real scene image; and the AR teaching instrument runs a pre-designed AR experiment model, superposes and fuses the reality scene image for display, and controls the attitude of the AR experiment model according to the input control parameters. A system framework of knowledge resources related to professional courses is formed in a knowledge graph form, theories and structures of knowledge of the professional courses are systematically presented, immersive teaching of an AR experiment model constructed based on a knowledge graph database is realized through an AR teaching instrument, and real scene images captured by a camera are superposed, so that the teaching effect of the AR experiment model is improved. And a highly interactive and immersive learning scene is created, so that the teaching interactivity and the learning effect are improved.

Description

Technical Field

[0001] The present invention relates to the field of computer education technology, and in particular to a virtual teaching system and method based on a dynamic knowledge graph. Background Art

[0002] Professional course teaching for students is the core educational task of technical colleges, and the technical fields involved in professional course teaching are very broad. For example, the teaching of automatic control courses involves a lot of theoretical and practical knowledge. How to systematically and visually display the course knowledge structure to clarify the relationship between knowledge points, and effectively map these abstract concepts into teaching practice, and provide a self-exploration learning ecology close to real scenes for course teaching, is the key difficulty of this type of teaching.

[0003] Taking the automatic control course as an example, knowledge graph, as an advanced data management and integration technology, models data through graph structure and represents entities and their relationships in the data in the form of nodes and edges. It can not only reflect the semantic relationship between data more intuitively, but also reveal implicit knowledge through the reasoning ability of graph, thereby improving data integration efficiency and query accuracy, and providing a feasible solution for the teaching of control theory, various components of control systems and their interrelationships in the automatic control principle course. In terms of teaching practice, the current automatic control principle course widely uses high-performance simulation tools such as MATLAB and Mathematics to process complex data and generate variable curve graphs, but its limitations in graphically expressing the overall picture of automatic control are particularly prominent. Augmented reality (AR) technology can seamlessly integrate virtual information and real environment information, giving users an immersive experience of intuitively operating 3D virtual objects, greatly improving the realism and interactivity of image display, and opening up a new way to develop low-cost, portable virtual teaching instruments.

[0004] However, the application of knowledge graphs and AR technology in the field of teaching still faces the following challenges:

[0005] 1. Lack of dynamic association between semantic reasoning and knowledge points: The existing teaching system has shortcomings in natural language processing and semantic analysis, and is unable to automatically extract implicit relationships in teaching data and achieve adaptive reconstruction and dynamic updating of knowledge graphs.

[0006] 2. Deep integration of system design and educational theory: The current development of AR systems relies heavily on engineers with strong technical backgrounds, but they lack a deep understanding of educational theory and practice, which makes it difficult for the system to accurately meet teaching needs and lacks targeted teaching design.

[0007] 3. Accurate positioning and stability in complex interactive scenarios: Ensuring accurate alignment and stable display of AR content in real environments involves complex target recognition and positioning algorithms, and it is necessary to overcome problems such as image jitter and positioning errors to ensure the continuity and immersion of the user experience.

[0008] Based on this problem, the present invention proposes a virtual teaching solution based on dynamic knowledge graph. Summary of the invention

[0009] In view of the technical problems in the prior art, the present invention provides a multi-station positioning method and system based on time-sharing inquiry.

[0010] In a first aspect, the present invention includes a virtual teaching system based on a dynamic knowledge graph, the system including a knowledge graph database, a camera and an AR teaching instrument; wherein:

[0011] The knowledge graph database is constructed based on the knowledge resources involved in the professional courses, and contains knowledge structures of different levels and logical relationships between nodes;

[0012] The camera is used to capture images of real scenes;

[0013] The AR teaching instrument runs a pre-designed AR experimental model and superimposes and fuses the real scene image for display, and controls the posture of the AR experimental model according to the input control parameters; the AR experimental model is an instance constructed based on the knowledge graph database.

[0014] Furthermore, the process of constructing the knowledge graph database based on the knowledge resources involved in the professional courses includes:

[0015] Taking the content of professional courses as the core, construct the overall architecture of the knowledge graph; the overall architecture at least includes the storage structure, data model and query language of the knowledge graph;

[0016] Use natural language processing technology to pre-process the text contained in knowledge resources and extract key concepts, entities and attributes;

[0017] The knowledge graph is represented by a graph data structure. Nodes are designed to represent key concepts and entities. The size and color of the nodes are used to reflect the knowledge structure at different levels. Edges are used to express the logical relationship between nodes to form the framework of the knowledge graph.

[0018] Furthermore, it also includes: the knowledge graph database is updated by regularly scanning the knowledge resources involved in the professional courses.

[0019] Furthermore, the AR experimental model is a spacecraft experimental model constructed based on the automatic control principle course; setting the spacecraft experimental model includes:

[0020] Set the attitude quaternion of spacecraft i relative to the inertial system to q 0，i ∈R,q i =[q 1，i ,q 2，i ,q 3，i ] T ∈R 3 , R is a real number;

[0021] The reference attitude quaternion is

[0022] The preset difference between the target posture and the reference posture is d 0，i ∈R,d i =[d 1，i , d 2，i , d 3，i ] T ∈R 3 ;

[0023] The target posture is set as:

[0024]

[0025] in,

[0026]

[0027] definition Q i The first-order derivative of the control torque on spacecraft i is defined as T i ∈R 3 , the disturbance torque is defined as δ Ti ∈R 3 , Ji∈R 3×3 Represents the moment of inertia matrix of spacecraft i. Define the following auxiliary variable matrix:

[0028]

[0029]

[0030] Then the attitude kinematics of spacecraft i is and attitude dynamics for:

[0031]

[0032]

[0033] in,

[0034]

[0035] Defining variables And define the first backstepping control state variable X 1，i and the second backstepping control state variable X 2，i satisfy:

[0036] X 1，i =ε i -Q * ;X 1，i ∈R 4 ;

[0037] X 2，i ∈R 4 ;

[0038]

[0039] Among them, f(X 1，i )∈R 4 , is a virtual control function, 5<k1<1 and k2>1 are the ratios of two odd integers, γ1>0, γ2>0;

[0040] Define the following parameters:

[0041]

[0042]

[0043]

[0044]

[0045]

[0046]

[0047]

[0048] Given the following assumptions:

[0049] Assumption 1 Considering the fuel consumption, the spacecraft moment of inertia matrix J i The spacecraft model should satisfy the following constraints:

[0050] ||h i (Q i , v i )||≤kh ||v i || 2 , where k J , k g , and k h is a positive constant, and I3 is represented by a 3*3 identity matrix;

[0051] Assume that 2 spacecraft i is subject to an unknown small disturbance torque δ Ti The following inequality should be satisfied:

[0052]

[0053] in,

[0054] The fixed-time attitude control law of spacecraft i is set as:

[0055] T i =sat(u i ),

[0056] in:

[0057]

[0058] Δu i =T i -u i ;

[0059]

[0060] Among them, u a，1i ,u a，2i and sat(u i )for:

[0061]

[0062]

[0063] and κ>1,m1>0,m2>0,k3>0,constant u imax ∈R ＞0 represents the maximum control torque of the single axis output of spacecraft i, and the adaptive compensation law x a，i Defined as:

[0064]

[0065] Among them, h1, h2, h3 and k3 satisfy the inequality: h3>1;

[0066] The Lyapunov function is defined as follows:

[0067] Define the first backstepping control state variable X 1，i The Lyapunov function V 1，i As follows:

[0068]

[0069] The power integration technique is used to define the second backstepping control state variable X 2，i The Lyapunov function V 2，i As follows:

[0070]

[0071] in:

[0072]

[0073] Definition and adaptive compensation law x a，i The Lyapunov function V 3，i As follows:

[0074]

[0075] Define the Lyapunov function V a，i as follows:

[0076] V a，i =V 1，i +V 2，i +V 3，i ;

[0077] This leads to the inequality:

[0078]

[0079] in,

[0080]

[0081] in and satisfy:

[0082]

[0083] satisfy:

[0084]

[0085]

[0086] The upper bound of ΔM is:

[0087]

[0088] Among them, λ min (J i ) is the matrix J i The minimum eigenvalue of Converges to the following residual set within a fixed time T1 that satisfies the constraints:

[0089]

[0090] Fixed time T1 satisfies:

[0091] Furthermore, the parameters for controlling the attitude of the spacecraft experimental model include: k1, k2, k3, κ, γ1, γ2, m1, m2, h1, h2, h3.

[0092] Furthermore, the AR teaching instrument runs a pre-designed AR experimental model and superimposes and fuses the real scene image for display, including:

[0093] Acquire real scene images and design positioning marks, extract feature descriptors from real scene images containing positioning marks, and calculate the position information and posture information of the positioning marks in the three-dimensional scene;

[0094] Using the position information and posture information of the positioning marker as a reference, the AR experimental model is drawn in the three-dimensional scene through a transformation matrix and jitter smoothing is performed;

[0095] The AR experimental model is superimposed and fused with the real scene image for display.

[0096] Furthermore, the position information and posture information of the positioning mark in the three-dimensional scene are expressed as:

[0097] Position information: R = [r1 r2 r3] T ∈R 3 ;

[0098] Posture information: q0∈R,q i =[q 1，i ,q 2，i ,q 3，i ] T ∈R 3 .

[0099] Furthermore, the AR experimental model is drawn in the three-dimensional scene by using the transformation matrix and performing jitter smoothing processing, with the position information and posture information of the positioning mark as a reference, including:

[0100] The jitter sampling time is t k , k∈N≥0 , the position information and posture information at time k are represented as R(k) and Q(k) respectively, and the definition is

[0101]

[0102] After smoothing, time t, t∈[t k , t k+1 The position information and posture information of the positioning mark of ] in the three-dimensional scene are expressed as:

[0103]

[0104] The position information R(t) and posture information Q(t) of the positioning marker in the three-dimensional scene are used as references, and the relative displacement matrix R t (t) = [r 1，t r 2,t r 3,t ] T and the relative attitude rotation matrix Q t (t) = [q 0，t ,q 1，t ,q 2，t ,q 3，t ] T Transformation, to obtain the position information R of the AR teaching instrument in the three-dimensional scene k (t) = [r 1，k r 2,k r 3,k ] T and attitude information Q k (t) = [q 0，k ,q 1,k ,q 2，k ,q 3，k ] T :

[0105] R k (t) = R(t) + R t (t);

[0106]

[0107] In the three-dimensional scene, the position information is R k (t) Q k (t) Plotting the spacecraft experimental model for attitude information.

[0108] Furthermore, the spacecraft experimental model is drawn using Blender software and adopts a cylindrical truss-type load-bearing structure to support the replacement of loads. The solar panels are installed on a fixed support plate on the side of the satellite platform, the antenna is deployed on the outer surface of the payload bay, and the two propellant tanks are placed between the diagonal bars inside the truss.

[0109] In a second aspect, the present invention also includes a virtual teaching method based on a dynamic knowledge graph, and the method is implemented based on the aforementioned virtual teaching system based on a dynamic knowledge graph.

[0110] The present invention provides a virtual teaching system and method based on a dynamic knowledge graph. The system includes a knowledge graph database, a camera and an AR teaching instrument. The knowledge graph database is constructed based on the knowledge resources involved in professional courses, and the knowledge graph database contains knowledge structures of different levels and logical relationships between nodes; the camera is used to capture real scene images; the AR teaching instrument runs a pre-designed AR experimental model and superimposes and fuses the real scene image for display, and controls the posture of the AR experimental model according to the input control parameters; the AR experimental model is an instance constructed based on the knowledge graph database. The present invention adopts the form of knowledge graph to form a system framework of knowledge resources involved in professional courses, systematically presents the theory and structure of professional course knowledge, can clearly show the logical relationship between knowledge points, realizes the organic aggregation of knowledge, and improves the teaching quality of professional courses. The present invention also realizes immersive teaching of AR experimental models based on knowledge graph database through AR teaching instruments, superimposes real-life scene images captured by cameras, creates highly interactive and immersive learning scenes, enhances the intuitiveness and operability of learning materials, and enhances the immersion and authenticity of learning. The comprehensive virtual teaching method effectively breaks through the limitations of traditional teaching, improves the efficiency of transforming theoretical knowledge into practical skills, and realizes the concretization of abstract concepts, providing students with theoretical and practical training opportunities anytime and anywhere, thereby improving teaching interactivity and learning outcomes. BRIEF DESCRIPTION OF THE DRAWINGS

[0111] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0112] Figure 1 A schematic diagram of the backbone architecture of a knowledge graph according to an embodiment of the present invention;

[0113] Figure 2 A schematic diagram of a process for optimizing a knowledge graph based on a genetic algorithm according to an embodiment of the present invention;

[0114] Figure 3 This is a display effect diagram of an AR experimental model according to an embodiment of the present invention. DETAILED DESCRIPTION

[0115] The following will be combined with the accompanying drawings in the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work belong to the protection scope of the present invention.

[0116] A virtual teaching system based on a dynamic knowledge graph in an embodiment of the present invention includes a knowledge graph database, a camera, and an AR teaching instrument; wherein:

[0117] The knowledge graph database is constructed based on the knowledge resources involved in professional courses. The knowledge graph database contains knowledge structures at different levels and the logical relationships between nodes.

[0118] A camera, used to capture images of real scenes;

[0119] AR teaching equipment runs a pre-designed AR experimental model and superimposes and fuses real-life scene images for display, and controls the posture of the AR experimental model according to input control parameters; the AR experimental model is an instance built based on a knowledge graph database.

[0120] The present invention adopts the form of knowledge graph to form a system framework of knowledge resources involved in professional courses, comprehensively covers the key concepts, entities and their complex logical relationships in this field, systematically presents the theory and structure of professional course knowledge, can clearly show the logical relationship between knowledge points, realizes the organic aggregation of knowledge, and improves the teaching quality of professional courses. The present invention also realizes immersive teaching of AR experimental models built based on knowledge graph database through AR teaching instruments, superimposes real-life scene images captured by cameras, creates highly interactive and immersive learning scenes, enhances the intuitiveness and operability of learning materials, and also enhances the immersion and authenticity of learning. The comprehensive virtual teaching method effectively breaks through the limitations of traditional teaching, improves the efficiency of transforming theoretical knowledge into practical skills, and also realizes the concretization of abstract concepts, providing students with theoretical and practical training opportunities anytime and anywhere, thereby improving teaching interactivity and learning outcomes.

[0121] This embodiment does not limit the specific technical direction of the professional courses corresponding to the knowledge graph database. The teaching system of the embodiment of the present invention is applicable to teaching work in all walks of life. For example, the undergraduate automatic control major established by universities includes the study of the professional course "Principles of Automatic Control". The knowledge graph database of the embodiment of the present invention can be based on the Principles of Automatic Control course to construct a highly structured knowledge graph database.

[0122] Specifically, the process of constructing the knowledge graph database based on the knowledge resources involved in professional courses includes:

[0123] (1) With the content of professional courses as the core, build the overall architecture of the knowledge graph; the overall architecture at least includes key technical elements such as the storage structure, data model, and query language of the knowledge graph.

[0124] (2) Use natural language processing technology to preprocess the text contained in knowledge resources and extract key concepts, entities, and attributes.

[0125] The preprocessing of this step may include word segmentation, part-of-speech tagging, named entity recognition and other steps.

[0126] (3) The knowledge graph is represented by a graph data structure. Nodes are designed to represent key concepts and entities. The size and color of the nodes are used to reflect the knowledge structure at different levels. Edges are used to express the logical relationship between nodes to form the framework of the knowledge graph. Logical relationships include but are not limited to prerequisites, interdependencies, and a clear conceptual hierarchy.

[0127] like Figure 1 As shown, it is a schematic diagram of the backbone architecture of a knowledge graph according to an embodiment of the present invention.

[0128] The constructed knowledge graph can also be designed with a quality evaluation mechanism to conduct quality assessment and optimization adjustments, check the accuracy, completeness and consistency of nodes and edges, ensure the reliability of the knowledge graph, and make necessary optimization adjustments to the knowledge graph based on the evaluation results, such as correcting erroneous relationships, supplementing missing information, etc. Therefore, step (3) may include:

[0129] Utilize NLP technology and database processing technology to analyze and calculate node m and node The number of repeated words in the keywords between them is used to obtain the association weight

[0130] The constraint condition for the self-organization reconstruction of the knowledge graph is that, based on the backbone architecture, there must be more than one (including one) path connecting the nodes whose association weight is greater than the preset threshold, so at least one connecting path is established between two nodes whose association weight value is greater than the preset threshold (for example, the threshold is set to 4); the objective function is defined as the number of edges in the knowledge graph is as small as possible, and the distribution of edges is as uniform as possible to improve the clarity of the knowledge graph. Therefore, the objective function of the edge in the knowledge graph is defined as:

[0131]

[0132] Among them, N S is the total number of nodes, x mnis the number of edges between node m and node n, E(·) is the variance function; n∈[1,N S ],m∈[1,N S ]. )

[0133] Genetic algorithms are used to optimize the topological structure of the knowledge graph. Through competition, selection, crossover and mutation among individuals in the population, the optimal solution of the topological structure of the knowledge graph is gradually optimized. Figure 2 The figure shows a flow chart of optimizing the knowledge graph based on the genetic algorithm of this embodiment, including: initialization, generating the initial population (generating a preliminary knowledge graph), evaluating the scheme, and determining whether to terminate according to the evaluation score. If so, outputting the knowledge graph edge configuration result; if not, firstly performing random individual screening, and then performing random gene hybridization operation on the screened population to obtain the offspring population after random hybridization, and then performing random gene mutation to obtain the offspring population after random mutation, and finally obtaining a new population (new knowledge graph), and then performing scheme evaluation until the termination condition is met to obtain the final knowledge graph edge configuration result, which is also the optimal solution of the knowledge graph topological structure.

[0134] The embodiment of the present invention relies on the intelligent recommendation function of the knowledge graph to effectively enhance students' understanding and memory of the core concepts, key structures and dynamic processes of automatic control principles. Combined with the in-depth application of AR teaching instruments, it accurately maps the computer-generated AR experimental model to the real teaching scene, and intuitively demonstrates the teaching experimental process, thereby enriching teaching methods and experimental resources in a low-cost and convenient way, and promoting the deep integration of theory and practice.

[0135] The embodiment of the present invention also includes: the knowledge graph database is updated by regularly scanning the knowledge resources involved in the professional courses.

[0136] In order to ensure the timeliness and accuracy of the knowledge graph, we use natural language processing technology to regularly scan various database resources, and combine it with advanced database processing and text understanding algorithms to regularly mine and integrate knowledge resources from multiple sources such as academic journals, textbooks, and online forums. By adopting various association methods such as event association and logical association, we organically aggregate new knowledge elements, further realizing the organic aggregation of knowledge points and the periodic self-organizing update of the knowledge graph.

[0137] The embodiment of the present invention designs and implements a periodic self-organizing update mechanism for the knowledge graph, automatically integrating new knowledge elements into the existing knowledge graph to form a dynamically updated knowledge network, thereby ensuring the timeliness and accuracy of the knowledge graph.

[0138] In addition, as an example, the AR experimental model of the embodiment of the present invention is a spacecraft experimental model constructed based on the automatic control principle course;

[0139] In terms of model structure, the spacecraft experimental model of the embodiment of the present invention can be designed and developed through Blender software, including a cylindrical truss load-bearing structure, replaceable loads, solar wing installation and antenna deployment, etc. Specifically, a cylindrical truss load-bearing structure is adopted to support the replacement of loads, the solar wing is installed on the fixed support plate on the side of the satellite platform, the antenna is deployed on the outer surface of the payload cabin, and two propellant tanks are placed between the diagonal bars inside the truss to realize an efficient propulsion system. Figure 3 , which is a display effect diagram of an AR experimental model according to an embodiment of the present invention.

[0140] In terms of model functions, a teaching experiment on attitude fixed-time control is designed, mathematical modeling is carried out based on the kinematics and dynamics of the real spacecraft, and the design of the control law is completed, covering attitude control tasks under different working conditions, such as attitude adjustment and stability maintenance, etc., and a control gain input window is provided. The dynamic process of spacecraft attitude change under the influence of different control parameters is displayed through AR teaching instruments to understand the influence of control parameters on system stability, response speed and convergence accuracy.

[0141] The specific design process of the spacecraft experimental model includes:

[0142] (1) Assume that the attitude quaternion of spacecraft i relative to the inertial system is q 0，i ∈R,q i =[q 1，i ,q 2，i ,q 3，i ] T ∈R 3 , R is a real number;

[0143] The reference attitude quaternion is

[0144] The preset difference between the target posture and the reference posture is d 0，i ∈R,d i =[d 1，i , d 2，i , d 3，i ] T ∈R 3 ;

[0145] The target posture is set as:

[0146]

[0147] in,

[0148]

[0149] (2) Definition Q i The first-order derivative of the control torque on spacecraft i is defined as T i ∈R 3 , the disturbance torque is defined as δ Ti ∈R 3 , J i ∈R 3×3 Represents the moment of inertia matrix of spacecraft i. Define the following auxiliary variable matrix:

[0150]

[0151] Then the attitude kinematics of spacecraft i is and attitude dynamics for:

[0152]

[0153]

[0154] in,

[0155]

[0156] (3) Define variables And define the first backstepping control state variable X 1，i and the second backstepping control state variable X 2，i satisfy:

[0157] X 1，i =ε i -Q * ;X 1，i ∈R 4 ;

[0158] X 2，i ∈R 4 ;

[0159]

[0160] Among them, f(X 1，i )∈R 4 , is a virtual control function, 5<k1<1 and k2>1 are the ratios of two odd integers, γ1>0, γ2>0;

[0161] (4) Define the following parameters:

[0162]

[0163] v1=0,v2=k2-k1,

[0164]

[0165]

[0166]

[0167]

[0168]

[0169]

[0170]

[0171] Given the following assumptions:

[0172] Assumption 1: Considering that fuel consumption may cause the spacecraft moment of inertia matrix J i The spacecraft model should satisfy the following constraints:

[0173] ||h i (Q i , v i )||≤k h ||v i || 2 , where k J , k g , and k h is a positive constant, and I3 is represented by a 3*3 identity matrix;

[0174] Assumption 2: Spacecraft i is subject to an unknown small disturbance torque δ Ti The following inequality should be satisfied:

[0175]

[0176] in,

[0177] (5) Assume that the fixed-time attitude control law of spacecraft i is:

[0178] T i =sat(u i );

[0179] in:

[0180]

[0181] Δu i =T i -u i ;

[0182]

[0183] Among them, u a，1i ,u a，2i and sat(u i )for:

[0184]

[0185] and κ>1,m1>0,m2>0,k3>0,constant u imax ∈R ＞0 represents the maximum control torque of the single axis output of spacecraft i, and the adaptive compensation law x a，i satisfy:

[0186]

[0187] Among them, h1, h2, h3 and k3 satisfy the inequality: h3>1;

[0188] (6) Define the Lyapunov function as follows:

[0189] Define the first backstepping control state variable X 1，i The Lyapunov function V 1，i As follows:

[0190]

[0191] The power integration technique is used to define the second backstepping control state variable X 2，i The Lyapunov function V 2，i As follows:

[0192]

[0193] in:

[0194]

[0195] Definition and adaptive compensation law x a，i The Lyapunov function V 3，i As follows:

[0196]

[0197] Define the Lyapunov function V a，i as follows:

[0198] V a，i =V 1，i +V 2，i +V 3，i.

[0199] (7) The inequality is derived:

[0200]

[0201] in,

[0202]

[0203] in and satisfy:

[0204]

[0205] satisfy:

[0206]

[0207]

[0208] The upper bound of ΔM is:

[0209]

[0210] Among them, λ min (J i ) is the matrix J i The minimum eigenvalue of Converges to the following residual set within a fixed time T1 that satisfies the constraints:

[0211]

[0212] Fixed time T1 satisfies:

[0213] That is, there exists a fixed time T1≥0 so that X 1，i and X 2，i Converges to the neighborhood of the origin within a fixed time T1. And the fixed time T1 is bounded by the upper limit of ΔM. It can be concluded that the spacecraft attitude can actually and stably converge to the reference attitude quaternion Q within the fixed time T1. * Preset attitude difference D * .

[0214] Based on the aforementioned design process of the spacecraft experimental model, the parameters for controlling the attitude of the spacecraft experimental model in the embodiment of the present invention include: k1, k2, k3, κ, γ1, γ2, m1, m2, h1, h2, h3.

[0215] Specifically, in an embodiment of the present invention, an AR teaching instrument runs a pre-designed AR experimental model and superimposes and fuses a real scene image for display, including the following steps:

[0216] Step S00 1: Acquire a real scene image and design a positioning mark, extract feature descriptors from the real scene image containing the positioning mark, and calculate the position information and posture information of the positioning mark in the three-dimensional scene.

[0217] This step captures the visible light information of the three-dimensional scene through the camera, designs the positioning mark, extracts the feature descriptor from the visible light image of the real scene containing the positioning mark, and calculates the position information and six-dimensional posture information of the positioning mark in the three-dimensional scene.

[0218] Step S002: Using the position information and posture information of the positioning mark as a reference, the AR experimental model is drawn in the three-dimensional scene through the transformation matrix and jitter smoothing is performed.

[0219] Step S003: Overlay and fuse the AR experimental model with the real scene image for display.

[0220] Using the position and posture information of the positioning marker as a reference, virtual objects are drawn in the scene through the transformation matrix, and jitter smoothing is performed. The virtual objects are displayed in the user's field of view in an overlaid manner, and the display device's resolution, refresh rate, field of view and other parameters are optimized to provide a high-quality visual experience.

[0221] In step S001, the position information and posture information of the positioning mark in the three-dimensional scene are expressed as:

[0222] Position information: R = [r1 r2 r3] T ∈R 3 ;

[0223] Posture information: q0∈R,q i =[q 1，i ,q 2，i ,q 3，i ] T ∈R 3 .

[0224] In step S002, the position information and posture information of the positioning mark are used as references, and the AR experimental model is drawn in the three-dimensional scene through the transformation matrix and jitter smoothing is performed, including:

[0225] The jitter sampling time is t k , k∈N ≥0 , the position information and posture information at time k are represented as R(k) and Q(k) respectively, and the definition is

[0226]

[0227]

[0228] [ω 1，i ,ω 2，i ,ω 3，i ] is the Euler angle representation, representing it as a three-dimensional vector.

[0229] After smoothing, time t, t∈[t k , t k+1 The position information and posture information of the positioning mark of ] in the three-dimensional scene are expressed as:

[0230]

[0231] The position information R(t) and posture information Q(t) of the positioning marker in the three-dimensional scene are used as references, and the relative displacement matrix R t (t) = [r 1，t r 2，t r 3，t ] T and the relative attitude rotation matrix Q t (t) = [q 0，t ,q 1，t ,q 2，t ,q 3，t ] T Transformation, to obtain the position information R of the AR teaching instrument in the three-dimensional scene k (t) = [r 1，k r 2，k r 3，k ] T and attitude information Q k (t) = [q 0，k ,q 1，k ,q 2，k ,q 3，k ] T :

[0232] R k (t) = R(t) + R t (t);

[0233]

[0234] In the three-dimensional scene, the position information is R k (t) Q k (t) Plotting the spacecraft experimental model for attitude information.

[0235] Through the above algorithm, the positioning data and complex scene information are deeply analyzed.

[0236] The spacecraft fixed-time attitude control AR teaching instrument has been realized to show the attitude stabilization process of the spacecraft under the classic control strategy in real time, and the AR experimental model can continue to maintain its correct spatial positioning relationship in a dynamic and changeable real environment, seamlessly integrate with the real scene image and omnidirectionally accurately align, significantly improve the positioning accuracy, and effectively eliminate the image jitter caused by environmental changes or device movement. Ultimately, users will deeply experience the high presence and perfect integration of virtual objects in the real environment from a sensory level, thereby enhancing the authenticity and immersion of the learning experience.

[0237] In addition, an embodiment of the present invention also includes a virtual teaching method based on a dynamic knowledge graph, which is applicable to the virtual teaching system based on a dynamic knowledge graph of the aforementioned embodiment to achieve teaching objectives.

[0238] The camera in the embodiment of the present invention is connected to the AR teaching instrument for communication. When the AR teaching instrument is input with certain control instructions or parameters, the camera can be automatically activated to capture the real scene image. The camera in this embodiment can be a separately configured camera, combined with other AR devices to achieve the design purpose of the present invention, or it can be a camera installed on a laptop computer. At this time, the laptop computer serves as the main control component of the AR teaching instrument to process data and control the corresponding display device for real-time display. When different model parameters are input, the controlled change process of the model posture is vividly displayed, which improves the intuitiveness and interactivity of teaching, so that the teaching of professional courses can transcend the boundaries of traditional numerical simulation, directly touch the practical level of engineering application, and effectively promote the transformation of theoretical knowledge into practical skills.

[0239] As a specific example, the satellite's moment of inertia is:

[0240]

[0241] The satellite's initial attitude quaternion is randomly selected and expressed in Euler angles as [13° 11° 11°] T , the target attitude quaternion is randomly selected and expressed in Euler angles as [10° -5° 10°] T . Using the above-mentioned design of the present invention, the parameters in the control torque are selected as k1=9 / 11, k2=11 / 9, k3=0.5, K=1.01, γ1=0.05, γ2=0.025, m1=5.1, m2=4.6, h1=1, h2=1, h3=5. After entering the control parameters, the AR teaching instrument will dynamically display the whole process of the satellite maneuvering from the initial attitude to the target attitude.

[0242] The embodiments of the present invention focus on the systematic construction of professional courses and the visual presentation of knowledge structure, effectively mapping highly abstract theoretical concepts into specific teaching practices, integrating knowledge graphs with AR technology, innovating the teaching mode of automatic control courses, realizing the systematic and visual presentation of knowledge, enhancing the realism and interactivity of teaching practices, and thus constructing an efficient and interactive autonomous exploration learning ecosystem.

[0243] The present invention is further described above with the aid of specific embodiments. However, it should be understood that the specific description here should not be construed as limiting the essence and scope of the present invention. Various modifications made to the above embodiments by ordinary technicians in the field after reading this specification are all within the scope of protection of the present invention.

Claims

1. A virtual teaching system based on dynamic knowledge graph, characterized in that: The system includes a knowledge graph database, a camera and an AR teaching instrument; wherein: The knowledge graph database is constructed based on the knowledge resources involved in the professional courses, and contains knowledge structures of different levels and logical relationships between nodes; The camera is used to capture images of real scenes; The AR teaching instrument runs a pre-designed AR experimental model and superimposes and fuses the real scene image for display, and controls the posture of the AR experimental model according to the input control parameters; the AR experimental model is an instance constructed based on the knowledge graph database.

2. A virtual teaching system based on a dynamic knowledge graph as claimed in claim 1, characterized in that: The process of constructing the knowledge graph database based on the knowledge resources involved in professional courses includes: Taking the content of professional courses as the core, construct the overall architecture of the knowledge graph; the overall architecture at least includes the storage structure, data model and query language of the knowledge graph; Use natural language processing technology to pre-process the text contained in knowledge resources and extract key concepts, entities and attributes; The knowledge graph is represented by a graph data structure. Nodes are designed to represent key concepts and entities. The size and color of the nodes are used to reflect the knowledge structure at different levels. Edges are used to express the logical relationship between nodes to form the framework of the knowledge graph.

3. A virtual teaching system based on dynamic knowledge graph as claimed in claim 2, characterized in that: Also includes: The knowledge graph database is updated by regularly scanning the knowledge resources involved in professional courses.

4. A virtual teaching system based on dynamic knowledge graph as claimed in claim 1, characterized in that: The AR experimental model is a spacecraft experimental model constructed based on the automatic control principle course; setting the spacecraft experimental model includes: Set the attitude quaternion of spacecraft i relative to the inertial system to q 0,i ∈R,q i =[q 1,i ,q 2,i ,q 3,i ] T ∈R 3 , R is a real number; The reference attitude quaternion is The preset difference between the target posture and the reference posture is d 0,i ∈R,d i =[d 1,i ,d 2,i ,d 3,i ] T ∈R 3 ; The target posture is set as: in, definition Q i The first-order derivative of the control torque on spacecraft i is defined as T i ∈R 3 , the disturbance torque is defined as δ Ti ∈R 3 , J i ∈R 3×3 Represents the moment of inertia matrix of spacecraft i; define the following auxiliary variable matrix: Then the attitude kinematics of spacecraft i is and attitude dynamics for: in, Defining variables And define the first backstepping control state variable X 1,i and the second backstepping control state variable X 2,i satisfy: X 1,i =ε i -Q * ;X 1,i ∈R 4 ; X 2,i ∈R 4 ; where, f(X 1,i ) ∈ R 4 , is a virtual control function, 5 < k1 < 1 and k2 > 1 are the ratios of two odd integers, γ1 > 0, γ2 > 0; Define the following parameters: <h2 style=";text-align:left;direction:ltr">v1=0,v2=k2-k1,<h2 style=";text-align:left;direction:ltr"> Given the following assumptions: Assumption 1: Considering that fuel consumption may cause the spacecraft moment of inertia matrix J i The spacecraft model should satisfy the following constraints: ||h i (Q i ,v i )||≤k h ||v i || 2 , where k J , k g , and k h is a positive constant, and I3 is represented by a 3*3 identity matrix; Assumption 2: Spacecraft i is subject to an unknown small disturbance torque δ Ti The following inequality should be satisfied: in, The fixed-time attitude control law of spacecraft i is set as: T i =sat(u i ); in: Δu i =T i -u i ; Among them, u a,1i ,u a,2i and st(u i )for: and κ>1,m1>0,m2>0,k3>0, constant u imax ∈R >0 represents the maximum control torque of the single axis output of spacecraft i, and the adaptive compensation law x a,i The definition is as follows: Among them, h1, h2, h3 and k3 satisfy the inequality: h3>1; The Lyapunov function is defined as follows: Define the first backstepping control state variable X 1,i The Lyapunov function V 1,i As follows: The power integration technique is used to define the second backstepping control state variable X 2,i The Lyapunov function V 2,i As follows: in: Definition and adaptive compensation law x a,i The Lyapunov function V 3,i As follows: Define the Lyapunov function V a,i as follows: V a,i =V 1,i +V 2,i +V 3,i ; This leads to the inequality: in, in and satisfy: satisfy: The upper bound of ΔM is: Among them, λ min (J i ) is the matrix J i The minimum eigenvalue of Converges to the following residual set within a fixed time T1 that satisfies the constraints: Fixed time T1 satisfies:

5. A virtual teaching system based on dynamic knowledge graph as claimed in claim 4, characterized in that: The parameters for controlling the attitude of the spacecraft experimental model include: k1, k2, k3, κ, γ1, γ2, m1, m2, h1, h2, h3.

6. A virtual teaching system based on dynamic knowledge graph as claimed in claim 4, characterized in that: The AR teaching instrument runs a pre-designed AR experimental model and superimposes and fuses the real scene image for display, including: Acquire real scene images and design positioning marks, extract feature descriptors from real scene images containing positioning marks, and calculate the position information and posture information of the positioning marks in the three-dimensional scene; Using the position information and posture information of the positioning marker as a reference, the AR experimental model is drawn in the three-dimensional scene through a transformation matrix and jitter smoothing is performed; The AR experimental model is superimposed and fused with the real scene image for display.

7. A virtual teaching system based on dynamic knowledge graph as claimed in claim 6, characterized in that: The position information and posture information of the positioning marker in the three-dimensional scene are expressed as: Position information: R = [r1r2r3] T ∈R 3 ; Posture information: q0∈R,q i =[q 1,i ,q 2,i ,q 3,i ] T ∈R 3 .

8. A virtual teaching system based on dynamic knowledge graph as claimed in claim 7, characterized in that: Using the position information and posture information of the positioning marker as a reference, the AR experimental model is drawn in the three-dimensional scene through a transformation matrix and jitter smoothing is performed, including: The jitter sampling time is t k , k∈N ≥0 , the position information and posture information at time k are represented as R(k) and Q(k) respectively, and the definition is After smoothing, time t, t∈[t k ,t k+1 The position information and posture information of the positioning mark of ] in the three-dimensional scene are expressed as: The position information R(t) and posture information Q(t) of the positioning marker in the three-dimensional scene are used as references, and the relative displacement matrix R t (t) = [r 1,t r 2,t r 3,t ] T and the relative attitude rotation matrix Q t (t) = [q 0,t ,q 1,t ,q 2,t ,q 3,t ] T Transformation, to obtain the position information R of the AR teaching instrument in the three-dimensional scene k (t) = [r 1,k r 2,k r 3,k ] T and attitude information Q k (t) = [q 0,k ,q 1,k ,q 2,k ,q 3,k ] T : R k (t)=R(t)+R t (t); In the three-dimensional scene, the position information is R k (t) Q k (t) Plotting the spacecraft experimental model for attitude information.

9. A virtual teaching system based on dynamic knowledge graph as claimed in claim 8, characterized in that: The spacecraft experimental model is drawn using Blender software and adopts a cylindrical truss load-bearing structure to support the replacement of loads. The solar panels are installed on a fixed support plate on the side of the satellite platform, the antenna is deployed on the outer surface of the payload bay, and the two propellant tanks are placed between the diagonal bars inside the truss.

10. A virtual teaching method based on dynamic knowledge graph, characterized in that: The method is implemented based on the virtual teaching system based on the dynamic knowledge graph as described in any one of claims 1 to 9.

Images