Intelligent classroom auxiliary teaching management system based on real-time feedback of intelligent handwriting data

Abstract

The application discloses a smart classroom auxiliary teaching management system based on intelligent handwriting data real-time feedback and belongs to the field of big data processing, and comprises the following steps: a signal acquisition and segmentation module is used for receiving sensor data of an intelligent handwriting device, a dynamic vibration energy threshold is set according to environmental background noise to capture pen-down and pen-up events, and a writing time interval is defined; an abnormal behavior recognition and cleaning module is used for extracting time sequence motion features of a writing state, classifying and marking the writing state, and removing sensor data marked as abnormal behavior; a trajectory reconstruction and drift elimination module is used for accumulating and transforming the removed sensor data to obtain handwriting trajectories, and correcting speed vectors to suppress cumulative errors; and a smart classroom management service module is used for obtaining writing feature statistics according to the output handwriting trajectories, generating real-time teaching prompt information, and pushing the real-time teaching prompt information to a teacher terminal, so that the smart classroom auxiliary teaching management effectiveness is improved.

Description

Technical Field

[0001] This invention relates to the field of big data processing technology, and in particular to a smart classroom-assisted teaching management system based on real-time feedback of intelligent handwritten data. Background Technology

[0002] Intelligent handwriting technology, as a natural and intuitive human-computer interaction method, has been widely used in the field of education in recent years. Intelligent handwriting terminals (such as smart pens and percentage boards) can accurately capture multi-dimensional information about students during the writing process. Specifically, intelligent handwriting data refers to process data that reflects the characteristics of students' writing behavior, collected using terminal devices such as smart pens. Intelligent handwriting data focuses on the process characteristics of writing, mainly including parameters such as writing speed, handwriting continuity, and behavioral sequences over time. It records students' behavior during answering questions, practicing, or discussing, reflecting changes in the fluency of thought, level of concentration, and cognitive load.

[0003] Smart classroom-assisted teaching management refers to a teaching model that uses modern information technology to collect, analyze, and provide feedback on various types of data generated in classroom teaching activities in real time, in order to assist teachers in optimizing teaching strategies, adjusting teaching pace, and implementing refined class management.

[0004] In scenarios such as synchronous classroom answering, group discussion recording, and instant practice feedback, the system uses smart handwriting terminals to acquire students' behavioral parameters in real time. Combining time-based metrics and group comparison algorithms, the system generates real-time curves for classroom participation and focus, and calculates evaluation indicators such as classroom response delay, average writing speed change rate, and group synchronization. Through dynamic visualization of data from the whole class or small groups, teachers can observe in real time nodes of overall writing lag, frequent pauses, or large-scale revisions. Simultaneously, the system supports automatic identification and alerts for individual anomalies (such as prolonged periods without writing or frequent erasures). This real-time feedback mechanism based on objective data provides teachers with precise decision-making support for adjusting the pace of instruction and focusing on individual students.

[0005] For example, Chinese invention patent CN118071553B discloses a method and system for intelligent recognition of corrected handwriting, which includes: handwriting information collected based on dot matrix codes and smart pens, using a long short-term memory network to perform sequential data processing of writing information, including time series analysis of pen tip speed, pressure and pause time, and using preprocessing techniques to perform normalization and noise reduction to generate a dynamic writing feature dataset.

[0006] For example, the Chinese invention patent application CN119273504A discloses an interactive teaching method and system based on smart paper and pen, which includes: configuring student-end devices: distributing a smart pen and a specially designed code book to each student; configuring teacher-end devices: the teacher initiates the paper-pen interactive session using a large-screen teaching tool; acquiring student-end writing information: the smart pen captures writing actions in real time through its built-in sensors and converts the handwriting into digital signals through the code book; data collection and transmission; data processing and storage; data display on the large-screen teacher-end; teacher-end grading and feedback; teaching interaction and evaluation; optimization and upgrading.

[0007] The inventor has noted / that at least the following technical problems exist in the prior art:

[0008] In existing technologies, when collecting raw writing signals, including coordinates, timestamps, pen pressure, and pen lift-off events, it is difficult to effectively eliminate invalid handwriting behaviors (such as repeated tracing, unconscious doodling, and momentary accidental pen touches) that are not related to answering questions in a real classroom environment. Because invalid handwriting and valid answer handwriting are highly mixed at the physical signal level, the system suffers serious distortion when extracting behavioral features (e.g., misjudging repeated tracing as high fluency and false touches as valid pen strokes). It cannot accurately distinguish between the intention of "real answering questions" and "random doodling" based solely on physical features, and cannot objectively reflect the student's true writing state.

[0009] Furthermore, existing technologies are mostly based on statistical analysis using a single phase plane or a fixed time window for dimensionality reduction processing. This can easily lead to the loss of motion direction and spatial structure information, making it difficult to distinguish heterogeneous motions with similar dynamic characteristics. Moreover, they are not suitable for non-stationary time-varying writing processes and multi-granularity analysis requirements, resulting in low effectiveness of smart classroom-assisted teaching management due to errors in handwritten data feedback. Summary of the Invention

[0010] To address the technical problem of low effectiveness of smart classroom-assisted teaching management due to errors in handwritten data feedback in existing technologies, this invention provides a smart classroom-assisted teaching management system based on real-time feedback of intelligent handwritten data. The system includes:

[0011] The signal acquisition and segmentation module receives sensor data from the smart handwriting device, including at least acceleration and gyroscope data. It dynamically sets a dynamic vibration energy threshold based on ambient background noise to capture pen-dropping and pen-lifting events and defines the writing time interval. The abnormal behavior recognition and cleaning module extracts the temporal motion features of the writing state based on the acceleration data of the smart handwriting device, classifies and labels the writing state based on these features, and removes sensor data marked as abnormal behavior. The trajectory reconstruction and drift elimination module performs cumulative transformation on the removed sensor data to obtain the handwriting trajectory and corrects the velocity vector to suppress cumulative errors when the pen tip is detected to be stationary. The smart classroom management service module obtains writing feature statistics based on the handwriting trajectory output by the trajectory reconstruction and drift elimination module, and, combined with the abnormal behavior recognition results output by the abnormal behavior recognition and cleaning module, generates real-time teaching prompts and / or control commands, which are then pushed to the teacher's terminal.

[0012] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0013] 1. The intelligent classroom auxiliary teaching management system based on real-time feedback of intelligent handwriting data provided by this invention, by dynamically setting a dynamic vibration energy threshold, can automatically adjust the judgment criteria according to the real-time changing environmental background noise, greatly improving the accuracy of capturing pen-dropping and pen-lifting events. Compared with the fixed dynamic vibration energy threshold in the prior art, which is prone to misjudgment or omission of pen-dropping and pen-lifting events, resulting in inaccurate and incomplete collected writing data, this invention extracts the temporal motion features of the writing state from the acceleration data of the intelligent handwriting device, classifies and marks the writing state, and then removes sensor data marked as abnormal behavior. This effectively avoids the interference of abnormal data on subsequent trajectory reconstruction, writing feature statistics, and learning analysis, making the system more efficient and accurate. The generated teaching prompts are more accurate and reliable. Then, the sensor data after removing abnormal data is cumulatively transformed to obtain the handwriting trajectory. When the pen tip is detected to be stationary, the velocity vector is corrected to suppress the cumulative error. This effectively suppresses the generation and expansion of the cumulative error, so that the reconstructed handwriting trajectory can more accurately reflect the student's actual writing situation. Finally, by comprehensively analyzing the writing feature statistics and abnormal behavior recognition results, it is possible to comprehensively and timely grasp the student's writing situation, classroom concentration, and whether there is any abnormal behavior in the classroom. This generates real-time teaching prompts, which can provide teachers with precise suggestions for classroom discipline management and learning analysis results, effectively improving the quality of classroom teaching and realizing the improvement of the effectiveness of smart classroom-assisted teaching management.

[0014] 2. This invention extracts high-frequency vibration carrier signals by performing high-frequency bandpass filtering on acceleration data and calculating high-frequency vibration energy. Simultaneously, low-pass filtering is performed to obtain the acceleration component of the preset writing sensing axis. This allows for the analysis of writing action characteristics from different angles. Then, the fluctuation amplitude of the acceleration component of the preset writing sensing axis is judged to accurately determine the pen's stationary state. By acquiring the current environmental noise situation and dynamically setting a threshold, the accuracy of pen placement judgment is improved. Next, the timing of pen placement and lifting is determined by combining high-frequency vibration energy and the cumulative duration of candidate states, effectively eliminating the influence of environmental noise and transient interference, thereby avoiding misjudgment as pen placement or lifting actions and further ensuring the reliability of writing action judgment. Finally, the time interval defined by the pen placement and lifting moments is recorded as the writing time interval, which can accurately record the start and end times of each student's writing behavior, helping teachers optimize teaching strategies and improve classroom teaching efficiency.

[0015] 3. By calculating the fluctuation amplitude and phase trajectory flatness of the acceleration component along the preset writing sensing axis, if the fluctuation amplitude of the acceleration component is less than the preset writing stability threshold and the phase trajectory flatness is less than the preset phase trajectory flatness threshold, it is determined to be a valid writing state. Otherwise, the current writing state is marked as a candidate abnormal behavior state. The writing state is comprehensively evaluated from two different dimensions of conditions, which improves the accuracy of judging the valid writing state and avoids misjudgment that may occur due to a single condition. Finally, the cumulative time of the candidate abnormal state and the writing state before and after the candidate abnormal state are judged. The short-term fluctuations that may occur during normal writing are fully considered, avoiding the rejection of normal data due to misjudgment, ensuring the continuity and integrity of writing data, improving the accuracy of judging abnormal states, and timely rejection of abnormal states that last for a long time can prevent these data from interfering with subsequent teaching analysis, thus improving the reliability of smart classroom-assisted teaching management. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A schematic diagram of the structure of the smart classroom auxiliary teaching management system based on real-time feedback of intelligent handwriting data provided in an embodiment of this application;

[0018] Figure 2 A flowchart for determining the moment of pen placement and pen lifting provided in an embodiment of this application;

[0019] Figure 3A flowchart for determining the writing state provided in this application embodiment;

[0020] Figure 4 A flowchart for identifying invalid outline behavior states provided in an embodiment of this application. Detailed Implementation

[0021] The technical solution provided in this application will now be described with reference to the accompanying drawings.

[0022] To facilitate understanding of the embodiments of this application, the following points will be explained first:

[0023] First, in this application, "at least one" means one or more, and "more than one" means two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A alone, A and B simultaneously, or B alone, where A and B can be singular or plural. The character " / " generally indicates an "or" relationship between the preceding and following related objects, but it does not exclude the possibility of indicating an "and" relationship; the specific meaning can be understood in context. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can mean: a, b, c; a and b; a and c; b and c; or a and b and c. Here, a, b, and c can be single or multiple.

[0024] Second, the use of prefixes such as "first" and "second" in this application is solely for the purpose of distinguishing and describing different things belonging to the same category, and does not constrain the order, size, or quantity of things. For example, "first message" and "second message" are simply different messages, and there is no chronological, size, or priority relationship between them.

[0025] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0026] In application scenarios such as synchronous classroom answering, group discussion recording, and instant practice feedback, intelligent handwriting terminals (such as percentage boards and smart pens) can acquire relevant student data in real time and process the student data to achieve smart classroom-assisted teaching management.

[0027] like Figure 1 The diagram shown is a structural schematic of a smart classroom-assisted teaching management system based on real-time feedback of intelligent handwriting data provided in an embodiment of this application. The method includes the following steps:

[0028] The signal acquisition and segmentation module receives sensor data from the smart handwriting device. It dynamically sets a dynamic vibration energy threshold based on ambient background noise to capture pen-drop and pen-lift events and defines the writing time interval. Sensor data includes at least acceleration and gyroscope data, reflecting the motion state of the handwriting device during writing. Ambient background noise represents sensor signal fluctuations caused by non-writing actions, such as hand tremors, device shaking, or table vibration. The smart handwriting device refers to a handwriting tool capable of collecting writing-related data (such as acceleration and gyroscope data), such as a smart pen. The writing time interval represents the time period from the pen-drop event to the pen-lift event. Dynamically setting the dynamic vibration energy threshold adapts to different ambient background noise conditions, improving the accuracy of pen-drop and pen-lift event capture. Accurately defining the writing time interval provides a clear time range for subsequent processing, ensuring that data processing targets the effective writing process.

[0029] like Figure 2 The flowchart shown illustrates the determination of the pen stroke and pen lift-off moments. Further, it dynamically sets a dynamic vibration energy threshold based on ambient background noise to capture pen stroke and pen lift-off events, specifically including:

[0030] Accelerometer data on the preset writing sensing axis is collected at a set sampling frequency. The sampling frequency is set according to the actual writing speed of students, which determines the precision of data collection. The preset writing sensing axis, i.e. the direction of the pen body axis, is a specific direction that is preset to sense writing actions, because the acceleration change in the direction of the pen body axis during writing can better reflect the pen-putting and pen-lifting actions.

[0031] Acceleration data undergoes high-frequency bandpass filtering to retain only specific high-frequency ranges in the signal, such as 50-500Hz, while removing interference signals from other frequency bands, i.e., high-frequency vibration carrier signals. The energy amplitude of the high-frequency vibration carrier signals is recorded as high-frequency vibration energy. The high-frequency vibration carrier signal represents the high-frequency vibration signal generated by the friction between the pen and the writing medium such as paper during writing. High-frequency vibration energy represents the energy possessed by the high-frequency vibration carrier signal and is used to measure the intensity of high-frequency vibration during writing. The high-frequency friction carrier energy acquisition technology directly converts microscopic vibration energy through the piezoelectric effect.

[0032] The acceleration data is low-pass filtered to retain low-frequency components in the signal, such as those less than 10Hz, while removing interference signals from other frequency bands. This yields the acceleration component along the preset writing sensing axis, reflecting the overall posture changes of the pen, such as lifting, lowering, and hovering, which is used to determine whether the pen is in a stationary state.

[0033] If the static state determination conditions are met within a preset time period, the smart pen is determined to be in a static state and enters calibration mode; otherwise, it continues to monitor the fluctuation amplitude of the acceleration component along the preset writing sensing axis. The static state determination conditions specifically include: the fluctuation amplitude of the acceleration component along the preset writing sensing axis is less than a preset static determination threshold; the smart pen's motion synthesis velocity is less than a preset static velocity threshold; and the angular velocity amplitude of the gyroscope data is less than a preset static angular velocity threshold. The motion synthesis velocity represents the result of cumulative transformation of the acceleration data, i.e., synthesizing the acceleration along the preset writing sensing axis, along the pen's length direction, and along the pen's width direction to obtain the motion acceleration magnitude, and integrating the motion acceleration to obtain the motion synthesis velocity. Calibration mode specifically involves: calculating the current ambient background noise energy and setting a dynamic vibration energy threshold for pen placement determination. The dynamic vibration energy threshold is the product of the background noise energy and the preset vibration coefficient; the fluctuation amplitude of the acceleration component along the preset writing sensing axis is the standard deviation of the acceleration component along the preset writing sensing axis; the preset stationary judgment threshold represents the preset limit of the acceleration component fluctuation amplitude; the preset stationary velocity threshold represents the preset limit of the motion synthesis velocity; the preset stationary angular velocity threshold represents the preset limit of the angular velocity amplitude of the gyroscope data; the current ambient background noise energy represents the energy of the noise generated by the environment when the smart pen is stationary, as reflected in the acceleration data; the preset vibration coefficient represents the coefficient set by the preset administrator, for example, it can be set to 3; thus accurately judging the stationary state of the smart pen, obtaining the current ambient noise situation through the calibration mode, and dynamically setting the dynamic vibration energy threshold can adapt to different ambient noise conditions and avoid false triggering or missed triggering.

[0034] If the high-frequency vibration energy is not less than the dynamic vibration energy threshold, it is marked as a candidate writing state, and the cumulative duration of the candidate writing state is counted, that is, the time required from the moment it is marked as a candidate writing state to the current moment; thus, the initial signal of the start of writing is identified, but time verification is required to eliminate instantaneous interference.

[0035] If the cumulative duration of a candidate writing state reaches the cumulative writing time threshold, the moment marked as a candidate writing state is marked as the writing moment; otherwise, it is not marked as the writing moment. The cumulative writing time threshold represents the pre-set length of time that can be determined as a writing moment. By judging the cumulative duration of candidate writing states, the reliability of writing event judgment is improved and false positives are reduced.

[0036] If the high-frequency vibration energy is less than the dynamic vibration energy threshold, it is marked as a candidate pen-lifting state, and the cumulative duration of the candidate pen-lifting state is counted, that is, the time required from the moment it is marked as a candidate pen-lifting state to the current moment; thus identifying the initial signal of writing pause or end.

[0037] If the cumulative duration of the candidate pen-lifting state reaches the cumulative pen-lifting time threshold, the moment marked as a candidate pen-lifting state will be marked as a pen-lifting moment; otherwise, it will not be marked as a pen-lifting moment. This avoids misjudging short pauses, such as thinking, as pen-lifting.

[0038] The time interval defined by the moment the pen is put down and the moment the pen is lifted is recorded as the writing time interval, providing structured data units for subsequent handwriting reconstruction, writing behavior analysis, and teaching feedback.

[0039] By extracting key signal features through various filtering processes and combining dynamic thresholds and cumulative time judgment, the influence of environmental noise and transient interference is effectively eliminated, improving the accuracy and reliability of pen-starting and pen-lifting judgments, thereby providing accurate writing time information for the smart classroom auxiliary teaching management system.

[0040] The abnormal behavior identification and cleaning module is used to extract temporal motion features of the writing state based on the acceleration data of the smart handwriting device. Based on these temporal motion features, the writing state is classified and labeled, and sensor data marked as abnormal behavior is removed. Temporal motion features represent the motion characteristics exhibited by the writing state over time, such as the variation of acceleration at different time points. Removal means removing data segments marked as abnormal from subsequent processing flows. Extracting temporal motion features accurately describes changes in the writing state, classification and labeling can identify abnormal writing behaviors, and removing abnormal data can prevent abnormal data from interfering with subsequent trajectory reconstruction and learning analysis, thus improving data quality.

[0041] Furthermore, the temporal motion features of the writing state are extracted based on the acceleration data of the smart handwriting device. The specific process is as follows:

[0042] Acceleration data within the writing time interval, i.e., the acceleration sequence, is cumulatively transformed to obtain a velocity sequence, which describes the change of the speed of the handwriting device over time within the writing time interval. The velocity sequence and acceleration sequence within the writing time interval are mapped to a mixed phase plane of velocity sequence-acceleration sequence. In this application, the cumulative transformation represents an integral operation; the mixed phase plane of velocity sequence-acceleration sequence represents the space formed by mapping the velocity sequence and acceleration sequence to the same plane, which is used to comprehensively analyze the relationship between velocity and acceleration to determine the writing state. The velocity sequence obtained by the cumulative transformation can more intuitively understand the speed change of the intelligent handwriting device, and mapping velocity and acceleration to the mixed phase plane of velocity sequence-acceleration sequence can comprehensively analyze the relationship between the two, providing a basis for subsequent judgment of the writing state.

[0043] The state determination and processing of signals within the mixed phase plane of velocity and acceleration sequences are performed by combining high-frequency vibration energy, specifically as follows:

[0044] If the detected high-frequency vibration energy is not less than the dynamic vibration energy threshold, it is determined to be a paper-pen contact state, and a first-order low-pass filter or moving average filter is used to smooth the velocity and acceleration sequences. Otherwise, it is determined to be a paper-pen non-contact state, and the velocity and acceleration sequences are attenuated to suppress noise drift. That is, each data in the velocity and acceleration sequences is multiplied by a preset attenuation coefficient, the value of which can be in the range of [0, 0.5], so that the phase trajectory converges to the origin quickly. By comparing the high-frequency vibration energy and the dynamic vibration energy threshold, the paper-pen contact state can be accurately determined. When the paper-pen is in contact, low-pass filtering and smoothing can remove high-frequency noise, making the data smoother and more accurately reflecting the writing motion. When the paper-pen is non-contact, attenuation can effectively suppress noise drift and improve the accuracy of acceleration data.

[0045] The confidence weight coefficient for each moment is calculated based on the high-frequency vibration energy and dynamic vibration energy threshold. Specifically, when the high-frequency vibration energy is less than the dynamic vibration energy threshold, the corresponding confidence weight coefficient is set to 0; when the high-frequency vibration energy is not less than the maximum high-frequency vibration energy in the historical time period, the corresponding confidence weight coefficient is set to 1; when the high-frequency vibration energy is between the dynamic vibration energy threshold and the maximum high-frequency vibration energy in the historical time period, the difference between the high-frequency vibration energy and the dynamic vibration energy threshold is calculated as a ratio to the difference between the maximum high-frequency vibration energy and the dynamic vibration energy threshold in the historical time period to obtain the corresponding confidence weight coefficient, which reflects the reliability of the acceleration data.

[0046] The velocity and acceleration sequences at each time point within the writing time interval are weighted and corrected using confidence weight coefficients at each time point to filter out noisy data points in the static state and improve the quality of acceleration data.

[0047] Based on the corrected velocity and acceleration sequences, phase trajectory flatness and phase trajectory compactness are calculated. Phase trajectory flatness is the ratio of the standard deviation of acceleration to the standard deviation of velocity. The mixed phase plane of the velocity-acceleration sequence is divided into several grid regions. The ratio of the number of data points falling into each grid region to the total number of data points falling into the grid region is denoted as the statistical probability. The statistical probability is substituted into the information entropy formula to calculate the information entropy, and the reciprocal of the information entropy is used as the phase trajectory compactness. Phase trajectory flatness and phase trajectory compactness are parameters that describe writing characteristics from different perspectives. Phase trajectory flatness can reflect the relative degree of acceleration and velocity changes during writing, while phase trajectory compactness can reflect the density of writing, providing an important basis for further in-depth analysis of writing states and characteristics.

[0048] like Figure 3The flowchart for determining the writing state is shown. Further, the writing state is classified and labeled based on its temporal motion features. The specific process is as follows:

[0049] A writing state is considered valid only when both the writing stability condition and the shape flatness condition are met simultaneously. The writing stability condition is that the fluctuation amplitude of the acceleration component along the preset writing sensing axis is less than the preset writing stability threshold, indicating that the writing along that axis is relatively stable. The shape flatness condition is that the phase trajectory flatness is less than the preset phase trajectory flatness threshold. The phase trajectory is a trajectory diagram drawn by plotting the acceleration signal in two orthogonal dimensions, such as x-axis acceleration and y-axis acceleration. If the trajectory is close to a straight line (i.e., flat), it indicates that the movement direction is unidirectional and there is no complex jitter, which meets the characteristics of valid writing. If the trajectory is circular or messy, it may be a non-writing action, such as shaking or hovering. The preset writing stability threshold and the preset phase trajectory flatness threshold are set by preset personnel based on the actual writing state of the students. The preset writing stability threshold is greater than the preset static judgment threshold. By judging the conditions in two different dimensions, the writing state is comprehensively evaluated from both the aspects of stability and shape, which improves the accuracy of judging the valid writing state and avoids misjudgment that may occur due to judging a single condition.

[0050] If any one or two of the above conditions are not met, the current writing state will be marked as a candidate abnormal behavior state. Further judgment based on subsequent conditions is needed to determine whether it is a true abnormal behavior state. This is to prepare for further identification and handling of abnormal states and to initially screen out writing states that may have problems.

[0051] Specifically, the process for removing sensor data marked as abnormal behavior is as follows:

[0052] If the cumulative time of a candidate abnormal state is less than the first abnormal cumulative time threshold, and the writing states before and after the candidate abnormal state are all valid writing states, then the sensor data within the current time interval is merged into the data stream of the adjacent valid writing states, and the candidate abnormal behavior state is removed from the data stream. The first and second abnormal cumulative time thresholds represent the boundary values ​​used to measure the duration of candidate abnormal states, set by preset personnel based on experience, with the second abnormal cumulative time threshold being greater than the first abnormal cumulative time threshold. "Time sequence" refers to the chronological order of writing states, used to describe the temporal relationship between various states during the writing process. Considering the possible brief fluctuations or minor abnormalities that may occur during normal writing, this process avoids misjudging and removing valid data due to these brief situations. Through the merging operation, the continuity and integrity of the writing data are ensured, enabling subsequent teaching analysis to more accurately reflect students' actual writing performance and learning status.

[0053] If the cumulative time of the candidate abnormal state is less than the first abnormal cumulative time threshold, and the writing states before and after the candidate abnormal state are not all valid writing states, then the current state is maintained and the writing state of the next time sequence is monitored. Since the states before and after the candidate abnormal state are not all valid writing states, it means that the nature of the current candidate abnormal state is not clear enough and a judgment cannot be made immediately. Maintaining the current state and continuing to monitor can provide more information for subsequent judgments, avoid making wrong decisions too early, and improve the accuracy of writing state judgment.

[0054] If the cumulative time of candidate abnormal states is not less than the second abnormal cumulative time threshold, the current writing state is marked as an abnormal behavior state, and the corresponding data is removed. Timely marking as abnormal behavior states and removing corresponding data can prevent abnormal data from interfering with subsequent teaching analysis, ensuring the purity and reliability of the data on which the analysis is based, thereby improving the accuracy of teaching analysis results.

[0055] If the cumulative time of a candidate abnormal state is between the first and second abnormal cumulative time thresholds, the current state is maintained and the writing state of the next time series is monitored. For candidate abnormal situations in the intermediate state, a certain observation time is given. By continuously monitoring the writing state of the next time series, the development trend of the state can be understood more comprehensively, so as to more accurately determine whether it is a real abnormal behavior state, and further improve the accuracy and rationality of the writing state judgment.

[0056] The trajectory reconstruction and drift elimination module is used to perform cumulative transformation on the rejected sensor data to obtain the handwriting trajectory, and to correct the velocity vector when the pen tip is detected to be stationary in order to suppress the cumulative error. The cumulative error represents the total error caused by position drift due to small deviations in the sensor during the trajectory reconstruction process. By performing cumulative transformation on the sensor data, the handwriting trajectory can be restored, and correcting the velocity vector can effectively suppress the cumulative error, making the reconstructed handwriting trajectory more accurate and closer to the actual writing trajectory.

[0057] Furthermore, the trajectory reconstruction and drift elimination module specifically includes:

[0058] The removed acceleration data is subjected to a secondary cumulative transformation, that is, a secondary integration operation to obtain displacement. Acceleration can be integrated to obtain velocity, and then integrated again to obtain displacement, thus obtaining the position change of the stylus during the writing process, realizing the conversion from acceleration data to displacement data.

[0059] During the cumulative transformation process, the high-frequency vibration energy and the gyroscope angular velocity magnitude are monitored in real time within the writing time interval. By monitoring these two parameters in real time, the vibration and rotation state of the stylus during the writing process can be understood in a timely manner, providing a basis for judging whether the pen tip is in a stationary state, which helps to accurately eliminate drift and improve the accuracy of trajectory reconstruction.

[0060] If the high-frequency vibration energy is less than the dynamic vibration energy threshold and the gyroscope angular velocity magnitude is less than the preset angular velocity threshold, the pen tip is determined to be stationary, and the horizontal velocity vector in the writing plane is set to zero, because theoretically, the horizontal velocity should be zero when the pen tip is stationary. The gyroscope is used to measure the rotational angular velocity of the stylus. Angular velocity is a vector with both value and direction. The angular velocity magnitude is the value of the angular velocity vector, reflecting the speed of the stylus's rotation. The preset angular velocity threshold is a pre-set reference value used to determine the gyroscope angular velocity magnitude. Accurately determining the stationary state of the pen tip and setting the horizontal velocity vector in the writing plane to zero can effectively eliminate drift caused by equipment errors, external interference, and other factors accumulated during the integration process, making the reconstructed trajectory more accurately reflect the actual writing situation and improving the quality and stability of trajectory reconstruction.

[0061] The smart classroom management service module uses handwriting feature statistics derived from the handwriting trajectory output by the trajectory reconstruction and drift elimination module, combined with the abnormal behavior recognition results output by the abnormal behavior identification and cleaning module, to generate real-time teaching prompts. These prompts and / or control commands are then pushed to the teacher's terminal. The handwriting feature statistics represent statistical data describing handwriting characteristics obtained after analyzing the handwriting trajectory. The real-time teaching prompts, generated based on the handwriting feature statistics and abnormal behavior recognition results, provide timely information to influence teachers' teaching behavior during class, such as reminding teachers to pay attention to a student's abnormal behavior. The handwriting feature statistics reflect students' handwriting performance, and combined with abnormal behavior recognition results, provide a comprehensive understanding of students' classroom performance. The generated real-time teaching prompts offer teachers timely guidance for classroom discipline management and student learning analysis, assisting them in better managing their teaching.

[0062] Furthermore, writing feature statistics are obtained based on the handwriting trajectory output by the trajectory reconstruction and drift elimination module. The specific acquisition method is as follows:

[0063] From the handwriting trajectory processed by the trajectory reconstruction and drift elimination module, the trajectory segments determined to be in a valid writing state are extracted. The total path length of all trajectory segments determined to be in a valid writing state is calculated and recorded as the valid writing distance. A valid writing state refers to the state actually used for writing content (such as text, graphics, etc.) during handwriting, distinguished from non-writing states such as lifting the pen or pausing. The effective writing distance can intuitively reflect the actual amount of content written by students within a certain period of time. By comparing the effective writing distances of different students or at different time periods, we can understand students' writing speed and completion of writing tasks, assisting teachers in assessing students' learning engagement.

[0064] The average speed of the trajectory segment judged as a valid writing state within a preset time statistical interval is calculated and recorded as the writing fluency; the writing fluency reflects the student's proficiency and consistency in writing.

[0065] Calculate the percentage of effective writing time within the preset time statistical interval and record it as classroom focus; classroom focus reflects the degree of student concentration in the classroom from the perspective of writing behavior.

[0066] The statistical measures of writing characteristics include effective writing area, writing fluency, and classroom concentration.

[0067] Specifically, the process for generating real-time teaching prompts is as follows:

[0068] Within a preset time interval, the effective writing distance from the start of writing to the current time window and the effective writing distance from the start of writing to the previous time window are calculated separately. Then, the effective writing distance from the start of writing to the current time window is subtracted from the effective writing distance from the start of writing to the previous time window, and then divided by the effective writing distance from the start of writing to the previous time window. The result is the effective writing distance growth rate. The effective writing distance growth rate can reflect the change in the amount of writing content of students in different time periods. A time window represents a smaller time period further divided within the preset time interval. For example, if the preset time interval is 10 minutes, it can be divided into 5 two-minute time windows to compare the changes in writing characteristics in different time periods.

[0069] Student behavior is monitored within a preset time period, and the number of abnormal behaviors is counted. If the number of abnormal behaviors exceeds a preset threshold, or the effective writing distance growth rate is less than the preset effective writing distance growth rate threshold, a attention prompt control instruction is generated for display on the teacher's end. The effective writing distance growth rate threshold represents a preset limit for judging the student's writing status.

[0070] If the average classroom attention level of all students is lower than a preset attention threshold, and the average writing fluency is lower than a preset fluency threshold, a control instruction is generated to prompt the teacher to increase the time spent explaining the questions. The preset attention threshold and preset fluency threshold are preset by the designated personnel based on the actual application scenario. By generating control instructions, teachers can adjust their teaching strategies in a timely manner according to the overall learning status of the class, thereby improving the effectiveness of smart classroom-assisted teaching management.

[0071] Example 2: Based on Example 1, further data processing is performed on the invalid outlining behavior state in the abnormal behavior state. The abnormal behavior identification and cleaning module also includes:

[0072] like Figure 4 The flowchart shown illustrates the invalid stroke behavior state recognition process. Within the writing time interval, acceleration data is cumulatively transformed to generate a velocity sequence. A first-order high-pass filter is then used to filter the velocity sequence to suppress DC component drift. The first-order high-pass filter allows high-frequency signals to pass through while suppressing low-frequency signals (the DC component is a low-frequency signal). In this application, it is used to filter the velocity sequence, suppressing DC component drift and making the velocity sequence more accurately reflect the actual movement speed changes of the smart pen. Converting acceleration data into velocity data provides a more intuitive reflection of the smart pen's movement speed changes. Suppressing DC component drift eliminates low-frequency interference in the velocity sequence, making the velocity data more accurate and reliable, providing a good foundation for subsequent analysis.

[0073] A speed threshold is set, and the speed amplitudes of data points whose filtered speed amplitudes are less than the threshold are zeroed out. The number of valid sampling points with speed amplitudes greater than zero is counted and recorded as the motion amplitude feature. The speed threshold represents a pre-defined speed value limit. By zeroing out meaningless speed fluctuations, speed data with obvious motion is highlighted. The motion amplitude feature obtained by counting the number of valid sampling points can simply and effectively measure the motion amplitude of the smart pen during writing.

[0074] Obtain the zero-crossing rate of the composite acceleration in two orthogonal directions within the writing plane; the zero-crossing rate refers to the number of times the signal value changes from positive to negative or from negative to positive within the signal writing time interval.

[0075] A relative displacement sequence is generated by cumulative transformation based on the velocity sequence. The standard deviations of displacement along the pen's length and width are calculated within the writing time interval. The relative displacement sequence is generated by cumulative transformation based on the velocity sequence. According to the physical relationship between velocity and displacement (displacement is the integral of velocity over time), velocity information is converted into displacement information, reflecting the positional change of the smart pen relative to the moment of pen stroke. Generating the relative displacement sequence reveals the positional changes of the smart pen; calculating the standard deviations of displacement in two directions measures the dispersion of displacement in these directions, reflecting the stability and regularity of the smart pen's movement.

[0076] If the motion amplitude characteristic is less than the preset sampling lower limit threshold, and the zero-crossing rate of the composite acceleration in the two orthogonal directions within the writing plane is greater than the preset zero-crossing upper limit threshold, and the standard deviation of displacement along the pen length direction and the standard deviation of displacement along the pen width direction are both less than the preset displacement aggregation threshold, then the current writing state is marked as an invalid stroke behavior state among the abnormal behavior states, and the corresponding data is removed. By comprehensively judging multiple feature conditions, invalid stroke behavior can be accurately identified. Removing abnormal data can avoid interference with subsequent teaching analysis and assistance, and improve the quality of handwritten data and the accuracy of analysis results.

[0077] Specifically, to further eliminate false positives and ensure the accuracy of invalid stroke behavior status determination, the current writing state is marked as the invalid stroke behavior state among abnormal behavior states, which also includes:

[0078] The effective writing distance is calculated by performing cumulative transformation on the writing time intervals where the speed amplitude is not less than the speed threshold. The speed amplitude refers to the numerical value of the speed. By setting the speed threshold to filter the data time periods, data that is too slow, belongs to slight jitter, or is an ineffective writing action is excluded. Only data with a certain movement speed and producing an effective writing effect are focused on. The cumulative transformation operation integrates the speed information of these effective data time periods into the effective writing distance, providing a quantitative basis for subsequent judgment on whether the writing behavior is effective.

[0079] If the effective writing distance is less than the preset effective displacement threshold, the current writing state is marked as an invalid stroke behavior state in the abnormal behavior state, and the corresponding data is removed. The preset effective displacement threshold represents a pre-set displacement value standard. By comparing with the preset effective displacement threshold, invalid stroke behavior can be accurately identified, avoiding interference from invalid data with subsequent teaching analysis and feedback. Marking abnormal behavior states can provide teachers with detailed information about students' writing behavior, helping them understand the problems students encounter during the writing process. Removing the corresponding data ensures that the data processed by the system is valid, improving the quality of handwritten data and the accuracy of subsequent analysis.

[0080] The various features and processes described above can be used independently of each other or can be combined in various ways. All possible combinations and sub-combinations are intended to fall within the scope of this disclosure. Furthermore, certain method or process blocks may be omitted in some embodiments. The methods and processes described herein are not limited to any particular order, and the blocks or states associated with them may be performed in other suitable orders. For example, the described blocks or states may be performed in an order different from the order specifically disclosed, or multiple blocks or states may be combined in a single block or state. Example blocks or states may be performed serially, in parallel, or in some other manner. Blocks or states may be added to or removed from the disclosed example embodiments. The exemplary systems and components described herein may be configured differently from those described. For example, elements may be added to, removed from, or rearranged compared to the disclosed example embodiments.

[0081] The various operations of the example methods described herein can be performed at least in part by an algorithm. This algorithm can be contained in program code or instructions stored in memory (e.g., the aforementioned non-transitory computer-readable storage medium). Such an algorithm may include a machine learning algorithm. In some embodiments, the machine learning algorithm may not be explicitly programmed into the computer to perform the function, but can learn from training data to create a predictive model that performs the function.

[0082] The various operations of the example methods described herein can be performed, at least in part, by one or more processors that are temporarily configured (e.g., by software) or permanently configured to perform the relevant operations. Whether temporarily or permanently configured, such processors can constitute the engine of a processor implementation that operates to perform one or more of the operations or functions described herein.

[0083] Similarly, the methods described herein can be implemented at least in part by a processor, where one or more specific processors are examples of hardware. For example, at least some operations of a method can be performed by one or more processors or an engine implemented by a processor. Furthermore, one or more processors can also be operated to support the performance of related operations in a “cloud computing” environment or as “Software as a Service” (SaaS). For example, at least some operations can be performed by a set of computers (as an example of a machine including processors), where these operations are accessible via a network (e.g., the Internet) and via one or more suitable interfaces (e.g., application programming interfaces (APIs)).

[0084] The performance of certain operations can be distributed across processors, residing not only within a single machine but also deployed across multiple machines. In some example embodiments, the processor or processor-implemented engine may reside in a single geographic location (e.g., within a home environment, office environment, or server cluster). In other example embodiments, the processor or processor-implemented engine may be distributed across multiple geographic locations.

[0085] In this specification, multiple instances may implement components, operations, or structures described as single instances. Although individual operations of one or more methods are shown and described as separate operations, one or more of the separate operations may be performed simultaneously and do not need to be performed in the order shown. Structures and functions presented as separate components in the example configuration may be implemented as composite structures or components. Similarly, structures and functions presented as single components may be implemented as separate components. These and other variations, modifications, additions, and improvements fall within the scope of this document.

[0086] While an overview of the subject matter has been described with reference to specific example embodiments, various modifications and changes can be made to these embodiments without departing from the broader scope of embodiments of this disclosure. Such embodiments of the subject matter are referred to herein, individually or collectively, by the term "invention," and are used for convenience only and are not intended to limit the scope of this application to any single disclosure or concept, should more than one disclosure or concept be disclosed in fact.

[0087] The embodiments described herein have been described in sufficient detail to enable those skilled in the art to practice the disclosed teachings. Other embodiments may be used and derived therefrom, such that structural and logical substitutions and changes may be made without departing from the scope of this disclosure. Therefore, the detailed description should not be construed as limiting, and the scope of the various embodiments is defined only by the appended claims and the full scope of their equivalents.

Claims

1. A smart classroom-assisted teaching management system based on real-time feedback of intelligent handwritten data, characterized in that: include: The signal acquisition and segmentation module is used to receive sensor data from the smart handwriting device. The sensor data includes at least acceleration data and gyroscope data. The module dynamically sets the dynamic vibration energy threshold based on the ambient background noise to capture pen-dropping and pen-lifting events and defines the writing time interval. The abnormal behavior identification and cleaning module is used to extract the temporal motion features of the writing state based on the acceleration data of the smart handwriting device, classify and label the writing state based on the temporal motion features of the writing state, and remove sensor data marked as abnormal behavior. The trajectory reconstruction and drift elimination module is used to perform cumulative transformation on the discarded sensor data to obtain the handwriting trajectory, and to correct the velocity vector to suppress cumulative error when the pen tip is detected to be stationary. The smart classroom management service module is used to obtain writing feature statistics based on the handwriting trajectory output by the trajectory reconstruction and drift elimination module, and combine the abnormal behavior recognition results output by the abnormal behavior recognition and cleaning module to generate real-time teaching prompts and push the displayed real-time teaching prompts and / or control commands to the teacher's terminal. The specific process for extracting the temporal motion features of the writing state based on the acceleration data of the smart handwriting device is as follows: Acceleration data within the writing time interval are cumulatively transformed to obtain a velocity sequence, and the velocity sequence and acceleration sequence within the writing time interval are mapped to a mixed phase plane of velocity sequence-acceleration sequence; The state determination and processing of signals within the mixed phase plane of velocity and acceleration sequences are performed by combining high-frequency vibration energy, specifically as follows: If the detected high-frequency vibration energy is not less than the dynamic vibration energy threshold, it is determined to be a paper-pen contact state, and the velocity sequence and acceleration sequence are subjected to low-pass filtering smoothing. Otherwise, it is determined to be a paper-pen non-contact state, and the velocity sequence and acceleration sequence are subjected to attenuation processing to suppress noise drift. The confidence weighting coefficient is calculated based on the high-frequency vibration energy and dynamic vibration energy threshold, and the confidence weighting coefficient is used to weight and correct the velocity sequence and acceleration sequence within the writing time interval to filter out noise data points in the static state. Based on the corrected velocity and acceleration sequences, the phase trajectory flatness and phase trajectory density are calculated. The phase trajectory flatness is the ratio of the acceleration standard deviation to the velocity standard deviation, and the phase trajectory density is obtained based on the data distribution information entropy of the mixed phase plane of the velocity-acceleration sequence.

2. The smart classroom auxiliary teaching management system based on real-time feedback of intelligent handwriting data as described in claim 1, characterized in that: The method of dynamically setting a dynamic vibration energy threshold based on environmental background noise to capture pen-dropping and pen-lifting events specifically includes: Accelerometer data at a preset writing sensing axis are collected at a set sampling frequency; The acceleration data is subjected to high-frequency bandpass filtering to extract the high-frequency vibration carrier signal, and the energy amplitude of the high-frequency vibration carrier signal is recorded as the high-frequency vibration energy. The acceleration data is low-pass filtered to extract the acceleration component along the preset writing sensing axis, which is used as a pressure change feature. If the static state determination condition is met within the preset time period, the smart pen is determined to be in a static state and enters the calibration mode; otherwise, the fluctuation amplitude of the acceleration component in the preset writing sensing axis continues to be monitored. The static state determination conditions specifically include: the fluctuation amplitude of the acceleration component along the preset writing sensing axis is less than the preset static determination threshold, the motion synthesis speed of the smart pen is less than the preset static speed threshold, and the angular velocity amplitude of the gyroscope data is less than the preset static angular velocity threshold. The motion synthesis speed represents the result obtained by cumulatively transforming the acceleration data. The calibration mode specifically involves: calculating the current ambient background noise energy and setting a dynamic vibration energy threshold for pen placement judgment, wherein the dynamic vibration energy threshold is the product of the background noise energy and a preset vibration coefficient. If the high-frequency vibration energy is not less than the dynamic vibration energy threshold, it is marked as a candidate pen-writing state, and the cumulative duration of the candidate pen-writing state is obtained. If the cumulative duration of the candidate writing state reaches the cumulative writing time threshold, then the moment marked as the candidate writing state will be marked as the writing moment; otherwise, it will not be marked as the writing moment. If the high-frequency vibration energy is less than the dynamic vibration energy threshold, it is marked as a candidate pen-lifting state, and the cumulative duration of the candidate pen-lifting state is obtained. If the cumulative duration of the candidate pen-lifting state reaches the cumulative pen-lifting time threshold, then the moment marked as the candidate pen-lifting state will be marked as the pen-lifting moment; otherwise, it will not be marked as the pen-lifting moment. The time interval defined by the moment the pen is put down and the moment the pen is lifted is recorded as the writing time interval.

3. The smart classroom auxiliary teaching management system based on real-time feedback of intelligent handwriting data as described in claim 1, characterized in that: The process of classifying and labeling writing states based on temporal motion features of writing states is as follows: A writing state is considered valid only when both the writing stability condition and the form flatness condition are met simultaneously. The writing stability condition is that the fluctuation amplitude of the acceleration component in the preset writing sensing axis is less than the preset writing stability threshold, and the morphological flatness condition is that the phase trajectory flatness is less than the preset phase trajectory flatness threshold. If the above conditions are not met, the current writing state will be marked as a candidate abnormal behavior state.

4. The smart classroom auxiliary teaching management system based on real-time feedback of intelligent handwriting data as described in claim 3, characterized in that: The specific process for removing sensor data marked as abnormal behavior is as follows: If the cumulative time of the candidate abnormal behavior state is less than the first abnormal cumulative time threshold, and the writing states before and after the candidate abnormal behavior state are both valid writing states, then the sensor data in the current time interval will be merged into the data stream of the adjacent valid writing states, and the label of the candidate abnormal behavior state will be removed. If the cumulative time of the candidate abnormal behavior state is less than the first abnormal cumulative time threshold, and the writing states of the time sequence before and after the candidate abnormal behavior state are not all valid writing states, then the current state is maintained and the writing state of the next time sequence is monitored. If the cumulative time of the candidate abnormal behavior state is not less than the second abnormal cumulative time threshold, then the current writing state is marked as an abnormal behavior state and the corresponding data is removed. If the cumulative time of the candidate abnormal behavior state is between the first abnormal cumulative time threshold and the second abnormal cumulative time threshold, then the current state is maintained and the writing state of the next time series is monitored. The second abnormal cumulative time threshold is greater than the first abnormal cumulative time threshold.

5. The smart classroom auxiliary teaching management system based on real-time feedback of intelligent handwriting data as described in claim 1, characterized in that: The trajectory reconstruction and drift elimination module specifically includes: The displacement is obtained by performing a secondary cumulative transformation on the removed acceleration data; During the cumulative transformation process, the high-frequency vibration energy and the gyroscope angular velocity magnitude are monitored in real time within the writing time interval; If the high-frequency vibration energy is less than the dynamic vibration energy threshold and the gyroscope angular velocity magnitude is less than the preset angular velocity threshold, then the pen tip is determined to be stationary, and the horizontal velocity vector in the writing plane is set to zero.

6. The smart classroom auxiliary teaching management system based on real-time feedback of intelligent handwriting data as described in claim 1, characterized in that: The abnormal behavior identification and cleanup module also includes: During the writing time interval, the acceleration data is cumulatively transformed to generate a velocity sequence, and a first-order high-pass filter is used to filter the velocity sequence to suppress DC component drift. Set a speed threshold, set the speed amplitude of data points whose speed amplitude is less than the speed threshold to zero, count the number of valid sampling points with speed amplitude greater than zero, and record it as the motion amplitude feature; Obtain the zero-crossing rate of the resultant acceleration in two orthogonal directions within the writing plane; The relative displacement sequence is generated by cumulative transformation based on the velocity sequence, and the standard deviation of displacement along the pen length direction and the standard deviation of displacement along the pen width direction are calculated within the writing time interval. If the motion amplitude feature is less than the preset sampling lower limit threshold, and the zero-crossing rate of the composite acceleration in the two orthogonal directions in the writing plane is greater than the preset zero-crossing upper limit threshold, and the displacement standard deviation along the pen body length direction and the displacement standard deviation along the pen body width direction are both less than the preset displacement aggregation threshold, then the current writing state is marked as an invalid stroke behavior state in the abnormal behavior state, and the corresponding data is removed.

7. The smart classroom auxiliary teaching management system based on real-time feedback of intelligent handwriting data as described in claim 6, characterized in that: The step of marking the current writing state as an invalid stroke behavior state in the abnormal behavior state also includes: The effective writing distance is calculated by performing cumulative transformation on the writing time interval where the speed amplitude is not less than the speed threshold. If the effective writing distance is less than the preset effective displacement threshold, the current writing state is marked as an invalid stroke behavior state in the abnormal behavior state, and the corresponding data is removed.

8. The smart classroom auxiliary teaching management system based on real-time feedback of intelligent handwriting data as described in claim 1, characterized in that: The writing feature statistics are obtained from the handwriting trajectory output by the trajectory reconstruction and drift elimination module. The specific method for obtaining these statistics is as follows: Extract the trajectory segments that are determined to be in a valid writing state, calculate the total path length of all trajectory segments that are determined to be in a valid writing state, and record it as the valid writing distance; Calculate the average speed of the trajectory segment judged as valid writing state within a preset time statistical interval, and record it as writing fluency; Calculate the percentage of effective writing time within the preset time statistical interval and record it as classroom focus; The statistical measures of writing characteristics include effective writing distance, writing fluency, and classroom focus.

9. The smart classroom auxiliary teaching management system based on real-time feedback of intelligent handwritten data as described in claim 1, characterized in that: The specific process for generating real-time teaching prompts is as follows: Calculate the growth rate of the effective writing distance in the current time window relative to the effective writing distance in the previous time window within the preset time statistical interval, and record it as the effective writing distance growth rate. If the number of abnormal behaviors detected within the preset time statistical interval exceeds the preset number threshold, or the effective writing distance growth rate is less than the preset effective writing distance growth rate threshold, then a attention prompt control instruction will be generated for display on the teacher's end. If the average classroom focus of all students is lower than the preset focus threshold and the average writing fluency is lower than the preset fluency threshold, a control instruction will be generated to prompt the teacher to increase the time for explaining the questions.

Images