Computer base inclination angle self-adaptive recommendation method and system based on user habit modeling

Abstract

The invention provides a computer base inclination angle self-adaptive recommendation method and system based on user habit modeling. The method comprises the steps that a state threshold value of a computer in a user use state is collected; establishing a computer base inclination angle recommendation model taking a human body state as a control variable by taking human engineering as a constraint condition of a computer base inclination angle; collecting the time sequence variable quantity of the user pose and a computer base inclination angle corresponding to the user pose; and training the prediction model through the time sequence variable quantity of the user pose and the computer base inclination angle corresponding to the user pose to obtain an optimal computer base inclination angle which enables a predicted inclination angle error to be smaller than a preset error. According to the method, the optimal base inclination angle recommendation value of the whole service cycle is generated at a time based on the initially set computer height and the initial pose parameters of the user through fusion optimization of the human engineering hard constraint and the historical habit inclination angle of the user, and collaborative adaptation of health specifications and personalized habits is achieved.

Description

Technical Field

[0001] This invention relates to the field of tilt angle recommendation technology, specifically to a computer stand tilt angle adaptive recommendation method and system based on user habit modeling. Background Technology

[0002] With the widespread use of portable computers, users' demands for office comfort and health are increasing. The tilt angle of the computer stand, as a key parameter affecting human posture, is directly related to cervical spine load and visual comfort.

[0003] Traditional computer stands use fixed tilt angles or manual mechanical adjustments, which cannot dynamically adjust according to the user's real-time posture. Users need to frequently intervene manually to match different usage scenarios (such as typing, reading, and drawing). The operation is cumbersome and relies on subjective experience. Long-term use can easily lead to cumulative muscle strain due to non-optimal tilt angles. Although existing smart computer stands can provide initial tilt angle recommendations based on ergonomic principles (such as eye level and cervical spine neutral position), they ignore the natural posture drift of users during long-term use. In addition, users do not frequently adjust the tilt angle in real time or during use when using a computer. Instead, they adjust the tilt angle based on the needs of the first use. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a computer stand tilt angle adaptive recommendation method and system based on user habit modeling, in order to solve the problems mentioned in the background.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a computer stand tilt angle adaptive recommendation method based on user habit modeling, comprising the following steps:

[0006] Collect the computer's status thresholds under user usage conditions;

[0007] The state thresholds include computer height and computer stand tilt angle;

[0008] Using ergonomics as a constraint on the tilt angle of the computer stand, a recommended model for the tilt angle of the computer stand with human body status as the control variable is established.

[0009] Collect the time-series changes in user pose and the corresponding tilt angle of the computer base;

[0010] The prediction model is trained by measuring the time-series changes in user pose and the tilt angle of the computer base corresponding to the user pose, and the optimal tilt angle of the computer base that makes the prediction tilt angle error less than the preset error is obtained.

[0011] Collect the user's initial pose and the set computer height, and input the initial pose and the set computer height into the trained prediction model, and output the predicted change in user pose over time.

[0012] The user's pose change over time is input into the recommendation model to obtain the predicted tilt angle of the computer stand, and the predicted tilt angle of the computer stand is recommended to the user.

[0013] As a preferred embodiment, the computer stand tilt angle recommendation model with human body state as the control variable is specifically as follows:

[0014] ;

[0015] in, Indicates the recommended tilt angle. Indicates the cervical spine flexion-extension angle. Indicates the angle of view (pitch). Indicates the ideal cervical spine flexion-extension angle. Indicates the ideal line-of-sight pitch angle. This represents the customary weighting coefficient. Indicates the current tilt angle of the computer stand. Indicates the historically customary angle of inclination;

[0016] The formula for hard constraints in ergonomics is:

[0017] ;

[0018] Where d represents the horizontal distance from the human eye to the screen.

[0019] In a preferred embodiment, the process of training the prediction model using the time-series changes in the user's pose and the corresponding tilt angle of the computer base to obtain the optimal tilt angle of the computer base that makes the prediction tilt angle error less than a preset error includes:

[0020] Acquire time-series training data over multiple complete cycles of user computer use, with each cycle starting from the moment the user begins using the computer;

[0021] The training data includes:

[0022] User initial pose parameters ,in, Indicates the cervical spine flexion-extension angle. d(0) represents the tilt angle of the view, d(0) represents the horizontal distance from the viewer's eye to the screen, and h is the computer height initially set by the user.

[0023] User pose change sequence over time ,in ;

[0024] The prediction model is trained using time-series training data;

[0025] Using the user's initial pose parameters and the user's initial computer height as model inputs, the output is a predicted sequence of the user's pose over time.

[0026] Optimize the trained prediction model and calculate the total loss function;

[0027] The total loss function is obtained by weighting the first loss term and the second loss term, and the calculation formula is as follows:

[0028] ;

[0029] in, Represents the total loss item. Indicates the first loss item. Indicates the second loss item. , All of these represent weighting coefficients.

[0030] In a preferred embodiment, the first loss term is specifically the mean square error between the predicted pose change and the actual pose change:

[0031] ;

[0032] Where T represents the period duration, This represents the predicted change in pose. Indicates the actual change in pose;

[0033] The second loss term represents the absolute error between the predicted tilt angle and the initial recommended tilt angle:

[0034] ;

[0035] in, Indicates the predicted dip angle. Indicates the initial recommended tilt angle;

[0036] Furthermore, the optimized prediction model must meet the following requirements. Converging to a stable value, Less than the preset threshold.

[0037] In a preferred embodiment, the specific process of acquiring the user's initial pose and the set computer height, inputting the initial pose and the set computer height into the trained prediction model, and outputting the predicted change in user pose over time includes:

[0038] Obtain the user's initial pose parameters when they start using the computer, and the computer height initially set by the user;

[0039] Using the user's initial pose and the initial computer height as input, a prediction model is used to predict the sequence of changes in the user's pose over time. T is the preset prediction period duration;

[0040] The sequence of user pose changes over time is discretized and stored according to the time step.

[0041] Among them, the sequence of changes in user pose over time Each time step t includes the cervical spine pitch angle and the line of sight pitch angle.

[0042] As a preferred embodiment, a computer stand tilt angle adaptive recommendation system based on user habit modeling is provided to implement the above-mentioned computer stand tilt angle adaptive recommendation method based on user habit modeling, including:

[0043] The data acquisition module is used to collect the computer's status thresholds under user usage conditions;

[0044] The state thresholds include computer height and computer stand tilt angle;

[0045] Model building model, used to build computer stand tilt angle recommendation and prediction models;

[0046] The tilt angle prediction module allows users to collect their initial pose and set computer height. The prediction model outputs the predicted change in user pose over time, and the change in user pose over time is input into the recommendation model to obtain the predicted tilt angle of the computer base.

[0047] The control and adjustment module is used to recommend the predicted tilt angle of the computer stand to the user.

[0048] As a preferred embodiment, the model construction model uses ergonomics as the constraint condition for the computer base tilt angle, and establishes a computer base tilt angle recommendation model with human body state as the control variable.

[0049] The computer stand tilt angle recommendation model, which uses human body state as the control variable, is specifically as follows:

[0050] ;

[0051] in, Indicates the recommended tilt angle. Indicates the cervical spine flexion-extension angle. Indicates the angle of view (pitch). Indicates the ideal cervical spine flexion-extension angle. Indicates the ideal line-of-sight pitch angle. This represents the customary weighting coefficient. Indicates the current tilt angle of the computer stand. Indicates the historically customary angle of inclination;

[0052] The formula for hard constraints in ergonomics is:

[0053] ;

[0054] Where d represents the horizontal distance from the human eye to the screen.

[0055] As a preferred embodiment, the model construction model is trained by the time series change of the user's pose and the tilt angle of the computer base corresponding to the user's pose, so as to obtain the optimal tilt angle of the computer base that makes the prediction tilt angle error less than the preset error.

[0056] The training process includes: acquiring time-series training data over multiple complete cycles of the user's computer use, with each cycle starting from the moment the user begins using the computer;

[0057] The training data includes:

[0058] User initial pose parameters ,in, Indicates the cervical spine flexion-extension angle. d(0) represents the tilt angle of the view, d(0) represents the horizontal distance from the viewer's eye to the screen, and h is the computer height initially set by the user.

[0059] User pose change sequence over time ,in ;

[0060] The prediction model is trained using time-series training data;

[0061] Using the user's initial pose parameters and the user's initial computer height as model inputs, the output is a predicted sequence of the user's pose over time.

[0062] Optimize the trained prediction model and calculate the total loss function;

[0063] The total loss function is obtained by weighting the first loss term and the second loss term, and the calculation formula is as follows:

[0064] ;

[0065] in, Represents the total loss item. Indicates the first loss item. Indicates the second loss item. , All of these represent weighting coefficients.

[0066] In a preferred embodiment, the first loss term is specifically the mean square error between the predicted pose change and the actual pose change:

[0067] ;

[0068] Where T represents the period duration, This represents the predicted change in pose. Indicates the actual change in pose;

[0069] The second loss term represents the absolute error between the predicted tilt angle and the initial recommended tilt angle:

[0070] ;

[0071] in, Indicates the predicted dip angle. Indicates the initial recommended tilt angle;

[0072] Furthermore, the optimized prediction model must meet the following requirements. Converging to a stable value, Less than the preset threshold.

[0073] As a preferred embodiment, the specific process by which the tilt prediction module obtains the predicted tilt angle of the computer base includes:

[0074] Obtain the user's initial pose parameters when they start using the computer, and the computer height initially set by the user;

[0075] Using the user's initial pose and the initial computer height as input, a prediction model is used to predict the sequence of changes in the user's pose over time. T is the preset prediction period duration;

[0076] The sequence of user pose changes over time is discretized and stored according to the time step.

[0077] Among them, the sequence of changes in user pose over time Each time step t includes the cervical spine pitch angle and the line of sight pitch angle.

[0078] This invention provides a computer stand tilt angle adaptive recommendation method and system based on user habit modeling, which has the following beneficial effects: By integrating and optimizing ergonomic hard constraints with the user's historical habit tilt angle, the optimal stand tilt angle recommendation value for the entire usage cycle is generated in one go based on the initially set computer height and the user's initial posture parameters. This achieves coordinated adaptation between health standards and personalized habits, effectively avoiding the risk of cervical strain and visual fatigue caused by continuous posture deviation during use. Furthermore, by using a recurrent neural network to accurately model the long-term posture evolution law of the user's initial posture and fixed height, the dynamic posture trajectory within the complete usage cycle is predicted in a forward-looking manner, driving the tilt angle recommendation model to output a stable tilt angle setting covering the entire cycle, significantly improving the timeliness and reliability of the recommendation. At the same time, by using a dual-loss constraint mechanism that integrates posture prediction error and healthy tilt angle deviation during the training phase, it ensures that the single recommendation result not only conforms to the user's natural habits but also strictly meets the ergonomic health boundary, preventing the cumulative impact of posture deterioration on health and comfort from the source. It is suitable for efficient and healthy tilt angle adaptive management in computer usage scenarios that require long-term stable operation. Attached Figure Description

[0079] Figure 1 This is a flowchart of the computer stand tilt angle adaptive recommendation method based on user habit modeling of the present invention;

[0080] Figure 2 This is a block diagram of the computer stand tilt angle adaptive recommendation system based on user habit modeling of the present invention. Detailed Implementation

[0081] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0082] The following disclosure provides many different embodiments or examples for implementing various structures of the invention. To simplify the disclosure, specific examples of components and arrangements are described below. These are merely examples and are not intended to limit the invention. Furthermore, reference numerals and / or letters may be repeated in different examples; such repetition is for simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or arrangements discussed. In addition, examples of various specific processes and materials are provided in this invention, but those skilled in the art will recognize the application of other processes and / or the use of other materials.

[0083] like Figure 1As shown, this embodiment of the invention provides a computer stand tilt angle adaptive recommendation method based on user habit modeling, including the following steps:

[0084] S1: Collect the computer's status threshold under user usage conditions;

[0085] The state thresholds include computer height and computer stand tilt angle;

[0086] Specifically, the system uses sensors to obtain the computer height and base tilt angle adjusted by the user in different usage scenarios as state thresholds, monitors changes in computer height and base tilt angle, and records a stable state when the rate of change is lower than the preset threshold and continues for a period of time.

[0087] S2: Using ergonomics as the constraint condition for the tilt angle of the computer base, establish a recommended model for the tilt angle of the computer base with human body status as the control variable;

[0088] It should be noted that the human body state specifically refers to the user's initial posture, which includes the cervical spine curvature angle, the line of sight pitch angle, and the horizontal distance from the human eye to the screen.

[0089] Specifically, at the beginning of a user's computer use, the initial computer stand tilt angle can be recommended based on the user's posture while using the computer, using ergonomics to reflect the computer's height.

[0090] In this embodiment, the computer stand tilt angle recommendation model with human body state as the control variable is specifically as follows:

[0091] ;

[0092] in, Indicates the recommended tilt angle. Indicates the cervical spine flexion-extension angle. Indicates the angle of view (pitch). Indicates the ideal cervical spine flexion-extension angle. Indicates the ideal line-of-sight pitch angle. This represents the customary weighting coefficient. Indicates the current tilt angle of the computer stand. Indicates the historical habit of tilt angle.

[0093] The formula for hard constraints in ergonomics is:

[0094] ;

[0095] Where d represents the horizontal distance from the human eye to the screen.

[0096] In this embodiment, hard constraints are established through ergonomic parameters to ensure that the recommended tilt angle conforms to the natural physiological curve of the human body, avoiding cervical fatigue or vision damage caused by long-term use, and taking into account both health and comfort.

[0097] S3: Collect the time-series changes in user pose and the tilt angle of the computer base corresponding to the user pose;

[0098] Specifically, the data is collected on the changes in the user's posture over time during several periods of computer use, and the corresponding tilt angle of the computer stand after the change in the user's posture is analyzed.

[0099] In each of these cycles, the starting point is when a user begins using a computer.

[0100] It's understandable that users might initially adjust the height of their computer, but they generally won't do so again afterward.

[0101] In this embodiment, a user-specific tilt angle recommendation model is constructed by collecting data on the user's stable usage status and posture changes over time in different scenarios. This model can learn the user's long-term usage habits and provide a base tilt angle recommendation that matches individual operating preferences, reducing the frequency of manual adjustments and improving ease of use.

[0102] S4: Train the prediction model by using the time series changes in user pose and the tilt angle of the computer base corresponding to the user pose, and obtain the optimal tilt angle of the computer base that makes the prediction tilt angle error less than the preset error.

[0103] Specifically, the system acquires time-series training data (time-series changes in user pose) over multiple complete cycles of user computer use, with each cycle starting from the moment the user begins using the computer.

[0104] The training data includes:

[0105] User initial pose parameters ,in, Indicates the cervical spine flexion-extension angle. d(0) represents the tilt angle of the view, d(0) represents the horizontal distance from the viewer's eye to the screen, and h is the computer height initially set by the user.

[0106] User pose change sequence over time ,in ;

[0107] The prediction model is trained using time-series training data;

[0108] Specifically, the model takes the user's initial pose parameters and the user's initial computer height as input and outputs a predicted sequence of the user's pose over time.

[0109] It should be noted that the prediction model uses a recurrent neural network architecture, which includes long short-term memory units or gated recurrent units.

[0110] Optimize the trained prediction model and calculate the total loss function;

[0111] The total loss function is obtained by weighting the first loss term and the second loss term, and the calculation formula is as follows:

[0112] ;

[0113] in, Represents the total loss item. Indicates the first loss item. Indicates the second loss item. , All represent weighting coefficients;

[0114] Specifically, the first loss term is the mean square error between the predicted pose change and the actual pose change:

[0115] ;

[0116] Where T represents the period duration, This represents the predicted change in pose. Indicates the actual change in pose;

[0117] The second loss term represents the absolute error between the predicted tilt angle and the initial recommended tilt angle:

[0118] ;

[0119] in, Indicates the predicted dip angle. Indicates the initial recommended tilt angle;

[0120] It should be noted that the predicted tilt angle is determined by the change in predicted pose. The predicted tilt angle generated after inputting the recommendation model.

[0121] Furthermore, the optimized prediction model must meet the following requirements. Converging to a stable value, Less than the preset threshold.

[0122] It should be noted that through periodic data collection and model training, the system can dynamically update the user habit model to adapt to changes in users' posture habits over a long period of time, ensuring that the recommendation method is effective in the long term.

[0123] S5: Collect the user's initial pose and the set computer height, input the initial pose and the set computer height into the trained prediction model, and output the predicted change in user pose over time.

[0124] Specifically, it involves obtaining the initial pose parameters of the user when they begin using the computer. The user's initial computer height h;

[0125] Using the user's initial pose and the initial computer height as input, a prediction model is used to predict the sequence of changes in the user's pose over time. T is the preset prediction period duration;

[0126] The sequence of user pose changes over time is discretized and stored according to the time step.

[0127] Among them, the sequence of changes in user pose over time Each time step t contains Indicates the cervical spine flexion-extension angle. Indicates the angle of view (tilt / slope).

[0128] S6: Input the change in user pose over time into the recommendation model to obtain the predicted tilt angle of the computer base, and recommend the predicted tilt angle of the computer base to the user.

[0129] Specifically, through an automated recommendation process, users no longer need to frequently adjust the base tilt angle manually, making it especially suitable for scenarios where users need to maintain a fixed posture for extended periods of time while working. This reduces the user's operational burden and, at the same time, prevents occupational diseases through ergonomic recommendations, thus improving the overall user experience.

[0130] The computer stand tilt angle adaptive recommendation method based on user habit modeling provided in this embodiment optimizes the tilt angle by fusing ergonomic hard constraints with the user's historical habit tilt angle. Based on the initially set computer height and the user's initial posture parameters, it generates the optimal tilt angle recommendation value for the entire usage cycle in one go, achieving a coordinated adaptation between health standards and personalized habits. This effectively avoids the risk of cervical strain and visual fatigue caused by continuous posture deviation during use. Furthermore, it accurately models the long-term posture evolution law of the user's initial posture and fixed height through a recurrent neural network, proactively predicting the posture dynamic trajectory within the complete usage cycle. This drives the tilt angle recommendation model to output a stable tilt angle setting covering the entire cycle, significantly improving the timeliness and reliability of the recommendation. At the same time, by leveraging a dual-loss constraint mechanism that integrates posture prediction error and healthy tilt angle deviation during the training phase, it ensures that the recommendation result of each time not only conforms to the user's natural habits but also strictly meets the ergonomic health boundaries. This prevents the cumulative impact of posture deterioration on health and comfort from the source, making it suitable for efficient and healthy tilt angle adaptive management in computer usage scenarios that require long-term stable operation.

[0131] like Figure 2 As shown, this embodiment also provides a computer stand tilt angle adaptive recommendation system based on user habit modeling, used to implement the above-mentioned computer stand tilt angle adaptive recommendation method based on user habit modeling, including:

[0132] The data acquisition module is used to collect the computer's status thresholds under user usage conditions;

[0133] The state thresholds include computer height and computer stand tilt angle;

[0134] Model building model, used to build computer stand tilt angle recommendation and prediction models;

[0135] The tilt angle prediction module allows users to collect their initial pose and set computer height. The prediction model outputs the predicted change in user pose over time, and the change in user pose over time is input into the recommendation model to obtain the predicted tilt angle of the computer base.

[0136] The tilt angle recommendation module is used to recommend the predicted tilt angle of the computer stand to the user.

[0137] It should be noted that the data acquisition module collects the computer's status thresholds under user conditions through sensors, and the specific type of sensor used can be selected according to the actual situation.

[0138] Furthermore, the model construction model uses ergonomics as the constraint condition for the tilt angle of the computer base, and establishes a computer base tilt angle recommendation model with human body state as the control variable.

[0139] The computer stand tilt angle recommendation model, which uses human body state as the control variable, is specifically as follows:

[0140] ;

[0141] in, Indicates the recommended tilt angle. Indicates the cervical spine flexion-extension angle. Indicates the angle of view (pitch). Indicates the ideal cervical spine flexion-extension angle. Indicates the ideal line-of-sight pitch angle. This represents the customary weighting coefficient. Indicates the current tilt angle of the computer stand. Indicates the historically customary angle of inclination;

[0142] The formula for hard constraints in ergonomics is:

[0143] ;

[0144] Where d represents the horizontal distance from the human eye to the screen.

[0145] Furthermore, the model construction model is trained by the time series change of the user's pose and the tilt angle of the computer base corresponding to the user's pose, so as to obtain the optimal tilt angle of the computer base that makes the prediction tilt angle error less than the preset error.

[0146] The training process includes: acquiring time-series training data over multiple complete cycles of the user's computer use, with each cycle starting from the moment the user begins using the computer;

[0147] The training data includes:

[0148] User initial pose parameters ,in, Indicates the cervical spine flexion-extension angle. d(0) represents the tilt angle of the view, d(0) represents the horizontal distance from the viewer's eye to the screen, and h is the computer height initially set by the user.

[0149] User pose change sequence over time ,in ;

[0150] The prediction model is trained using time-series training data;

[0151] Using the user's initial pose parameters and the user's initial computer height as model inputs, the output is a predicted sequence of the user's pose over time.

[0152] Optimize the trained prediction model and calculate the total loss function;

[0153] The total loss function is obtained by weighting the first loss term and the second loss term, and the calculation formula is as follows:

[0154] ;

[0155] in, Represents the total loss item. Indicates the first loss item. Indicates the second loss item. , All of these represent weighting coefficients.

[0156] Furthermore, the first loss term specifically comprises the mean square error between the predicted pose change and the actual pose change:

[0157] ;

[0158] Where T represents the period duration, This represents the predicted change in pose. Indicates the actual change in pose;

[0159] The second loss term represents the absolute error between the predicted tilt angle and the initial recommended tilt angle:

[0160] ;

[0161] in, Indicates the predicted dip angle. Indicates the initial recommended tilt angle;

[0162] Furthermore, the optimized prediction model must meet the following requirements. Converging to a stable value, Less than the preset threshold.

[0163] Furthermore, the specific process by which the tilt prediction module obtains the predicted tilt angle of the computer base includes:

[0164] Obtain the user's initial pose parameters when they start using the computer, and the computer height initially set by the user;

[0165] Using the user's initial pose and the initial computer height as input, a prediction model is used to predict the sequence of changes in the user's pose over time. T is the preset prediction period duration;

[0166] The sequence of user pose changes over time is discretized and stored according to the time step.

[0167] Among them, the sequence of changes in user pose over time Each time step t includes the cervical spine pitch angle and the line of sight pitch angle.

[0168] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A computer stand tilt angle adaptive recommendation method based on user habit modeling, characterized in that, Includes the following steps: Collect the computer's status thresholds under user usage conditions; The state thresholds include computer height and computer stand tilt angle; Using ergonomics as a constraint on the tilt angle of the computer stand, a recommended model for the tilt angle of the computer stand with human body status as the control variable is established. Collect the time-series changes in user pose and the corresponding tilt angle of the computer base; The prediction model is trained by measuring the time-series changes in user pose and the tilt angle of the computer base corresponding to the user pose, and the optimal tilt angle of the computer base that makes the prediction tilt angle error less than the preset error is obtained. Collect the user's initial pose and the set computer height, and input the initial pose and the set computer height into the trained prediction model, and output the predicted change in user pose over time. The user's pose change over time is input into the recommendation model to obtain the predicted tilt angle of the computer stand, and the predicted tilt angle of the computer stand is recommended to the user.

2. The computer stand tilt angle adaptive recommendation method based on user habit modeling according to claim 1, characterized in that, The computer stand tilt angle recommendation model, which uses human body state as the control variable, is specifically as follows: ； in, Indicates the recommended tilt angle. Indicates the cervical spine flexion-extension angle. Indicates the angle of view (pitch). Indicates the ideal cervical spine flexion-extension angle. Indicates the ideal line-of-sight pitch angle. This represents the customary weighting coefficient. Indicates the current tilt angle of the computer stand. Indicates the historically customary angle of inclination; The formula for hard constraints in ergonomics is: ； Where d represents the horizontal distance from the human eye to the screen.

3. The computer stand tilt angle adaptive recommendation method based on user habit modeling according to claim 1, characterized in that, The process of training the prediction model using the time-series changes in user pose and the corresponding tilt angle of the computer base to obtain the optimal tilt angle of the computer base that makes the prediction tilt angle error less than a preset error includes: Acquire time-series training data over multiple complete cycles of user computer use, with each cycle starting from the moment the user begins using the computer; The training data includes: User initial pose parameters ,in, Indicates the cervical spine flexion-extension angle. d(0) represents the tilt angle of the view, d(0) represents the horizontal distance from the viewer's eye to the screen, and h is the computer height initially set by the user. User pose change sequence over time ,in ; The prediction model is trained using time-series training data; Using the user's initial pose parameters and the user's initial computer height as model inputs, the output is a predicted sequence of the user's pose over time; Optimize the trained prediction model and calculate the total loss function; The total loss function is obtained by weighting the first loss term and the second loss term, and the calculation formula is as follows: ； in, Represents the total loss item. Indicates the first loss item. Indicates the second loss item. , All of these represent weighting coefficients.

4. The computer stand tilt angle adaptive recommendation method based on user habit modeling according to claim 3, characterized in that, The first loss term is specifically the mean square error between the predicted pose change and the actual pose change: ； Where T represents the period duration, This represents the predicted change in pose. Indicates the actual change in pose; The second loss term represents the absolute error between the predicted tilt angle and the initial recommended tilt angle: ； in, Indicates the predicted dip angle. Indicates the initial recommended tilt angle; Furthermore, the optimized prediction model must meet the following requirements. Converging to a stable value, Less than the preset threshold.

5. The computer stand tilt angle adaptive recommendation method based on user habit modeling according to claim 1, characterized in that, The specific process of acquiring the user's initial pose and the set computer height, inputting the initial pose and the set computer height into the trained prediction model, and outputting the predicted change in user pose over time includes: Obtain the user's initial pose parameters when they start using the computer, and the computer height initially set by the user; Using the user's initial pose and the initial computer height as input, a prediction model is used to predict the sequence of changes in the user's pose over time. T is the preset prediction period duration; The sequence of user pose changes over time is discretized and stored according to the time step. Among them, the sequence of changes in user pose over time Each time step t includes the cervical spine pitch angle and the line of sight pitch angle.

6. A computer stand tilt angle adaptive recommendation system based on user habit modeling, used to implement the computer stand tilt angle adaptive recommendation method based on user habit modeling as described in any one of claims 1-5, characterized in that, include: The data acquisition module is used to collect the computer's status thresholds under user usage conditions; The state thresholds include computer height and computer stand tilt angle; Model building model, used to build computer stand tilt angle recommendation and prediction models; The tilt angle prediction module allows users to collect their initial pose and set computer height. The prediction model outputs the predicted change in user pose over time, and the change in user pose over time is input into the recommendation model to obtain the predicted tilt angle of the computer base. The control and adjustment module is used to recommend the predicted tilt angle of the computer stand to the user.

7. The computer stand tilt angle adaptive recommendation system based on user habit modeling according to claim 6, characterized in that, The model construction model uses ergonomics as the constraint condition for the tilt angle of the computer base, and establishes a recommended model for the tilt angle of the computer base with human body status as the control variable. The computer stand tilt angle recommendation model, which uses human body state as the control variable, is specifically as follows: ； in, Indicates the recommended tilt angle. Indicates the cervical spine flexion-extension angle. Indicates the angle of view (pitch). Indicates the ideal cervical spine flexion-extension angle. Indicates the ideal line-of-sight pitch angle. This represents the customary weighting coefficient. Indicates the current tilt angle of the computer stand. Indicates the historically customary angle of inclination; The formula for hard constraints in ergonomics is: ； Where d represents the horizontal distance from the human eye to the screen.

8. The computer stand tilt angle adaptive recommendation system based on user habit modeling according to claim 7, characterized in that, The model construction model is trained by the time series changes of the user's pose and the tilt angle of the computer base corresponding to the user's pose, so as to obtain the optimal tilt angle of the computer base that makes the prediction tilt angle error less than the preset error. The training process includes: acquiring time-series training data over multiple complete cycles of the user's computer use, with each cycle starting from the moment the user begins using the computer; The training data includes: User initial pose parameters ,in, Indicates the cervical spine flexion-extension angle. d(0) represents the tilt angle of the view, d(0) represents the horizontal distance from the viewer's eye to the screen, and h is the computer height initially set by the user. User pose change sequence over time ,in ; The prediction model is trained using time-series training data; Using the user's initial pose parameters and the user's initial computer height as model inputs, the output is a predicted sequence of the user's pose over time; Optimize the trained prediction model and calculate the total loss function; The total loss function is obtained by weighting the first loss term and the second loss term, and the calculation formula is as follows: ； in, Represents the total loss item. Indicates the first loss item. Indicates the second loss item. , All of these represent weighting coefficients.

9. A computer stand tilt angle adaptive recommendation system based on user habit modeling according to claim 8, characterized in that, The first loss term is specifically the mean square error between the predicted pose change and the actual pose change: ； Where T represents the period duration, This represents the predicted change in pose. Indicates the actual change in pose; The second loss term represents the absolute error between the predicted tilt angle and the initial recommended tilt angle: ； in, Indicates the predicted dip angle. Indicates the initial recommended tilt angle; Furthermore, the optimized prediction model must meet the following requirements. Converging to a stable value, Less than the preset threshold.

10. A computer stand tilt angle adaptive recommendation system based on user habit modeling according to claim 6, characterized in that, The specific process by which the tilt prediction module obtains the predicted tilt angle of the computer base includes: Obtain the user's initial pose parameters when they start using the computer, and the computer height initially set by the user; Using the user's initial pose and the initial computer height as input, a prediction model is used to predict the sequence of changes in the user's pose over time. T is the preset prediction period duration; The sequence of user pose changes over time is discretized and stored according to the time step. Among them, the sequence of changes in user pose over time Each time step t includes the cervical spine pitch angle and the line of sight pitch angle.

Images