CN118949364A - Intelligent control circuit board and control method for treadmill - Google Patents
Intelligent control circuit board and control method for treadmill Download PDFInfo
- Publication number
- CN118949364A CN118949364A CN202411000354.6A CN202411000354A CN118949364A CN 118949364 A CN118949364 A CN 118949364A CN 202411000354 A CN202411000354 A CN 202411000354A CN 118949364 A CN118949364 A CN 118949364A
- Authority
- CN
- China
- Prior art keywords
- running speed
- time sequence
- sequence
- heart rate
- vector
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H05—ELECTRIC TECHNIQUES NOT OTHERWISE PROVIDED FOR
- H05K—PRINTED CIRCUITS; CASINGS OR CONSTRUCTIONAL DETAILS OF ELECTRIC APPARATUS; MANUFACTURE OF ASSEMBLAGES OF ELECTRICAL COMPONENTS
- H05K1/00—Printed circuits
- H05K1/02—Details
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63B—APPARATUS FOR PHYSICAL TRAINING, GYMNASTICS, SWIMMING, CLIMBING, OR FENCING; BALL GAMES; TRAINING EQUIPMENT
- A63B22/00—Exercising apparatus specially adapted for conditioning the cardio-vascular system, for training agility or co-ordination of movements
- A63B22/02—Exercising apparatus specially adapted for conditioning the cardio-vascular system, for training agility or co-ordination of movements with movable endless bands, e.g. treadmills
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63B—APPARATUS FOR PHYSICAL TRAINING, GYMNASTICS, SWIMMING, CLIMBING, OR FENCING; BALL GAMES; TRAINING EQUIPMENT
- A63B24/00—Electric or electronic controls for exercising apparatus of preceding groups; Controlling or monitoring of exercises, sportive games, training or athletic performances
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63B—APPARATUS FOR PHYSICAL TRAINING, GYMNASTICS, SWIMMING, CLIMBING, OR FENCING; BALL GAMES; TRAINING EQUIPMENT
- A63B24/00—Electric or electronic controls for exercising apparatus of preceding groups; Controlling or monitoring of exercises, sportive games, training or athletic performances
- A63B24/0075—Means for generating exercise programs or schemes, e.g. computerized virtual trainer, e.g. using expert databases
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63B—APPARATUS FOR PHYSICAL TRAINING, GYMNASTICS, SWIMMING, CLIMBING, OR FENCING; BALL GAMES; TRAINING EQUIPMENT
- A63B24/00—Electric or electronic controls for exercising apparatus of preceding groups; Controlling or monitoring of exercises, sportive games, training or athletic performances
- A63B24/0087—Electric or electronic controls for exercising apparatus of groups A63B21/00 - A63B23/00, e.g. controlling load
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63B—APPARATUS FOR PHYSICAL TRAINING, GYMNASTICS, SWIMMING, CLIMBING, OR FENCING; BALL GAMES; TRAINING EQUIPMENT
- A63B71/00—Games or sports accessories not covered in groups A63B1/00 - A63B69/00
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63B—APPARATUS FOR PHYSICAL TRAINING, GYMNASTICS, SWIMMING, CLIMBING, OR FENCING; BALL GAMES; TRAINING EQUIPMENT
- A63B2230/00—Measuring physiological parameters of the user
- A63B2230/04—Measuring physiological parameters of the user heartbeat characteristics, e.g. ECG, blood pressure modulations
- A63B2230/06—Measuring physiological parameters of the user heartbeat characteristics, e.g. ECG, blood pressure modulations heartbeat rate only
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63B—APPARATUS FOR PHYSICAL TRAINING, GYMNASTICS, SWIMMING, CLIMBING, OR FENCING; BALL GAMES; TRAINING EQUIPMENT
- A63B2230/00—Measuring physiological parameters of the user
- A63B2230/04—Measuring physiological parameters of the user heartbeat characteristics, e.g. ECG, blood pressure modulations
- A63B2230/06—Measuring physiological parameters of the user heartbeat characteristics, e.g. ECG, blood pressure modulations heartbeat rate only
- A63B2230/062—Measuring physiological parameters of the user heartbeat characteristics, e.g. ECG, blood pressure modulations heartbeat rate only used as a control parameter for the apparatus
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2123/00—Data types
- G06F2123/02—Data types in the time domain, e.g. time-series data
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Physical Education & Sports Medicine (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Mathematical Physics (AREA)
- Vascular Medicine (AREA)
- Cardiology (AREA)
- Computational Linguistics (AREA)
- Microelectronics & Electronic Packaging (AREA)
- Measuring Pulse, Heart Rate, Blood Pressure Or Blood Flow (AREA)
Abstract
The application discloses an intelligent control circuit board and a control method for a running machine, which are used for monitoring and collecting running speed of the running machine and heart rate values of an athlete in real time through a sensor group, inputting the data into the intelligent control circuit board, analyzing and interactively processing time sequence data of the running speed and the heart rate time sequence data by utilizing a data processing and analyzing algorithm based on artificial intelligence and deep learning in the intelligent control circuit board, so that time sequence correlation relation and interactive influence between the running speed and the heart rate of the athlete are learned and captured, and gradient of the running machine is adaptively controlled by utilizing interactive fusion semantics of the running speed and the heart rate of the athlete. Therefore, more intelligent exercise experience can be provided for the user according to the user's ability, so that personalized requirements of the user can be met, safety accidents caused by excessive exercise of the user can be avoided, and the safety of the running machine is improved.
Description
Technical Field
The application relates to the field of intelligent control, in particular to an intelligent control circuit board for a running machine and a control method.
Background
Running is becoming a highly efficient aerobic exercise and is becoming popular with many people worldwide. With the acceleration of the pace of life and the pursuit of healthy lifestyles, a treadmill as a convenient indoor exercise device is becoming an indispensable part of families and gymnasiums. The indoor running machine is not limited by weather, time and places, and a user can run at any time according to the time arrangement and the exercise requirement of the user, so that a stable running environment is provided for the user.
Currently, the main functions of treadmills are focused on providing a motion platform, a driving device, and simple data monitoring. Some high-end treadmills are also equipped with touch screens that can display athletic data, play music or video, etc. However, the existing running machine can only provide a few fixed exercise modes, such as jogging, fast running, and the like, and the interaction and adjustment of the running machine are completed through physical buttons or touch screens, so that the personalized exercise requirements of users cannot be met, and more intelligent and personalized exercise experience cannot be provided. Furthermore, in some cases, conventional treadmills may not respond in time to abnormal conditions of the user, such as excessive fatigue or abnormally elevated heart rates, thereby increasing exercise risk.
Accordingly, an optimized control scheme for a treadmill is desired.
Disclosure of Invention
The present application has been made to solve the above-mentioned technical problems. The embodiment of the application provides an intelligent control circuit board and a control method for a running machine, which are used for monitoring and collecting running speed of the running machine and heart rate values of an athlete in real time through a sensor group, inputting the data into the intelligent control circuit board, analyzing and interactively processing running speed time sequence data and heart rate time sequence data by utilizing an artificial intelligence and deep learning-based data processing and analyzing algorithm in the intelligent control circuit board, so that time sequence correlation relation and interactive influence between the running speed and the heart rate of the athlete are learned and captured, and gradient of the running machine is adaptively controlled by utilizing interactive fusion semantics of the two. Therefore, more intelligent exercise experience can be provided for the user according to the user's ability, so that personalized requirements of the user can be met, safety accidents caused by excessive exercise of the user can be avoided, and the safety of the running machine is improved.
According to one aspect of the present application, there is provided a control method of an intelligent control circuit board for a treadmill, comprising:
acquiring a time queue of running speeds and a time queue of heart rate values acquired by a sensor group;
performing time sequence coding on the time sequence of the running speed and the time sequence of the heart rate value to obtain a sequence of running speed local time sequence associated implicit characteristic vectors and a sequence of heart rate local time sequence associated implicit characteristic vectors;
Inputting the sequence of the running speed local time sequence associated implicit characteristic vector and the sequence of the heart rate local time sequence associated implicit characteristic vector into a node characteristic propagation network based on a node energy attenuation mechanism to obtain a running speed time sequence propagation aggregate expression vector and a heart rate time sequence propagation aggregate expression vector;
performing characteristic interaction response processing on the running speed time sequence propagation aggregation expression vector and the heart rate time sequence propagation aggregation expression vector to obtain a remarkable running speed-heart rate time sequence interaction fusion expression vector;
Based on the significant running speed-heart rate timing interaction fusion representation vector, control instructions are generated for representing increasing the treadmill grade value, decreasing the treadmill grade value, and maintaining the treadmill grade value.
Compared with the prior art, the intelligent control circuit board for the running machine and the control method thereof provided by the application monitor and collect running speed of the running machine and heart rate value of an athlete in real time through the sensor group, input the data into the intelligent control circuit board, analyze and interactively process the running speed time sequence data and the heart rate time sequence data in the intelligent control circuit board by utilizing a data processing and analyzing algorithm based on artificial intelligence and deep learning, so as to learn and capture time sequence correlation relation and interactive influence between the running speed and the heart rate of the athlete, and adaptively control gradient of the running machine by utilizing interactive fusion semantics of the running speed and the heart rate value of the athlete. Therefore, more intelligent exercise experience can be provided for the user according to the user's ability, so that personalized requirements of the user can be met, safety accidents caused by excessive exercise of the user can be avoided, and the safety of the running machine is improved.
Drawings
The above and other objects, features and advantages of the present application will become more apparent by describing embodiments of the present application in more detail with reference to the attached drawings. The accompanying drawings are included to provide a further understanding of embodiments of the application and are incorporated in and constitute a part of this specification, illustrate the application and together with the embodiments of the application, and not constitute a limitation to the application. In the drawings, like reference numerals generally refer to like parts or steps.
FIG. 1 is a flow chart of a control method of an intelligent control circuit board for a treadmill according to an embodiment of the present application;
FIG. 2 is a data flow diagram of a control method of an intelligent control circuit board for a treadmill according to an embodiment of the present application;
Fig. 3 is a flowchart of a training phase of a control method of an intelligent control circuit board for a treadmill according to an embodiment of the present application.
Detailed Description
Hereinafter, exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be apparent that the described embodiments are only some embodiments of the present application and not all embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described herein.
As used in the specification and in the claims, the terms "a," "an," "the," and/or "the" are not specific to a singular, but may include a plurality, unless the context clearly dictates otherwise. In general, the terms "comprises" and "comprising" merely indicate that the steps and elements are explicitly identified, and they do not constitute an exclusive list, as other steps or elements may be included in a method or apparatus.
Although the present application makes various references to certain modules in a system according to embodiments of the present application, any number of different modules may be used and run on a user terminal and/or server. The modules are merely illustrative, and different aspects of the systems and methods may use different modules.
A flowchart is used in the present application to describe the operations performed by a system according to embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in order precisely. Rather, the various steps may be processed in reverse order or simultaneously, as desired. Also, other operations may be added to or removed from these processes.
Hereinafter, exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be apparent that the described embodiments are only some embodiments of the present application and not all embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described herein.
The existing running machine can only provide a few fixed exercise modes, such as jogging, sprinting and the like, and can not meet the personalized exercise requirements of users or provide more intelligent and personalized exercise experience through interaction and adjustment of the running machine by physical buttons or touch screens. Furthermore, in some cases, conventional treadmills may not respond in time to abnormal conditions of the user, such as excessive fatigue or abnormally elevated heart rates, thereby increasing exercise risk. Accordingly, an optimized control scheme for a treadmill is desired.
In the technical scheme of the application, a control method of an intelligent control circuit board for a running machine is provided. FIG. 1 is a flow chart of a control method of an intelligent control circuit board for a treadmill according to an embodiment of the present application;
Fig. 2 is a data flow diagram of a control method of an intelligent control circuit board for a treadmill according to an embodiment of the present application. As shown in fig. 1 and 2, the control method of the intelligent control circuit board for the running machine according to the embodiment of the application includes the steps of: s1, acquiring a time queue of running speeds and a time queue of heart rate values acquired by a sensor group; s2, performing time sequence coding on the time sequence of the running speed and the time sequence of the heart rate value to obtain a sequence of running speed local time sequence associated implicit characteristic vectors and a sequence of heart rate local time sequence associated implicit characteristic vectors; s3, inputting the sequence of the running speed local time sequence associated implicit characteristic vector and the sequence of the heart rate local time sequence associated implicit characteristic vector into a node characteristic propagation network based on a node energy attenuation mechanism to obtain a running speed time sequence propagation aggregate expression vector and a heart rate time sequence propagation aggregate expression vector; s4, performing characteristic interaction response processing on the running speed time sequence propagation aggregation expression vector and the heart rate time sequence propagation aggregation expression vector to obtain a remarkable running speed-heart rate time sequence interaction fusion expression vector; s5, generating control instructions based on the significant running speed-heart rate time sequence interaction fusion representation vector, wherein the control instructions are used for representing increasing the gradient value of the running machine, decreasing the gradient value of the running machine and maintaining the gradient value of the running machine.
In particular, the S1 and the S2 acquire a time queue of running speeds and a time queue of heart rate values acquired by a sensor group; and performing time sequence coding on the time sequence of the running speed and the time sequence of the heart rate value to obtain a sequence of running speed local time sequence associated implicit characteristic vectors and a sequence of heart rate local time sequence associated implicit characteristic vectors. In a specific example of the present application, the time series of running speeds and the time series of heart rate values are input to a 1D-CNN model based sequence encoder to obtain the sequence of running speed local timing related implicit characteristic vectors and the sequence of heart rate local timing related implicit characteristic vectors. Namely, inputting the time sequence of the running speed and the time sequence of the heart rate value into a sequence encoder based on a 1D-CNN model for feature mining so as to extract local time sequence implicit association feature information of the running speed and the heart rate value in the time dimension respectively, thereby obtaining a sequence of the running speed local time sequence implicit association feature vector and a sequence of the heart rate local time sequence implicit association feature vector. It is worth mentioning that 1D-CNN is a one-dimensional convolutional neural network, suitable for processing sequence data, such as time sequence data or text data.
In particular, the step S3 is to input the sequence of the running speed local time sequence associated implicit feature vectors and the sequence of the heart rate local time sequence associated implicit feature vectors into a node feature propagation network based on a node energy attenuation mechanism to obtain a running speed time sequence propagation aggregate representation vector and a heart rate time sequence propagation aggregate representation vector. It should be appreciated that the variation in the time dimension is a time sequence process for running speed, with temporal continuity and correlation. The local time sequence association implicit characteristic vector of each running speed captures the local mode and trend of speed change in the local time period in the running process, and the time sequence global association relation and characteristic exist between the time sequence characteristics of the running speed in the different local time periods. Therefore, in order to fuse these running speed local time sequence features to form a more comprehensive representation of the running speed time sequence variation, in the technical scheme of the application, the sequence of running speed local time sequence associated implicit feature vectors is further input into a node feature propagation network based on a node energy attenuation mechanism to obtain a running speed time sequence propagation aggregate representation vector. Through the processing of the node characteristic propagation network based on the node energy attenuation mechanism, the node energy attenuation process between the running speed local time sequence characteristics in each local time period can be simulated, and the time sequence space aggregation representation of the node characteristics is realized, so that the understanding capability of the model on the running speed full time domain characteristics and the change modes is improved.
Specifically, the node characteristic propagation network based on the node energy attenuation mechanism uses each running speed local time sequence associated implicit characteristic vector as a node characteristic vector to calculate an energy statistical norm value of a node characteristic vector sequence, wherein the energy statistical norm value is related to the maximum value, the average value and the variance of the node characteristic vector and is used for quantifying the energy level of the node characteristic. Then, by counting propagation space span values among different node feature vectors, the topological structure and the space relation among the nodes, namely node association features among the implicit feature vectors of each running speed local time sequence association, are captured. These span values are combined with the node energy statistical paradigm values of the historical running speed local timing to determine the node energy propagation attenuation coefficient of the running speed local timing, which is inversely related to the historical node energy of the running speed local timing, simulating the attenuation characteristics of each running speed local timing node energy over time and space. Further, using these running speed local timing decay coefficients as weights (i.e., taking into account energy decay effects), a weighted sum of the sequence of node feature vectors is further calculated based on the energy statistical paradigm value of the running speed local timing to generate a historical node energy decay timing aggregate feature vector of the running speed local timing to aggregate running speed local timing historical node feature information. Finally, by fusing the energy statistical paradigm value of the current node of the running speed local time sequence, the weighted sum of the historical running speed local time sequence aggregate characteristic vector and the current running speed local time sequence node characteristic vector is calculated to generate the node energy attenuation time sequence space aggregate expression vector of the running speed local time sequence, and the information of time sequence dynamic and space structure is integrated. In summary, the node characteristic propagation network based on the node energy attenuation mechanism captures the topological structure among nodes by dynamically evaluating the energy level of the node characteristic of each running speed local time sequence, simulates the energy propagation attenuation, and aggregates the time sequence spatial characteristics. In addition, by dynamically adjusting the node energy weights and the aggregate time sequence spatial characteristics, the model can more accurately identify and process the running speed local time sequence key information in the data, thereby obtaining better performance in the follow-up self-adaptive control task of the running machine.
In an embodiment of the present application, inputting the sequence of running speed local time sequence associated implicit feature vectors and the sequence of heart rate local time sequence implicit associated feature vectors into a node feature propagation network based on a node energy attenuation mechanism to obtain a running speed time sequence propagation aggregate representation vector and a heart rate time sequence propagation aggregate representation vector, comprising: calculating node energy statistical pattern values of the running speed local time sequence associated implicit feature vectors based on the maximum value, the average value and the variance of the running speed local time sequence associated implicit feature vectors in the sequence of the running speed local time sequence associated implicit feature vectors to obtain a sequence of running speed local time sequence node energy statistical pattern values, wherein the running speed local time sequence node energy statistical pattern values corresponding to the current running speed local time sequence associated implicit feature vectors in the sequence of running speed local time sequence node energy statistical pattern values are used as current node energy statistical pattern values, and other running speed local time sequence node energy statistical pattern values are used as historical node energy statistical pattern values to obtain a sequence of current running speed local time sequence node energy statistical pattern values and historical running speed local time sequence node energy statistical pattern values; counting node propagation space span values between each other running speed local time sequence correlation implicit feature vector in the sequence of running speed local time sequence correlation implicit feature vectors and the current running speed local time sequence correlation implicit feature vector to obtain a sequence of running speed local time sequence node propagation space span values; determining node energy propagation attenuation coefficient values of other each running speed local time sequence associated implicit feature vector in the sequence of running speed local time sequence associated implicit feature vectors based on the sequence of running speed local time sequence node propagation space span values and the sequence of historical running speed local time sequence node energy statistical paradigm values to obtain a sequence of running speed local time sequence node energy propagation attenuation coefficient values; taking the sequence of the running speed local time sequence node energy transmission attenuation coefficient values as a weight sequence, and calculating the weighted sum among all other running speed local time sequence associated hidden characteristic vectors in the sequence of the running speed local time sequence associated hidden characteristic vectors so as to obtain a historical running speed local time sequence node energy attenuation time sequence aggregation characteristic vector; calculating a weighted sum of the historical running speed local timing node energy decay timing aggregate feature vector and the current running speed local timing correlation implicit feature vector based on the current running speed local timing node energy statistical paradigm value to obtain the running speed timing propagation aggregate representation vector.
Wherein calculating the node energy statistical pattern values for each running speed local timing related implicit feature vector based on the maximum, average, and variance of each running speed local timing related implicit feature vector in the sequence of running speed local timing related implicit feature vectors to obtain the sequence of running speed local timing node energy statistical pattern values comprises: calculating the maximum value, the average value and the variance of the running speed local time sequence associated implicit characteristic vector to obtain a running speed local time sequence maximum value, a running speed local time sequence average value and a running speed local time sequence variance; calculating the sum of the running speed local time sequence average value and the running speed local time sequence variance, and multiplying the sum by a constant 4 to obtain a first running speed local time sequence node energy statistical factor; calculating the product between the running speed local time sequence average value and the running speed local time sequence variance and a constant 2 to obtain a double-modulation running speed local time sequence average value and a double-modulation running speed local time sequence variance; after calculating the square of the difference between the running speed local time sequence maximum value and the running speed local time sequence average value, adding the running speed local time sequence maximum value, the twice-modulated running speed local time sequence average value and the twice-modulated running speed local time sequence variance to obtain a second running speed local time sequence node energy statistical factor; a division between the first running speed local timing node energy statistics factor and the second running speed local timing node energy statistics factor is calculated to obtain a running speed local timing node energy statistic Fan Shizhi. And determining node energy propagation attenuation coefficient values for other respective ones of the sequence of running speed local timing related implicit feature vectors based on the sequence of running speed local timing node propagation spatial span values and the sequence of historical running speed local timing node energy statistical paradigm values to obtain a sequence of running speed local timing node energy propagation attenuation coefficient values comprises: calculating an exponential function value based on a natural constant e by taking each running speed local time sequence node propagation space span value in the sequence of running speed local time sequence node propagation space span values as an exponential power so as to obtain a sequence of running speed local time sequence node propagation class support space span values; calculating the position-wise summation of the sequence of running speed local time sequence node propagation type supporting space span values and the sequence of running speed local time sequence node propagation space span values to obtain a sequence of running speed local time sequence node propagation space span modulation coefficients; a position division between the sequence of historical running speed local timing node energy statistical paradigm values and the sequence of running speed local timing node propagation spatial span modulation coefficients is calculated to obtain a sequence of running speed local timing node energy propagation attenuation coefficient values.
To sum up, in the above embodiment, inputting the sequence of the running speed local time sequence associated implicit feature vector and the sequence of the heart rate local time sequence associated implicit feature vector into a node feature propagation network based on a node energy attenuation mechanism to obtain a running speed time sequence propagation aggregate representation vector and a heart rate time sequence propagation aggregate representation vector, including: inputting the sequence of running speed local time sequence associated implicit characteristic vectors into the node characteristic propagation network based on the node energy attenuation mechanism to process according to the following node characteristic propagation formula so as to obtain the running speed time sequence propagation aggregate expression vector; the node characteristic propagation formula is as follows:
X＝{x1,x2,…xk-1,xk}
Wherein X is the sequence of running speed local time sequence associated implicit feature vectors, X i is the ith running speed local time sequence associated implicit feature vector in the sequence of running speed local time sequence associated implicit feature vectors, For the ith position feature value in the ith running speed local time sequence association hidden feature vector, L is the length of the ith running speed local time sequence association hidden feature vector, mu i is the mean value of the ith running speed local time sequence association hidden feature vector, sigma 2 i is the variance of the ith running speed local time sequence association hidden feature vector, epsilon is a regular term super-parameter, max (x i) represents the maximum value in the ith running speed local time sequence association hidden feature vector,For the node energy statistical range value of the i-th running speed local time sequence associated implicit characteristic vector, count (·) represents a node propagation space span value, alpha, beta, gamma and delta are trainable super-parameters, and x k+1 is the running speed time sequence propagation aggregate representation vector.
Likewise, the sequence of heart rate local time sequence implicit correlation feature vectors is input into a node feature propagation network based on a node energy attenuation mechanism to obtain heart rate time sequence propagation aggregate representation vectors, and feature propagation aggregate representation information among the heart rate local time sequence features is captured in a dependent manner.
In particular, the S4 performs a feature interaction response process on the running speed time series propagation aggregate representation vector and the heart rate time series propagation aggregate representation vector to obtain a significant running speed-heart rate time series interaction fusion representation vector. Since running speed and heart rate are two important physiological indicators, including exercise intensity and cardiovascular system response, respectively, they can reflect the exercise intensity, endurance level, and physical condition of the user. Therefore, in order to provide more comprehensive exercise state information and physiological state information of the user so as to adaptively adjust the gradient of the treadmill according to the personalized needs and capabilities of the user, in the technical scheme of the application, the running speed time sequence transmission aggregate expression vector and the heart rate time sequence transmission aggregate expression vector are further input into a characteristic interaction response module based on an adaptive distinguishable mechanism so as to obtain a significant running speed-heart rate time sequence interaction fusion expression vector. Through the processing of the characteristic interaction response module based on the self-adaptive distinguishable mechanism, the expression capability of the characteristics and the overall performance of the model can be obviously improved through the fine analysis and the self-adaptive weight adjustment of each position. This may enable a characteristic interaction between running speed and heart rate to discover potential timing associations and effects between the two. This helps reveal the dynamic relationship between running speed and heart rate, providing a more thorough and careful data analysis basis for intelligent control systems.
Specifically, firstly, by calculating the position-by-position response between the running speed time sequence propagation aggregation expression vector and the heart rate time sequence propagation aggregation expression vector, a running speed-heart rate time sequence position-by-position response characteristic vector is generated, and the step effectively captures the local time sequence characteristic interaction between running speed time sequence aggregation information and heart rate time sequence aggregation information. And then, normalizing the running speed-heart rate time sequence position-by-position responses by using a Softmax function to form probability distribution, so that the standardization of characteristic responses is ensured, and preparation is provided for subsequent gating function weight screening. The normalization process not only balances the scale of the characteristic response, but also enables the expression of the running speed-heart rate time sequence characteristics in the model to be more balanced. By inputting the normalized running speed-heart rate time sequence position-by-position response characteristic vector into a learnable gating function, the model can adaptively learn and output the running speed-heart rate time sequence response screening weight mask vector, and the dynamic weighting process realizes time sequence characteristic selection and reinforcement between the running speed and the heart rate through a gating mechanism and dynamically adjusts the importance of the characteristic position. This step is the core to achieve feature optimization because it directly affects the sensitivity and discrimination ability of the model to the running speed-heart rate timing feature response. Further, by calculating the position-wise point multiplication between the running speed-heart rate time sequence response screening weight mask vector and the normalized running speed-heart rate time sequence position-by-position response characteristic vector, the running speed-heart rate time sequence position-by-position response distinguishable weight mask vector is obtained, weight distribution is further refined, and accurate weight control is provided for final running speed-heart rate time sequence interactive fusion significance characterization and optimization. This refinement of the position-by-position weights ensures that the model can assign the most appropriate weight for each position of the running speed-heart rate timing responsive feature, thereby capturing useful information in the data more effectively. Finally, a salient running speed-heart rate timing interactive fusion representation vector is generated by dot multiplying the running speed-heart rate timing position-by-position response distinguishable weight mask vector with the running speed-heart rate timing position-by-position response feature vector. The step integrates the characteristic response after weight adjustment, generates the final running speed-heart rate time sequence interactive fusion representation necklace after the saliency expression reinforcement, remarkably improves the capturing capability of the model on key running speed-heart rate time sequence interactive response information, and provides a basis for the gradient self-adaptive control of the follow-up running machine.
To sum up, in the above embodiment, inputting the running speed timing propagation aggregate representation vector and the heart rate timing propagation aggregate representation vector into a feature interaction response module based on an adaptive distinguishable mechanism to obtain the significant running speed-heart rate timing interaction fusion representation vector includes: inputting the running speed time sequence transmission aggregation expression vector and the heart rate time sequence transmission aggregation expression vector into the characteristic interaction response module based on the self-adaptive distinguishable mechanism to process according to the following characteristic interaction response formula so as to obtain the obvious running speed-heart rate time sequence interaction fusion expression vector; the characteristic interaction response formula is as follows:
vr＝xk+1/v1
vn＝softmax(vr)
vs＝vn⊙va
Wherein v 1 and x k+1 represent the heart rate timing propagation aggregate representation vector and the running speed timing propagation aggregate representation vector, respectively, v r is a running speed-heart rate timing position-by-position response feature vector, softmax (·) represents a softmax function, v n is a normalized running speed-heart rate timing position-by-position response feature vector, exp (·) is a natural exponential function value, v a is a running speed-heart rate timing response screening weight mask vector, as-by-position point multiplication, v s is a running speed-heart rate timing position-by-position response distinguishable weight mask vector, A representation vector is fused for the salient running speed-heart rate timing interactions.
In particular, the S5 generates control instructions for indicating increasing the treadmill grade value, decreasing the treadmill grade value, and maintaining the treadmill grade value based on the significant running speed-heart rate timing interaction fusion representation vector. In particular, in one specific example of the present application, the significant running speed-heart rate timing interaction fusion representation vector is input to a classifier-based grade controller for control instructions for representing increasing the treadmill grade value, decreasing the treadmill grade value, and maintaining the treadmill grade value. That is, the running speed and heart rate time sequence interaction fusion characteristics after the significance expression optimization are utilized to conduct classification processing, so that the gradient of the running machine is adaptively controlled. Therefore, more intelligent exercise experience can be provided for the user according to the user's ability, so that personalized requirements of the user can be met, safety accidents caused by excessive exercise of the user can be avoided, and the safety of the running machine is improved.
It should be appreciated that training of the 1D-CNN model-based sequence encoder, the node characteristic propagation network based on the node energy attenuation mechanism, the adaptively distinguishable mechanism-based characteristic interaction response module, and the classifier-based grade controller is required before the inference is made using the neural network model described above. That is, the control method for the intelligent control circuit board of the running machine further comprises a training stage for training the 1D-CNN model-based sequence encoder, the node characteristic propagation network based on the node energy attenuation mechanism, the characteristic interaction response module based on the adaptive distinguishable mechanism and the gradient controller based on the classifier.
Fig. 3 is a flowchart of a training phase of a control method of an intelligent control circuit board for a treadmill according to an embodiment of the present application. As shown in fig. 3, a control method of an intelligent control circuit board for a treadmill according to an embodiment of the present application includes: a training phase comprising: s110, training data is acquired, wherein the training data comprises a time queue of training running speeds and a time queue of training heart rate values, which are acquired by a sensor group; s120, inputting the time queue of the training running speed and the time queue of the training heart rate value into a sequence encoder based on a 1D-CNN model to obtain a sequence of training running speed local time sequence associated implicit characteristic vectors and a sequence of training heart rate local time sequence associated implicit characteristic vectors; s130, inputting the sequence of the training running speed local time sequence associated implicit characteristic vector and the sequence of the training heart rate local time sequence implicit associated characteristic vector into a node characteristic propagation network based on a node energy attenuation mechanism to obtain a training running speed time sequence propagation aggregate expression vector and a training heart rate time sequence propagation aggregate expression vector; s140, inputting the training running speed time sequence transmission aggregation expression vector and the training heart rate time sequence transmission aggregation expression vector into a characteristic interaction response module based on an adaptive distinguishable mechanism to obtain a training obvious running speed-heart rate time sequence interaction fusion expression vector; s150, enabling the training significant running speed-heart rate time sequence interaction fusion expression vector to pass through a gradient controller based on a classifier so as to obtain a classification loss function value; s160, calculating a significant running speed-heart rate time sequence interactive fusion loss function value based on a significant running speed-heart rate time sequence interactive fusion representation loss item; s170, calculating a weighted sum of the classified loss function value and the significant running speed-heart rate time sequence interaction fusion loss function value to obtain a final loss function value; s180, training the 1D-CNN model-based sequence encoder, the node characteristic propagation network based on the node energy attenuation mechanism, the characteristic interaction response module based on the self-adaptive distinguishable mechanism and the gradient controller based on the classifier based on the final loss function value.
In particular, it is preferred that a new loss function value is further introduced outside the classification loss function value, wherein the construction of the new loss function value comprises the steps of:
Calculating a training salient running speed-heart rate time sequence interaction fusion sum matrix and a training salient running speed-heart rate time sequence interaction fusion difference matrix based on the training salient running speed-heart rate time sequence interaction fusion representation vector, wherein the characteristic value of the (i, j) th position of the training salient running speed-heart rate time sequence interaction fusion sum matrix is the mean value of the i th characteristic value and the j th characteristic value of the training salient running speed-heart rate time sequence interaction fusion representation vector, and the characteristic value of the (i, j) th position of the training salient running speed-heart rate time sequence interaction fusion difference matrix is one half of the difference absolute value of the i th characteristic value and the j th characteristic value of the training salient running speed-heart rate time sequence interaction fusion representation vector;
The training significant running speed-heart rate time sequence interactive fusion expression vector is multiplied by the training significant running speed-heart rate time sequence interactive fusion sum matrix and the training significant running speed-heart rate time sequence interactive fusion difference matrix respectively to obtain a training significant running speed-heart rate time sequence interactive fusion query sum vector and a training significant running speed-heart rate time sequence interactive fusion query difference vector;
Calculating a vector inner product of the training salient running speed-heart rate time sequence interaction fusion representation and the vector and the training salient running speed-heart rate time sequence interaction fusion query difference vector to obtain a first salient running speed-heart rate time sequence interaction fusion representation loss term;
Performing matrix multiplication on the training significant running speed-heart rate time sequence interaction fusion sum matrix and the training significant running speed-heart rate time sequence interaction fusion difference matrix, and calculating the Frobenius norm of the product matrix to obtain a second significant running speed-heart rate time sequence interaction fusion representation loss term; and
Subtracting a product of a predetermined weight super parameter and the second significant running speed-heart rate timing cross fusion representation loss term from the first significant running speed-heart rate timing cross fusion representation loss term to obtain the new loss function value.
Wherein the new loss function value, for example, referred to as a significant running speed-heart rate timing interaction fusion loss function value, is specifically expressed as:
Wherein the method comprises the steps of
M1(i,j)＝vi+vj/2
M2(i,j)＝|vi-vj|/2
Wherein V is the training significant running speed-heart rate time sequence interactive fusion expression vector, M μ and M σ are the training significant running speed-heart rate time sequence interactive fusion sum matrix and the training significant running speed-heart rate time sequence interactive fusion difference matrix respectively, M μ (i, j) and M σ (i, j) are the training significant running speed-heart rate time sequence interactive fusion sum matrix and the characteristic value of the (i, j) th position of the training significant running speed-heart rate time sequence interactive fusion difference matrix respectively, V i and V j are the i characteristic value and the j characteristic value of the training significant running speed-heart rate time sequence interactive fusion expression vector respectively,For matrix multiplication, |·| F is the Frobenius norm of the computation matrix, a is a predetermined weight hyper-parameter,The loss function values are fused for significant running speed-heart rate timing interactions.
Here, considering that the training running speed time series propagation aggregate representative vector and the training heart rate time series propagation aggregate representative vector represent time series node energy decay propagation characteristics of one-dimensional local time series correlation characteristics of running speed and heart rate values respectively, feature interaction responses based on an adaptive distinguishable mechanism are performed on the training running speed time series propagation aggregate representative vector and the training heart rate time series propagation aggregate representative vector, and the training significant running speed-heart rate time series interaction fusion representative vector also has classification regression recognition difficulties due to the self-adaptive distinguishable interaction response weight differences caused by time series characteristic distribution differences of different data, thereby influencing the accuracy of classification results.
The applicant performs query composition of detail inner product space in the training significant running speed-heart rate time sequence interactive fusion expression vector through the short-distance cross-scale detail linked structural feature expression of the training significant running speed-heart rate time sequence interactive fusion expression vector, approximates the low-rank independent observable composition of the link detail composition provided by the structural detail interaction of the training significant running speed-heart rate time sequence interactive fusion expression vector, so that the detail group decomposition can be performed on the basis of detail complexity through the distributed detail group of the training significant running speed-heart rate time sequence interactive fusion expression vector by training through the significant running speed-heart rate time sequence interactive fusion loss function, thereby promoting classification regression decomposition identification of the training significant running speed-heart rate time sequence interactive fusion expression vector based on the complex feature expression and improving the accuracy of control instructions obtained by a gradient controller based on a classifier. Like this, can carry out the self-adaptation steerable of treadmill more accurately to improve the intelligent degree of treadmill control, provide more intelligent motion experience for the user, in order to satisfy user's individualized demand, can avoid the user to move excessively and lead to the incident simultaneously, increased the security of treadmill.
In summary, the control method of the intelligent control circuit board for the running machine according to the embodiment of the application is explained, wherein the running speed of the running machine and the heart rate value of an athlete are monitored and collected in real time through the sensor group, and the data are input into the intelligent control circuit board, so that the running speed time sequence data and the heart rate time sequence data are analyzed and interacted in the intelligent control circuit board by utilizing the data processing and analysis algorithm based on artificial intelligence and deep learning, and the time sequence correlation relation and the interaction influence between the running speed and the heart rate of the athlete are learned and captured, so that the gradient of the running machine is adaptively controlled by utilizing the interaction fusion semantics of the two. Therefore, more intelligent exercise experience can be provided for the user according to the user's ability, so that personalized requirements of the user can be met, safety accidents caused by excessive exercise of the user can be avoided, and the safety of the running machine is improved.
The foregoing description of the embodiments of the present disclosure has been presented for purposes of illustration and description, and is not intended to be exhaustive or limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the various embodiments described. The terminology used herein was chosen in order to best explain the principles of the embodiments, the practical application, or the improvement of technology in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein.
Claims (10)
1. A control method of an intelligent control circuit board for a running machine is characterized by comprising the following steps:
acquiring a time queue of running speeds and a time queue of heart rate values acquired by a sensor group;
performing time sequence coding on the time sequence of the running speed and the time sequence of the heart rate value to obtain a sequence of running speed local time sequence associated implicit characteristic vectors and a sequence of heart rate local time sequence associated implicit characteristic vectors;
Inputting the sequence of the running speed local time sequence associated implicit characteristic vector and the sequence of the heart rate local time sequence associated implicit characteristic vector into a node characteristic propagation network based on a node energy attenuation mechanism to obtain a running speed time sequence propagation aggregate expression vector and a heart rate time sequence propagation aggregate expression vector;
performing characteristic interaction response processing on the running speed time sequence propagation aggregation expression vector and the heart rate time sequence propagation aggregation expression vector to obtain a remarkable running speed-heart rate time sequence interaction fusion expression vector;
Based on the significant running speed-heart rate timing interaction fusion representation vector, control instructions are generated for representing increasing the treadmill grade value, decreasing the treadmill grade value, and maintaining the treadmill grade value.
2. The method of claim 1, wherein the time series of running speeds and the time series of heart rate values are time-series encoded to obtain a sequence of running speed local time series implicit characteristic vectors and a sequence of heart rate local time series implicit characteristic vectors, comprising: inputting the time queue of the running speed and the time queue of the heart rate value into a sequence encoder based on a 1D-CNN model to obtain a sequence of the running speed local time sequence associated implicit characteristic vector and a sequence of the heart rate local time sequence implicit associated characteristic vector.
3. The control method for an intelligent control circuit board of a treadmill of claim 2, wherein inputting the sequence of running speed local time sequence implicit characteristic vectors and the sequence of heart rate local time sequence implicit characteristic vectors into a node characteristic propagation network based on a node energy decay mechanism to obtain a running speed time sequence propagation aggregate representative vector and a heart rate time sequence propagation aggregate representative vector comprises:
Calculating node energy statistical pattern values of the running speed local time sequence associated implicit feature vectors based on the maximum value, the average value and the variance of the running speed local time sequence associated implicit feature vectors in the sequence of the running speed local time sequence associated implicit feature vectors to obtain a sequence of running speed local time sequence node energy statistical pattern values, wherein the running speed local time sequence node energy statistical pattern values corresponding to the current running speed local time sequence associated implicit feature vectors in the sequence of running speed local time sequence node energy statistical pattern values are used as current node energy statistical pattern values, and other running speed local time sequence node energy statistical pattern values are used as historical node energy statistical pattern values to obtain a sequence of current running speed local time sequence node energy statistical pattern values and historical running speed local time sequence node energy statistical pattern values;
Counting node propagation space span values between each other running speed local time sequence correlation implicit feature vector in the sequence of running speed local time sequence correlation implicit feature vectors and the current running speed local time sequence correlation implicit feature vector to obtain a sequence of running speed local time sequence node propagation space span values;
Determining node energy propagation attenuation coefficient values of other each running speed local time sequence associated implicit feature vector in the sequence of running speed local time sequence associated implicit feature vectors based on the sequence of running speed local time sequence node propagation space span values and the sequence of historical running speed local time sequence node energy statistical paradigm values to obtain a sequence of running speed local time sequence node energy propagation attenuation coefficient values;
Taking the sequence of the running speed local time sequence node energy transmission attenuation coefficient values as a weight sequence, and calculating the weighted sum among all other running speed local time sequence associated hidden characteristic vectors in the sequence of the running speed local time sequence associated hidden characteristic vectors so as to obtain a historical running speed local time sequence node energy attenuation time sequence aggregation characteristic vector;
Calculating a weighted sum of the historical running speed local timing node energy decay timing aggregate feature vector and the current running speed local timing correlation implicit feature vector based on the current running speed local timing node energy statistical paradigm value to obtain the running speed timing propagation aggregate representation vector.
4. The control method for an intelligent control circuit board for a treadmill of claim 3, wherein calculating the node energy statistical pattern values for each running speed local timing related implicit feature vector based on the maximum, average, and variance of each running speed local timing related implicit feature vector in the sequence of running speed local timing related implicit feature vectors to obtain the sequence of running speed local timing node energy statistical pattern values comprises:
calculating the maximum value, the average value and the variance of the running speed local time sequence associated implicit characteristic vector to obtain a running speed local time sequence maximum value, a running speed local time sequence average value and a running speed local time sequence variance;
calculating the sum of the running speed local time sequence average value and the running speed local time sequence variance, and multiplying the sum by a constant 4 to obtain a first running speed local time sequence node energy statistical factor;
calculating the product between the running speed local time sequence average value and the running speed local time sequence variance and a constant 2 to obtain a double-modulation running speed local time sequence average value and a double-modulation running speed local time sequence variance;
After calculating the square of the difference between the running speed local time sequence maximum value and the running speed local time sequence average value, adding the running speed local time sequence maximum value, the twice-modulated running speed local time sequence average value and the twice-modulated running speed local time sequence variance to obtain a second running speed local time sequence node energy statistical factor;
A division between the first running speed local timing node energy statistics factor and the second running speed local timing node energy statistics factor is calculated to obtain a running speed local timing node energy statistic Fan Shizhi.
5. The method of claim 4, wherein determining the node energy propagation attenuation coefficient values for other respective ones of the sequence of running speed local timing related implicit feature vectors to obtain the sequence of running speed local timing node energy propagation attenuation coefficient values based on the sequence of running speed local timing node propagation spatial span values and the sequence of historical running speed local timing node energy statistical paradigm values comprises:
Calculating an exponential function value based on a natural constant e by taking each running speed local time sequence node propagation space span value in the sequence of running speed local time sequence node propagation space span values as an exponential power so as to obtain a sequence of running speed local time sequence node propagation class support space span values;
Calculating the position-wise summation of the sequence of running speed local time sequence node propagation type supporting space span values and the sequence of running speed local time sequence node propagation space span values to obtain a sequence of running speed local time sequence node propagation space span modulation coefficients;
A position division between the sequence of historical running speed local timing node energy statistical paradigm values and the sequence of running speed local timing node propagation spatial span modulation coefficients is calculated to obtain a sequence of running speed local timing node energy propagation attenuation coefficient values.
6. The method of claim 5, wherein performing a characteristic interactive response process on the running speed time series propagation aggregate representation vector and the heart rate time series propagation aggregate representation vector to obtain a significant running speed-heart rate time series interactive fusion representation vector comprises: inputting the running speed time sequence propagation aggregate representation vector and the heart rate time sequence propagation aggregate representation vector into a characteristic interaction response module based on an adaptive distinguishable mechanism to obtain the obvious running speed-heart rate time sequence interaction fusion representation vector.
7. The method of claim 6, wherein inputting the running speed time series propagation aggregate representation vector and the heart rate time series propagation aggregate representation vector into a characteristic interaction response module based on an adaptive distinguishable mechanism to obtain the salient running speed-heart rate time series interaction fusion representation vector comprises:
calculating a position-by-position response between the heart rate timing propagation aggregate representative vector and the running speed timing propagation aggregate representative vector to obtain a running speed-heart rate timing position-by-position response feature vector;
Normalizing the running speed-heart rate time sequence position-by-position response characteristic vector by using a Softmax function to obtain a normalized running speed-heart rate time sequence position-by-position response characteristic vector;
taking the negative number of each position characteristic value in the normalized running speed-heart rate time sequence position-by-position response characteristic vector as a power, and calculating a natural exponential function value based on a natural constant e to obtain a normalized running speed-heart rate time sequence position-by-position response type support characteristic vector;
calculating the reciprocal of the sum of each position characteristic value and a constant I in the normalized running speed-heart rate time sequence position-by-position response type support characteristic vector to obtain a running speed-heart rate time sequence response screening weight mask vector;
Calculating the position-wise point multiplication between the running speed-heart rate time sequence response screening weight mask vector and the normalized running speed-heart rate time sequence position-by-position response feature vector to obtain a running speed-heart rate time sequence position-by-position response distinguishable weight mask vector;
Calculating a per-position point multiplication between the running speed-heart rate time sequence position-by-position response distinguishable weight mask vector and the running speed-heart rate time sequence position-by-position response feature vector results in a significant running speed-heart rate time sequence interactive fusion representation vector.
8. The control method of the intelligent control circuit board for the running machine according to claim 7, wherein the generating control instructions for indicating increasing the running machine gradient value, decreasing the running machine gradient value, and maintaining the running machine gradient value based on the significant running speed-heart rate timing interaction fusion representation vector includes: the significant running speed-heart rate timing interactive fusion representation vector is input to a classifier-based grade controller to derive control instructions for representing increasing the treadmill grade value, decreasing the treadmill grade value, and maintaining the treadmill grade value.
9. The control method of an intelligent control circuit board for a treadmill of claim 8, further comprising the training step of: training the 1D-CNN model-based sequence encoder, the node characteristic propagation network based on the node energy attenuation mechanism, the characteristic interaction response module based on the self-adaptive distinguishable mechanism and the gradient controller based on the classifier;
Wherein, the training steps are:
acquiring training data, wherein the training data comprises a time queue of training running speeds and a time queue of training heart rate values, which are acquired by a sensor group;
Inputting the time queue of the training running speed and the time queue of the training heart rate value into a sequence encoder based on a 1D-CNN model to obtain a sequence of training running speed local time sequence associated implicit characteristic vectors and a sequence of training heart rate local time sequence associated implicit characteristic vectors;
inputting the sequence of the training running speed local time sequence associated implicit characteristic vector and the sequence of the training heart rate local time sequence associated implicit characteristic vector into a node characteristic propagation network based on a node energy attenuation mechanism to obtain a training running speed time sequence propagation aggregate expression vector and a training heart rate time sequence propagation aggregate expression vector;
inputting the training running speed time sequence transmission aggregation expression vector and the training heart rate time sequence transmission aggregation expression vector into a characteristic interaction response module based on an adaptive distinguishable mechanism to obtain a training significant running speed-heart rate time sequence interaction fusion expression vector;
The training significant running speed-heart rate time sequence interactive fusion expression vector passes through a gradient controller based on a classifier to obtain a classification loss function value;
Calculating a significant running speed-heart rate time sequence interactive fusion loss function value based on the significant running speed-heart rate time sequence interactive fusion representation loss item;
Calculating a weighted sum of the classification loss function value and the significant running speed-heart rate timing interaction fusion loss function value to obtain a final loss function value;
training the 1D-CNN model-based sequence encoder, the node characteristic propagation network based on the node energy attenuation mechanism, the characteristic interaction response module based on the self-adaptive distinguishable mechanism and the gradient controller based on the classifier based on the final loss function value.
10. An intelligent control circuit board for a running machine, which is controlled by the control method of the intelligent control circuit board for a running machine according to any one of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411000354.6A CN118949364B (en) | 2024-07-24 | 2024-07-24 | Intelligent control circuit board and control method for treadmill |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411000354.6A CN118949364B (en) | 2024-07-24 | 2024-07-24 | Intelligent control circuit board and control method for treadmill |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118949364A true CN118949364A (en) | 2024-11-15 |
| CN118949364B CN118949364B (en) | 2025-04-01 |
Family
ID=93389150
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202411000354.6A Active CN118949364B (en) | 2024-07-24 | 2024-07-24 | Intelligent control circuit board and control method for treadmill |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118949364B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6033344A (en) * | 1994-02-04 | 2000-03-07 | True Fitness Technology, Inc. | Fitness apparatus with heart rate control system and method of operation |
| CN101708360A (en) * | 2009-12-17 | 2010-05-19 | 青岛英派斯（集团）有限公司 | Method for automatically controlling speed and gradient of electric treadmill according to heart rate of body-builder |
| CN103892814A (en) * | 2014-03-31 | 2014-07-02 | 无锡首康科技有限公司 | Treadmill control method for exercise rehabilitation therapy system |
| CN106139516A (en) * | 2015-03-27 | 2016-11-23 | 耐森雷特电子科技(上海)有限公司 | Automatically control the method for treadmill exercise intensity according to heart rate |
| JP6725731B1 (en) * | 2019-06-24 | 2020-07-22 | 株式会社Ｏｆｆｉｃｅ Ｙａｇｉ | Training support method and training support system |
-
2024
- 2024-07-24 CN CN202411000354.6A patent/CN118949364B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6033344A (en) * | 1994-02-04 | 2000-03-07 | True Fitness Technology, Inc. | Fitness apparatus with heart rate control system and method of operation |
| CN101708360A (en) * | 2009-12-17 | 2010-05-19 | 青岛英派斯（集团）有限公司 | Method for automatically controlling speed and gradient of electric treadmill according to heart rate of body-builder |
| CN103892814A (en) * | 2014-03-31 | 2014-07-02 | 无锡首康科技有限公司 | Treadmill control method for exercise rehabilitation therapy system |
| CN106139516A (en) * | 2015-03-27 | 2016-11-23 | 耐森雷特电子科技(上海)有限公司 | Automatically control the method for treadmill exercise intensity according to heart rate |
| JP6725731B1 (en) * | 2019-06-24 | 2020-07-22 | 株式会社Ｏｆｆｉｃｅ Ｙａｇｉ | Training support method and training support system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118949364B (en) | 2025-04-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112508044A (en) | Artificial intelligence AI model evaluation method, system and equipment | |
| CN114095381B (en) | Multitask model training method, multitask prediction method and related products | |
| CN102213957A (en) | Control method, and device and system for providing virtual private sport coach | |
| CN117522479B (en) | Accurate Internet advertisement delivery method and system | |
| CN116668198B (en) | Flow playback test method, device, equipment and medium based on deep learning | |
| CN113701825A (en) | Body-building facility abnormity detection method and system based on artificial intelligence | |
| CN117273769A (en) | Clothing data collection and analysis method and system | |
| CN118375997B (en) | Indoor environment monitoring system and method based on intelligent sensing technology | |
| CN117312941A (en) | Rolling bearing fault diagnosis method and device under sample unbalance condition | |
| CN117331803A (en) | Method and system for supporting multi-chip server resource management | |
| CN119044806A (en) | Method and device for predicting remaining life of lithium battery | |
| CN119007169B (en) | Cerebral apoplexy patient driving ability analysis system based on eye movement tracking technology | |
| CN119002664B (en) | Management system and method for intelligent computing platform | |
| Zhou et al. | Hybrid neural network-based exploration on the influence of continuous sensor data for the balancing ability of aerobics students | |
| CN120086993A (en) | Life prediction method and system of main shaft based on application analysis | |
| CN118949364B (en) | Intelligent control circuit board and control method for treadmill | |
| JP2020140226A (en) | General artificial intelligence apparatus and general artificial intelligence program | |
| CN119312198A (en) | Monitoring method, device, electronic equipment and medium based on digital twin model | |
| CN117648572A (en) | Cognitive assessment method and system based on virtual reality and ensemble learning | |
| CN117115596A (en) | Training method, device, equipment and medium of object action classification model | |
| US20220040532A1 (en) | Utilizing machine learning and cognitive state analysis to track user performance | |
| CN117350366A (en) | Network model construction method and related equipment | |
| CN113987261B (en) | A video recommendation method and system based on dynamic trust perception | |
| Yuan et al. | RETRACTED ARTICLE: Application of motion recognition algorithm based on optical detection iot devices in college student motion monitoring platform | |
| CN114356730A (en) | A simulator for training reinforcement learning models in the recommendation domain |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |