CN110897425A - Child sitting posture detection cushion, device interacting with cushion and sitting posture monitoring method - Google Patents
Child sitting posture detection cushion, device interacting with cushion and sitting posture monitoring method Download PDFInfo
- Publication number
- CN110897425A CN110897425A CN201911125903.1A CN201911125903A CN110897425A CN 110897425 A CN110897425 A CN 110897425A CN 201911125903 A CN201911125903 A CN 201911125903A CN 110897425 A CN110897425 A CN 110897425A
- Authority
- CN
- China
- Prior art keywords
- sitting posture
- data
- cushion
- module
- control chip
- 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.)
- Pending
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A47—FURNITURE; DOMESTIC ARTICLES OR APPLIANCES; COFFEE MILLS; SPICE MILLS; SUCTION CLEANERS IN GENERAL
- A47C—CHAIRS; SOFAS; BEDS
- A47C31/00—Details or accessories for chairs, beds, or the like, not provided for in other groups of this subclass, e.g. upholstery fasteners, mattress protectors, stretching devices for mattress nets
- A47C31/12—Means, e.g. measuring means for adapting chairs, beds or mattresses to the shape or weight of persons
- A47C31/126—Means, e.g. measuring means for adapting chairs, beds or mattresses to the shape or weight of persons for chairs
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
- A61B5/1036—Measuring load distribution, e.g. podologic studies
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/45—For evaluating or diagnosing the musculoskeletal system or teeth
- A61B5/4538—Evaluating a particular part of the muscoloskeletal system or a particular medical condition
- A61B5/4561—Evaluating static posture, e.g. undesirable back curvature
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Surgery (AREA)
- Biomedical Technology (AREA)
- Veterinary Medicine (AREA)
- Public Health (AREA)
- Physics & Mathematics (AREA)
- Dentistry (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Engineering & Computer Science (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Physical Education & Sports Medicine (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Rheumatology (AREA)
- Orthopedic Medicine & Surgery (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The invention discloses a child sitting posture detection cushion, a device interacting with the cushion and a sitting posture monitoring method.A pressure sensor is uniformly arranged in the cushion and used for monitoring the data of the pressure distribution of a human body sitting posture; after the pressure distribution map is obtained through processing, modeling is carried out on the pressure distribution map data through a convolution algorithm, and the sitting posture data of the user is obtained through a modeling model; after the device processes sitting posture data of a user, and the duration time of the detected poor sitting posture exceeds a preset time threshold, the device sends out vibration and/or makes visual and audio prompts to the user; the method is based on the seat cushion and the device, and the sitting posture of the child is detected in real time. The invention utilizes convolution algorithm to process the electronic data in the form of image array, replaces the artificial characteristic building model of the data detected by the sensor, and directly utilizes convolution algorithm to carry out intelligent modeling on the data without human intervention.
Description
Technical Field
The invention relates to the field of intelligent interaction, in particular to a sitting posture monitoring cushion, a sitting posture detection method based on the cushion and an interaction device intelligently interacting with the cushion.
Background
With the progress of society, the learning pressure of children is increased, and the sitting for a long time becomes a normal state. The bad sitting posture causes more and more children to have visual deterioration, humpback, spinal injury and the like in different degrees, and the physical and mental development of the children is seriously influenced.
Common sitting posture detection methods for children mainly comprise wearable type, distance detection type, infrared detection type and binocular camera type in form; common sitting posture prompting or feedback methods include vibration, sound and light, screen display and the like. But the children are not suitable for active activities and psychological states due to the characteristics of inconvenient and quick wearing, easy loss and the like. The existing sitting posture detection and sensor technology is relatively mature, but for children, how to achieve accurate sitting posture monitoring and how to meet the preference and psychological needs of children through an emerging interactive design, and users have a mutual excitation effect when using the sitting posture detection device.
Disclosure of Invention
The invention aims to solve the technical problems that a depth camera is required to be arranged to acquire sitting posture data in the prior art, most of the sitting posture data is suitable for office staff and is not suitable for children to use, and related data is required to be worn on children to acquire, so that the actions of the children are limited, uncomfortable feelings are easy to generate, and a children sitting posture detection cushion is provided and comprises a cushion body, wherein a measurement monitoring module, a plurality of pressure sensors and a data transmission module are arranged in the cushion in an array mode; wherein:
the pressure sensors are uniformly arranged in the cushion in an array form of n rows and m columns and used for monitoring the pressure distribution data of the human body sitting posture;
the measurement monitoring module comprises a control chip and a sensor measuring circuit which are sequentially connected; wherein: one end of the sensor measuring circuit is connected to the pressure sensor array, and the other end of the sensor measuring circuit is connected to the control chip and used for receiving the electric signals which are transmitted by the pressure sensor array and reflect the pressure distribution data of the sitting posture of the human body and further transmitting the electric signals to the control chip;
the control chip is used for modeling the pressure distribution diagram data by utilizing a convolution algorithm after reducing the received pressure array data into a pressure distribution diagram on one hand; on the other hand, aiming at the input human sitting posture pressure distribution data, the sitting posture data of the user is obtained by utilizing a modeling model, wherein automatic marking and parameter adjustment are carried out according to the specific sitting posture conditions of different users in the modeling process, the generated pressure distribution diagram is accurately classified in the sitting posture, and the misjudgment rate of the system is reduced; the obtained user sitting posture data is further transmitted to a device interacting with the cushion through a data transmission module for further processing.
In the invention, different from other scattered data processing methods, the received electronic data is processed by using a convolution algorithm without manually finding a characteristic establishing model for the data detected by a sensor, and the intelligent modeling is directly carried out on the electronic data by using the convolution algorithm without manual intervention; the convolution algorithm is applied to the electronic data, wherein the electronic data can be arranged into an image array for processing, and the advantages of intelligence and matching of deep learning are further embodied.
And (3) constructing a convolutional neural network model, and quickly training data detected by the cushion for analyzing and judging the sitting posture of the child. The deep learning algorithm has a learning function, can automatically label and adjust parameters according to specific sitting posture conditions of different users in the using process, accurately classifies the sitting postures of the generated pressure distribution diagram, and reduces the misjudgment rate of the system.
Furthermore, the measurement monitoring module is connected with the pressure sensor array through a flat cable interface, so that the stability of data transmission is further ensured.
Further, a data acquisition module is arranged in the control chip;
the data acquisition module is used for acquiring pressure distribution map data and establishing a sitting posture data sample library based on the acquired data;
a sitting posture recognition module is also arranged in the control chip; the sitting posture recognition module is used for constructing a sitting posture recognition model based on a convolution algorithm; the data in the sitting posture data sample library is used as a training sample to train the sitting posture recognition module;
the control chip is also provided with a posture data output module which is used for obtaining the sitting posture data of the user aiming at the received pressure array data by utilizing the sitting posture recognition model obtained by training, and the sitting posture of the user is further transmitted to the data transmission module.
Further, based on the convolution algorithm, the process of constructing the sitting posture recognition model comprises the following steps:
s1, collecting sitting posture pressure distribution diagram samples from the sitting posture data sample library, classifying according to different sitting postures of the human body, and further constructing a training set and a testing set;
s2, carrying out forward propagation once by using a training set sample in the currently constructed sitting posture recognition model; after the features are extracted through the convolutional layer, the data dimensionality is reduced through the pooling layer, and the features are extracted again through the full-connection layer, the sitting posture state is output, wherein the network parameters are updated according to the error between the output sitting posture state and the actual sitting posture state; in the current step, repeating iteration until the output sitting posture state is consistent with the actual sitting posture state, and stopping network training;
s3, inputting a test set into the sitting posture recognition model obtained by training in the step S2, and testing; when the test classification accuracy does not reach the preset standard value, repeatedly executing the step S2 and the step S3; when the final data classification accuracy reaches a preset standard value, taking a currently obtained sitting posture recognition model as a basic algorithm model; and an application and attitude data output module.
The method is different from the traditional scattered data processing method, and adopts a convolution algorithm to directly carry out intelligent modeling on data without human intervention, so that the investment of manpower is effectively saved.
The invention provides a device interacting with a cushion, which is used for receiving sitting posture data of a user; the device comprises a data receiving and transmitting module, a main control chip and a video and audio decoding module; wherein:
the data receiving and transmitting module is used for receiving the sitting posture data and further transmitting the sitting posture data to the main control chip;
the main control chip is used for starting a timing mode when the received bad sitting posture information is judged after data is received, and controlling the video and audio decoding module to send out vibration reminding and/or make video and audio reminding after the detected bad sitting posture duration time exceeds a preset time threshold.
The intelligent device interacting with the seat cushion provided by the invention can further effectively remind a child when the situation that the child maintains a bad sitting posture for more than a certain time is monitored.
Furthermore, the device also comprises a display screen, wherein the display screen is connected to the main control chip and used for displaying data in real time.
Furthermore, the platform also comprises an SD card storage module which is respectively connected with the video and audio decoding module and the main control chip and is used for storing the video and audio data output by the video and audio decoding module.
In the invention, the SD card storage module is used as a resource library for storing video and audio. The main control module selects and plays related video and audio by accessing data of the SD card.
Furthermore, the main control chip is connected to the cloud server side, and data processed by the main control chip is further uploaded to the cloud server side for storage.
The invention provides a sitting posture monitoring method, which is further realized based on the child sitting posture detection cushion and the device interacting with the cushion, and comprises the following steps:
a1, a user starts to use the cushion, and the data of the pressure sensor array is detected through a sensor measuring circuit arranged in the cushion;
a2, a control chip arranged in the cushion, and used for collecting data transmitted by the sensor measuring circuit; aiming at received pressure array data, the control chip firstly collects pressure distribution map data based on a data collection module;
secondly, based on a sitting posture recognition module, taking data in a sitting posture data sample library as a training sample, and constructing a sitting posture recognition model through a convolution algorithm;
finally, based on a posture data output module, taking a sitting posture recognition model obtained by training as a basic algorithm model to obtain user posture data; the user posture data is further transmitted to a device interacting with the cushion through a data transmission module;
a3, receiving the sitting posture data through the data receiving and transmitting module arranged in the device and further transmitting the sitting posture data to the main control chip;
a4, after receiving the sitting posture data, the main control chip uploads the data to a cloud server; on the other hand, when the received poor sitting posture information is judged, the timing mode is started, and after the duration time of the detected poor sitting posture exceeds the preset time threshold, the video and audio decoding module is controlled to send out vibration reminding and/or video and audio reminding.
Further, the cloud server side uploads the sitting posture state information of the user to a website or an APP at intervals of t.
In the child sitting posture detection cushion, the device interacting with the cushion and the sitting posture monitoring method, the practical effects are as follows:
1. the pressure sensor detection part is arranged in a detachable and replaceable cushion by adopting a cushion type design, and only needs to be arranged on a seat when in use, and the design is convenient to carry;
2. the sitting posture detection is realized by adopting the low-cost pressure sensor, and the cost performance is high.
3. A device which is provided with a display device and interacts with the cushion is adopted to display sitting posture detection and processing results and simultaneously complete the functions of mode selection, correction reminding and the like;
4. the received electronic data is processed by utilizing the convolution algorithm, and a model is established without manually finding characteristics of the data detected by the sensor, so that the intelligent modeling of the electronic data is directly carried out by utilizing the convolution algorithm without manual intervention;
5. the convolution algorithm is applied to the electronic data, wherein the electronic data can be arranged into an image array for processing, and the advantages of intelligence and matching of deep learning are further embodied.
Drawings
The invention will be further described with reference to the accompanying drawings and examples, in which:
FIG. 1 is a view showing the arrangement of the devices installed in a seat cushion;
FIG. 2 is a schematic view of a pressure sensor distribution;
FIG. 3 is a flow chart of a convolution algorithm training model;
fig. 4 is a flowchart of the program operation of the device provided in the seat cushion;
FIG. 5 is a block diagram of the apparatus interacting with the seat cushion;
figure 6 is a schematic flow chart of a program for a seat cushion and devices interacting with the seat cushion.
Detailed Description
For a more clear understanding of the technical features, objects and effects of the present invention, embodiments of the present invention will now be described in detail with reference to the accompanying drawings.
Some existing sitting posture detection products are physically effective and need to be worn as required, so that the movement of children is limited, and uncomfortable feeling is easy to generate. In addition, some related sitting posture detection products have intellectualization, but have some technical defects. The main embodiments are as follows:
(1) the system is complex, occupies large space and is not suitable for children to learn scenes.
(2) The child-resistant glasses are required to be worn on a child to collect related data, so that the actions of the child are limited, and discomfort is easily caused.
(3) The accuracy of detecting the sitting posture is not high, and the condition of false detection is easy to occur.
(4) The interaction platform is a personal mobile phone or a tablet personal computer, and the application scene does not conform to learning environments such as classrooms and computer rooms.
(5) The price is expensive and is not suitable for popularization.
The invention aims to design a sitting posture detection cushion which is high in accuracy and strong in interestingness and is suitable for being used in multiple scenes of children, an intelligent interaction system and a sitting posture detection method. The invention has the following beneficial effects:
(1) the invention realizes the sitting posture detection only by adopting the low-cost pressure sensor and has high cost performance.
(2) The pressure sensor detection part is placed in a detachable and replaceable cushion by adopting a cushion type design, and only needs to be placed on a seat when in use, and the design is convenient to carry.
(3) The pressure sensor array data are converted into the sitting posture pressure distribution map, the deep learning convolutional neural network is used for carrying out sample training, an algorithm model with a learning effect is constructed, the sitting posture state is classified efficiently, and the sitting posture detection accuracy is improved.
(4) The invention is designed by an interactive platform, and completes the functions of mode selection, correction reminding and the like.
(5) According to the characteristics of the psychological activities of children, audio and video contents with strong interestingness are designed, and meanwhile, various modes such as outward playing, vibration and the like are set, so that the method is suitable for various environments, particularly learning environments such as classrooms, machine rooms and the like.
Example 1:
the invention provides a child sitting posture detection cushion which comprises a cushion body, wherein a measurement monitoring module L1, a plurality of pressure sensors L2 and a data transmission module L3 are arranged in the cushion body, and the pressure sensors L2 and the data transmission module L3 are distributed in the cushion body in an array form; wherein (please refer to fig. 1 for the distribution structure diagram of the device in the cushion):
the pressure sensors are arranged in 36 rows and 36 columns to form a pressure sensor array L2 which is uniformly arranged in the seat cushion and used for monitoring the pressure distribution data of the human sitting posture; (please refer to fig. 2 for a distribution diagram of the pressure sensors); wherein each small square, such as square "1", represents an array of pressure sensors in 6 rows and 6 columns.
The measurement monitoring module L1 comprises a control chip L11 and a sensor measurement circuit L12 which are connected in sequence; in this embodiment, the measurement monitoring module L1 is connected to the pressure sensor array L2 through a flat cable interface. Wherein:
one end of the sensor measuring circuit L12 is connected to the pressure sensor array L2, the other end of the sensor measuring circuit L8926 is connected to the control chip L11, and the sensor measuring circuit receives the electric signal which is transmitted by the pressure sensor array L2 and reflects the pressure distribution data of the sitting posture of the human body and further transmits the electric signal to the control chip L11;
for the received pressure array data, on one hand, the control chip L11 utilizes a convolution algorithm to model the pressure distribution diagram data after reducing the pressure array data into a pressure distribution diagram; on the other hand, aiming at the input human sitting posture pressure distribution data, the sitting posture data of the user is obtained by utilizing a modeling model, wherein automatic marking and parameter adjustment are carried out according to the specific sitting posture conditions of different users in the modeling process, the generated pressure distribution diagram is accurately classified in the sitting posture, and the misjudgment rate of the system is reduced; the obtained sitting posture data of the user is further transmitted to a device interacting with the seat cushion through a data transmission module L3 for further processing.
In this embodiment, the control chip L11 is provided with a data acquisition module L111;
the data acquisition module L111 is used for acquiring pressure distribution map data and establishing a sitting posture data sample library based on the acquired data;
a sitting posture recognition module L112 is also arranged in the control chip L11; the sitting posture recognition module L112 is used for constructing a sitting posture recognition model based on a convolution algorithm; the data in the sitting posture data sample library is used as a training sample to train the sitting posture recognition module; wherein:
1. the generation principle of the sitting posture recognition model is as follows:
firstly, an array is formed by adopting denser pressure sensors and is used for collecting sitting pressure information of the buttocks of a user. Regarding the data of each pressure sensor as a pixel, all the data collected by the pressure sensor array each time can be regarded as a sitting posture pressure distribution graph;
finding the relationship between the sitting posture and the sitting posture pressure distribution diagram can reversely deduce the sitting posture type from the sitting posture pressure distribution diagram. The system constructs a training model by using a convolutional neural network method, and classifies five states of 'left inclination', 'right inclination', 'normal', 'forward leaning' and 'backward leaning' for the sitting posture by extracting the characteristics of the sitting posture pressure distribution diagram. A large number of sitting posture pressure distribution map samples of different crowds are collected for training to obtain a basic model with a sitting posture classification function. When the system is used by a user, the system collects the actual sitting posture data of the user and repeatedly trains and adjusts the basic model. The longer the user is used, the higher the algorithm fitting degree is, and finally the higher classification accuracy can be achieved.
2. Based on the convolution algorithm, the process of constructing the sitting posture recognition model includes (refer to fig. 3 specifically):
s1, collecting sitting posture pressure distribution diagram samples from the sitting posture data sample library, classifying according to different sitting postures of the human body, and further constructing a training set and a testing set; the sitting posture pressure distribution diagram samples of a large number of different crowds can be collected, and classified according to left inclination, right inclination, normal, front leaning and back leaning, and a training set and a test set with a certain proportion are sorted out and used as input data of the neural network.
S2, carrying out forward propagation once by using a training set sample in the currently constructed sitting posture recognition model; after the features are extracted through the convolutional layer, the data dimensionality is reduced through the pooling layer, and the features are extracted again through the full-connection layer, the sitting posture state is output, wherein the network parameters are updated according to the error between the output sitting posture state and the actual sitting posture state; in the current step, repeating iteration until the output sitting posture state is consistent with the actual sitting posture state, and stopping network training;
s3, inputting a test set into the sitting posture recognition model obtained by training in the step S2, and testing; when the test classification accuracy does not reach the preset standard value, repeatedly executing the step S2 and the step S3; when the final data classification accuracy reaches a preset standard value, taking a currently obtained sitting posture recognition model as a basic algorithm model; and an application and attitude data output module;
when the user starts to use the system, the user firstly inputs the self sitting posture sample. The system trains the algorithm model by using the self-sitting posture data of the user on the basis of the basic algorithm model so as to improve the accuracy of the system in detecting the self-sitting posture of the user. In the use process, the user can also actively calibrate and upload the self sitting posture sample. The system constructs a user personal database and continuously trains an optimization algorithm.
The control chip L11 is further provided with a posture data output module L113, configured to obtain sitting posture data of the user with respect to the received pressure array data by using the sitting posture recognition model obtained through training, where the sitting posture of the user is further transmitted to the data transmission module L3, and the data transmission module L3 is not limited to a wired, wireless, or bluetooth connection manner, and performs corresponding data transmission.
In the embodiment, the pressure sensor array is arranged in the seat cushion, the deep learning algorithm is combined, the pressure distribution diagram output by the pressure sensor array is used as training data, a basic algorithm model is constructed, the sitting posture of the user is further pre-judged and output, and the final data output accuracy is effectively improved.
Please refer to fig. 4, which is a flowchart illustrating a program operation of a device installed in a seat cushion, wherein it can be seen that, based on a main control chip (equivalent to the control chip L11 described above), the data output by the pressure sensor array and the training data (pressure distribution map) output by the database sample are received and processed, and then the sitting posture of the human body is further determined, wherein the determined output is output to the terminal display device or the only device performing data interaction with the terminal display device through a wireless data transmission module (equivalent to the data transmission module described above, in this embodiment, the transmission mode is set as wireless transmission), and in the sample database, in order to enrich the sample database, personal sitting posture data and multi-person sitting posture data are respectively input and stored; the human body sitting posture is divided into five states of left-leaning, right-leaning, normal, leaning forward and leaning backward through the algorithm model. The algorithm model is obtained by training a large amount of sample data through a deep learning neural network method, and when the algorithm model is used, a personal database can be distributed to a user by using a cloud server, and the algorithm model is continuously trained and optimized according to data fed back by the user or automatically calibrated.
Example 2:
the invention discloses a device interacting with a cushion, which can further feed back data detected by the cushion to a user.
The device comprises a data receiving and transmitting module A1, a main control chip A2 and a video and audio decoding module A3 (please refer to FIG. 5 for the specific structure diagram); wherein:
the data receiving and transmitting module A1 is connected to the main control chip A2 and is used for receiving the sitting posture data and further transmitting the sitting posture data to the main control chip A2;
the main control chip A2 is connected to the video and audio decoding module A3, and is used for starting a timing mode when the received information is judged to be bad sitting posture after receiving data, and controlling the video and audio decoding module A3 to send out vibration reminding and/or make video and audio reminding after the detected bad sitting posture duration time exceeds a preset time threshold.
In order to further display the data in real time, a display screen a4 is further provided in the device, and the display screen a4 is connected to the main control chip a2 for displaying the data in real time.
In order to further store the video and audio data output by the video and audio decoding module A3, the device further comprises an SD card storage module a5, wherein the SD card storage module a5 is respectively connected to the video and audio decoding module A3 and the main control chip a2 to store the video and audio data output by the video and audio decoding module.
In this embodiment, the main control chip a2 is connected to a cloud server, and further uploads the data processed by the main control chip a2 to the cloud server for storage.
Please refer to fig. 6, which is a schematic program flow diagram of a seat cushion and a device interacting with the seat cushion according to embodiments 1 and 2 of the present invention, wherein the device interacting with the seat cushion is configured as an interactive platform (watch), and the interactive platform is designed in a watch style so as to be conveniently worn by a child. Meanwhile, the multifunctional sitting posture reminding device has multiple adjustable modes, can be used and carried in different scenes, is particularly suitable for environments such as classrooms and machine rooms and the like which cannot use electronic equipment such as a personal mobile phone and a tablet personal computer, selects a vibration mode to realize sitting posture reminding, sets a user target as a child in the implementation process, collects sitting posture data of the child through a pressure sensor array, transmits the currently monitored sitting posture data to a control chip arranged in a cushion after low-pass filtering (interference of noise data is avoided), and further judges a sitting posture structure based on a basic sitting posture judgment model obtained by a convolution algorithm in the control chip; and then, the judgment result is further transmitted to the watch in a wireless mode, the sitting posture state of the user is displayed through a display screen of the watch, and the reminding function is further achieved through the video module and the audio module which are set in the watch according to the setting of the user.
In the embodiment, the characteristics that the sitting posture pressure distribution diagram on the cushion has the same overall area and approximately same stress area are utilized, a watch type interactive platform design is adopted, a 2.8-inch liquid crystal display screen is used as terminal equipment, the sitting posture detection and processing result is displayed, and functions of mode selection, correction reminding and the like are completed; the method is characterized in that an algorithm based on a convolutional neural network is adopted for the first time, a large number of samples rapidly collected by a microprocessor are trained and verified, and the sitting posture of a user is accurately detected.
In this embodiment, the used convolution algorithm has a "learning" function, and can automatically adjust relevant parameters according to the data feedback condition of the system, and automatically label the classification model.
The method for detecting the sitting posture of the child can be suitable for children with different body types and is also beneficial to the same user who grows and develops rapidly; the method for detecting the sitting posture of the child can also design audio and video contents with strong interestingness according to the characteristics of the psychological activities of the child, and can be suitable for various environments, particularly learning environments such as classrooms, machine rooms and the like by setting various modes such as outward playing, vibration and the like.
Example 3:
this embodiment will further the children's position of sitting that embodiment 1 provided detect the cushion to and embodiment 2 provides a device interactive with the cushion, further realize children's position of sitting and detect, specifically include following step:
1. a user starts to use the cushion, and the data of the pressure sensor array is detected through a sensor measuring circuit arranged in the cushion;
2. the control chip is arranged in the cushion and used for collecting data transmitted by the sensor measuring circuit; aiming at received pressure array data, the control chip firstly collects pressure distribution map data based on a data collection module;
secondly, based on a sitting posture recognition module, taking data in a sitting posture data sample library as a training sample, and constructing a sitting posture recognition model through a convolution algorithm;
finally, based on a posture data output module, taking a sitting posture recognition model obtained by training as a basic algorithm model to obtain user posture data; the user posture data is further transmitted to a device interacting with the cushion through a data transmission module;
3. the data receiving and transmitting module arranged in the device is used for receiving the sitting posture data and further transmitting the sitting posture data to the main control chip;
4. after receiving the sitting posture data, the main control chip uploads the data to a cloud server; on the other hand, when the received poor sitting posture information is judged, the timing mode is started, and after the duration time of the detected poor sitting posture exceeds the preset time threshold, the video and audio decoding module is controlled to send out vibration reminding and/or video and audio reminding. The cloud server uploads the sitting posture state information of the user to a website or an APP at intervals of t.
In this embodiment, children sit on the position of sitting detects and stacks up, can set up switch, self-checking lamp on the cushion, when pressing switch, the spare part of effect enters into the operation module in the cushion. At the moment, the sitting posture detection pad starts self-checking, and can be normally used after the self-checking normal lamp is turned on;
in the embodiment, the user can set and select the reminding mode to be the vibration reminding mode or the video and audio reminding mode through the watch type interactive platform, so that the entertainment of the interactive platform is improved.
In this embodiment, the low-pass filtering processing can be performed on the sitting posture pressure data collected by the pressure sensor array of the sitting posture detection pad to filter out interference noise.
In this embodiment, after the data is uploaded to the cloud server, the data can be restored to a pressure distribution map in the cloud server, and the current algorithm model can acquire training data and test data from the cloud server for further network training; the user can feed back the sitting posture detection effect to the cloud server through the watch type interactive platform, and the cloud server continuously updates the data according to the user feedback data.
While the present invention has been described with reference to the embodiments shown in the drawings, the present invention is not limited to the embodiments, which are illustrative and not restrictive, and it will be apparent to those skilled in the art that various changes and modifications can be made therein without departing from the spirit and scope of the invention as defined in the appended claims.
Claims (10)
1. A child sitting posture detection cushion comprises a cushion body and is characterized in that a measurement monitoring module, a plurality of pressure sensors distributed in the cushion in an array mode and a data transmission module are arranged in the cushion; wherein:
the pressure sensors are uniformly arranged in the cushion in an array form of n rows and m columns and used for monitoring the pressure distribution data of the human body sitting posture; wherein n and m are more than or equal to 1 and are positive integers;
the measurement monitoring module comprises a control chip and a sensor measuring circuit which are sequentially connected; wherein: one end of the sensor measuring circuit is connected to the pressure sensor array, and the other end of the sensor measuring circuit is connected to the control chip and used for receiving the electric signals which are transmitted by the pressure sensor array and reflect the pressure distribution data of the sitting posture of the human body and further transmitting the electric signals to the control chip;
the control chip is used for modeling the pressure distribution diagram data by utilizing a convolution algorithm after reducing the received pressure array data into a pressure distribution diagram on one hand; on the other hand, aiming at the input human sitting posture pressure distribution data, the sitting posture data of the user is obtained by utilizing a modeling model, wherein automatic marking and parameter adjustment are carried out according to the specific sitting posture conditions of different users in the modeling process, and the generated pressure distribution diagram is accurately classified in the sitting posture; the obtained user sitting posture data is further transmitted to a device interacting with the cushion through a data transmission module for further processing.
2. The child sitting posture detecting cushion as claimed in claim 1, wherein the measuring and monitoring module is connected with the pressure sensor array through a flat cable interface.
3. The child sitting posture detecting cushion as claimed in claim 1, wherein a data acquisition module is arranged in the control chip;
the data acquisition module is used for acquiring pressure distribution map data and establishing a sitting posture data sample library based on the acquired data;
a sitting posture recognition module is also arranged in the control chip; the sitting posture recognition module is used for constructing a sitting posture recognition model based on a convolution algorithm; the data in the sitting posture data sample library is used as a training sample to train the sitting posture recognition module;
the control chip is also provided with a posture data output module which is used for obtaining the sitting posture data of the user aiming at the received pressure array data by utilizing the sitting posture recognition model obtained by training, and the sitting posture of the user is further transmitted to the data transmission module.
4. The child sitting posture detection cushion of claim 3, wherein the process of constructing the sitting posture recognition model based on the convolution algorithm comprises:
s1, collecting sitting posture pressure distribution diagram samples from the sitting posture data sample library, classifying according to different sitting postures of the human body, and further constructing a training set and a testing set;
s2, carrying out forward propagation once by using a training set sample in the currently constructed sitting posture recognition model; after the features are extracted through the convolutional layer, the data dimensionality is reduced through the pooling layer, and the features are extracted again through the full-connection layer, the sitting posture state is output, wherein the network parameters are updated according to the error between the output sitting posture state and the actual sitting posture state; in the current step, repeating iteration until the output sitting posture state is consistent with the actual sitting posture state, and stopping network training;
s3, inputting a test set into the sitting posture recognition model obtained by training in the step S2, and testing; when the test classification accuracy does not reach the preset standard value, repeatedly executing the step S2 and the step S3; when the final data classification accuracy reaches a preset standard value, taking a currently obtained sitting posture recognition model as a basic algorithm model; and an application and attitude data output module.
5. A device for interacting with a seat cushion, the device being configured to receive sitting posture data of a user; the device comprises a data receiving and transmitting module, a main control chip and a video and audio decoding module; wherein:
the data receiving and transmitting module is used for receiving the sitting posture data and further transmitting the sitting posture data to the main control chip;
the main control chip is used for starting a timing mode when the received bad sitting posture information is judged after data is received, and controlling the video and audio decoding module to send out vibration reminding and/or make video and audio reminding after the detected bad sitting posture duration time exceeds a preset time threshold.
6. A device for interacting with a seat cushion as claimed in claim 5, further comprising a display screen connected to the main control chip for real-time data presentation.
7. The device of claim 5, wherein the platform further comprises an SD card storage module, the SD card storage module is respectively connected to the video and audio decoding module and the main control chip, and is configured to store the video and audio data output by the video and audio decoding module.
8. A device interacting with a seat cushion as claimed in claim 5, wherein the main control chip is connected to a cloud server, and the data processed by the main control chip is further uploaded to the cloud server for storage.
9. A sitting posture monitoring method further implemented based on a child sitting posture detecting cushion as claimed in claims 1-4 and a device interacting with the cushion as claimed in claims 5-8, comprising the steps of:
a1, a user starts to use the cushion, and the data of the pressure sensor array is detected through a sensor measuring circuit arranged in the cushion;
a2, a control chip arranged in the cushion, and used for collecting data transmitted by the sensor measuring circuit; aiming at received pressure array data, the control chip firstly collects pressure distribution map data based on a data collection module;
secondly, based on a sitting posture recognition module, taking data in a sitting posture data sample library as a training sample, and constructing a sitting posture recognition model through a convolution algorithm;
finally, based on a posture data output module, taking a sitting posture recognition model obtained by training as a basic algorithm model to obtain user posture data; the user posture data is further transmitted to a device interacting with the cushion through a data transmission module;
a3, receiving the sitting posture data through the data receiving and transmitting module arranged in the device and further transmitting the sitting posture data to the main control chip;
a4, after receiving the sitting posture data, the main control chip uploads the data to a cloud server; on the other hand, when the received poor sitting posture information is judged, the timing mode is started, and after the duration time of the detected poor sitting posture exceeds the preset time threshold, the video and audio decoding module is controlled to send out vibration reminding and/or video and audio reminding.
10. The sitting posture monitoring method as claimed in claim 9, wherein the cloud server uploads the sitting posture information of the user to the website or APP at intervals of time t.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911125903.1A CN110897425A (en) | 2019-11-18 | 2019-11-18 | Child sitting posture detection cushion, device interacting with cushion and sitting posture monitoring method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911125903.1A CN110897425A (en) | 2019-11-18 | 2019-11-18 | Child sitting posture detection cushion, device interacting with cushion and sitting posture monitoring method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110897425A true CN110897425A (en) | 2020-03-24 |
Family
ID=69817698
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911125903.1A Pending CN110897425A (en) | 2019-11-18 | 2019-11-18 | Child sitting posture detection cushion, device interacting with cushion and sitting posture monitoring method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110897425A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112120464A (en) * | 2020-10-28 | 2020-12-25 | 上海健康医学院 | Health seat and working method thereof |
| CN112651865A (en) * | 2020-12-30 | 2021-04-13 | 北京市商汤科技开发有限公司 | Behavior state prompting method and device, electronic equipment and storage medium |
| CN112773356A (en) * | 2020-11-16 | 2021-05-11 | 深圳数联天下智能科技有限公司 | Sitting posture detection method, system, terminal and storage medium |
| CN112949434A (en) * | 2021-02-19 | 2021-06-11 | 清华大学 | Human body sitting posture recognition system based on pressure cushion |
| CN113349599A (en) * | 2021-07-26 | 2021-09-07 | 合肥工业大学 | Sitting posture adjusting method and sitting posture adjusting seat |
| CN113384266A (en) * | 2021-07-26 | 2021-09-14 | 合肥工业大学 | Intelligent sitting posture monitoring and analyzing system and method |
| CN117838107A (en) * | 2024-03-07 | 2024-04-09 | 亿慧云智能科技（深圳）股份有限公司 | Healthy sitting posture monitoring method, device and equipment of intelligent cushion and storage medium |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103682078A (en) * | 2012-09-21 | 2014-03-26 | 国家纳米科学中心 | Pressure sensor array and manufacturing method thereof |
| CN103745577A (en) * | 2014-01-16 | 2014-04-23 | 中国工程物理研究院电子工程研究所 | Sedentary reminding and sitting posture detection device based on smart mobile device |
| CN105902049A (en) * | 2016-06-17 | 2016-08-31 | 心动网络股份有限公司 | Pressure testing device and chair cushion for correcting sitting posture |
| CN106406521A (en) * | 2016-08-30 | 2017-02-15 | 乐视控股（北京）有限公司 | Sitting posture detection method and apparatus |
| CN107232822A (en) * | 2016-08-31 | 2017-10-10 | 浙江玛拉蒂智能家具科技有限公司 | A kind of Intelligent seat based on gesture recognition, intelligent interactive system and method |
| CN107296427A (en) * | 2017-08-24 | 2017-10-27 | 云康新智能科技（深圳）有限公司 | A kind of cushion and sitting posture analysis method |
| US20170325599A1 (en) * | 2016-05-13 | 2017-11-16 | Srigiri Shankar Bellam | Posture Detection and Correction Cushion |
| CN109171315A (en) * | 2018-06-12 | 2019-01-11 | 深圳市格兰莫尔寝室用品有限公司 | Automatically adjust mattress and intelligent bed |
| CN109820373A (en) * | 2019-03-28 | 2019-05-31 | 重庆邮电大学 | Sitting posture self-adapting regulation method based on Intelligent seat |
| CN110174213A (en) * | 2019-05-29 | 2019-08-27 | 上海交通大学 | The calibration method of pliable pressure sensor array |
-
2019
- 2019-11-18 CN CN201911125903.1A patent/CN110897425A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103682078A (en) * | 2012-09-21 | 2014-03-26 | 国家纳米科学中心 | Pressure sensor array and manufacturing method thereof |
| CN103745577A (en) * | 2014-01-16 | 2014-04-23 | 中国工程物理研究院电子工程研究所 | Sedentary reminding and sitting posture detection device based on smart mobile device |
| US20170325599A1 (en) * | 2016-05-13 | 2017-11-16 | Srigiri Shankar Bellam | Posture Detection and Correction Cushion |
| CN105902049A (en) * | 2016-06-17 | 2016-08-31 | 心动网络股份有限公司 | Pressure testing device and chair cushion for correcting sitting posture |
| CN106406521A (en) * | 2016-08-30 | 2017-02-15 | 乐视控股（北京）有限公司 | Sitting posture detection method and apparatus |
| CN107232822A (en) * | 2016-08-31 | 2017-10-10 | 浙江玛拉蒂智能家具科技有限公司 | A kind of Intelligent seat based on gesture recognition, intelligent interactive system and method |
| CN107296427A (en) * | 2017-08-24 | 2017-10-27 | 云康新智能科技（深圳）有限公司 | A kind of cushion and sitting posture analysis method |
| CN109171315A (en) * | 2018-06-12 | 2019-01-11 | 深圳市格兰莫尔寝室用品有限公司 | Automatically adjust mattress and intelligent bed |
| CN109820373A (en) * | 2019-03-28 | 2019-05-31 | 重庆邮电大学 | Sitting posture self-adapting regulation method based on Intelligent seat |
| CN110174213A (en) * | 2019-05-29 | 2019-08-27 | 上海交通大学 | The calibration method of pliable pressure sensor array |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112120464A (en) * | 2020-10-28 | 2020-12-25 | 上海健康医学院 | Health seat and working method thereof |
| CN112773356A (en) * | 2020-11-16 | 2021-05-11 | 深圳数联天下智能科技有限公司 | Sitting posture detection method, system, terminal and storage medium |
| CN112651865A (en) * | 2020-12-30 | 2021-04-13 | 北京市商汤科技开发有限公司 | Behavior state prompting method and device, electronic equipment and storage medium |
| CN112949434A (en) * | 2021-02-19 | 2021-06-11 | 清华大学 | Human body sitting posture recognition system based on pressure cushion |
| CN113349599A (en) * | 2021-07-26 | 2021-09-07 | 合肥工业大学 | Sitting posture adjusting method and sitting posture adjusting seat |
| CN113384266A (en) * | 2021-07-26 | 2021-09-14 | 合肥工业大学 | Intelligent sitting posture monitoring and analyzing system and method |
| CN117838107A (en) * | 2024-03-07 | 2024-04-09 | 亿慧云智能科技（深圳）股份有限公司 | Healthy sitting posture monitoring method, device and equipment of intelligent cushion and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110897425A (en) | Child sitting posture detection cushion, device interacting with cushion and sitting posture monitoring method | |
| CN106984027B (en) | Action comparison analysis method and device and display | |
| CN104157107B (en) | A kind of human posture's apparatus for correcting based on Kinect sensor | |
| CN108814616A (en) | A kind of sitting posture knows method for distinguishing and Intelligent seat | |
| CN109069350B (en) | Infant feeding system | |
| CN106991619A (en) | A kind of diseases and pests of agronomic crop intelligent diagnosis system and diagnostic method | |
| CN107754225A (en) | A kind of intelligent body-building coaching system | |
| CN108601133A (en) | A kind of intelligent desk lamp and the sitting posture correction function method based on intelligent desk lamp | |
| CN108937965B (en) | Attention evaluation system and method based on sitting posture analysis | |
| CN103845038B (en) | A kind of sign signal acquisition method and apparatus | |
| KR20210121616A (en) | Health care device, system and method | |
| CN111243742B (en) | Intelligent glasses capable of analyzing eye habit of children | |
| CN112101124B (en) | Sitting posture detection method and device | |
| CN111444748B (en) | Sitting posture detection method, device, equipment and storage medium | |
| KR102466438B1 (en) | Cognitive function assessment system and method of assessing cognitive funtion | |
| CN106774861B (en) | Intelligent device and behavior data correction method and device | |
| CN114343574A (en) | Sleep quality evaluation method, client, server and storage medium | |
| CN110321820A (en) | A kind of sight drop point detection method based on contactless device | |
| CN110737201A (en) | monitoring method, device, storage medium and air conditioner | |
| CN114067435A (en) | Sleep behavior detection method and system based on pseudo-3D convolutional network and attention mechanism | |
| CN108492855A (en) | A kind of apparatus and method for training the elderly's attention | |
| CN115586834A (en) | Intelligent cardio-pulmonary resuscitation training system | |
| CN107019501B (en) | Remote tumble detection method and system based on genetic algorithm and probabilistic neural network | |
| CN111768863A (en) | Artificial intelligence-based infant development monitoring system and method | |
| CN113743255A (en) | Neural network-based child sitting posture identification and correction method and system |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200324 |
|
| RJ01 | Rejection of invention patent application after publication |