Data processing method, system and medium for smart seat

By collecting and storing pressure signals and timestamps in smart seats and building a timing database, the training efficiency of the cushion hardness adjustment model is improved, and the problem of low model training efficiency in the existing technology is solved, so that users of smart seats are not allowed Perceived hardness adjustment.

CN118708633BActive Publication Date: 2024-11-29SHENZHEN SIHOO INTELLIGENT FURNITURE CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202411187762.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-28
Publication Date
2024-11-29
Estimated Expiration
2044-08-28

AI Technical Summary

Technical Problem

Among existing smart seats, machine learning models used to adjust the hardness of the seat cushion are less efficient and difficult to improve efficiently.

Method used

By setting up pressure sensors and main control units in the smart seats, the user's pressure signals are periodically collected and the corresponding timestamps are stored, the timing database is built, and the sample data is matched according to the preset time range is improved to improve the sample number and diversity, and used for training Cush hardness adjustment model.

Benefits of technology

It effectively improves the training efficiency of the cushion hardness adjustment model, improves the prediction accuracy and applicability of the model, and realizes the unsensible cushion hardness adjustment of the user.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118708633B_ABST
    Figure CN118708633B_ABST
Patent Text Reader

Abstract

The disclosed embodiments disclose a data processing method, system and medium for a smart seat, wherein a pressure sensor of the smart seat periodically collects pressure signals, and sends the pressure signals and corresponding collection timestamps to a main control unit; the main control unit converts the pressure signals into cushion sensing data, obtains an account ID and a seat ID, creates candidate data entries according to the account ID, the seat ID, the cushion sensing data and the collection timestamp, and stores the candidate data entries in a time series database; when the main control unit receives a query instruction, in the time series database, matches corresponding sample data entries according to a preset time range, and sends the sample data entries to a server, which trains a cushion hardness adjustment model based on the cushion sensing data, and stores the trained cushion hardness adjustment model in association with the account ID and the seat ID, thereby effectively improving the training efficiency of the cushion hardness adjustment model.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to the field of artificial intelligence technology, and in particular to a data processing method, system and medium for a smart seat. Background Art

[0002] With the development of science and technology, seats are becoming more and more intelligent. In related technologies, there are solutions that input the user's sitting posture and the actual support of the seat cushion into a machine learning model for prediction, thereby automatically adjusting the hardness of the seat cushion. However, such technical solutions rely on trained machine learning models. Therefore, how to effectively improve the training efficiency of such machine learning models for adjusting the hardness of seat cushions is an urgent problem to be solved. Summary of the invention

[0003] The following is a summary of the subject matter of the detailed description of the present disclosure. This summary is not intended to limit the scope of the claims.

[0004] The disclosed embodiments provide a data processing method, system and medium for a smart seat, which can improve the training efficiency of a seat cushion hardness adjustment model.

[0005] On the one hand, an embodiment of the present disclosure provides a data processing method of a smart seat, wherein the smart seat is provided with a main control unit and a pressure sensor, and the data processing method includes:

[0006] The pressure sensor periodically collects the pressure signal of the target user based on a preset detection frequency, and whenever a pressure signal is collected, sends the pressure signal and the collection timestamp of the pressure signal to the main control unit;

[0007] Whenever the main control unit receives a pressure signal and an acquisition timestamp, it converts the pressure signal to obtain cushion sensing data, obtains the account ID of the target user's currently logged-in account and the seat ID of the smart seat, creates candidate data entries based on the account ID, seat ID, cushion sensing data and acquisition timestamp, and stores the candidate data entries in a time series database;

[0008] The main control unit receives a query instruction sent by the server, wherein the query instruction carries a plurality of preset time ranges of different lengths;

[0009] The main control unit matches the multiple candidate data entries in the time series database with the collection timestamp according to the preset time range, obtains the sample data entry corresponding to the preset time range, and sends the sample data entry to the server;

[0010] The server constructs multiple seat cushion sensing data into sample time series feature data based on the collection timestamp, trains a seat cushion hardness adjustment model based on the sample time series feature data, and associates the trained seat cushion hardness adjustment model with the account ID and the seat ID for storage.

[0011] On the other hand, an embodiment of the present disclosure provides a data processing system of a smart seat, including a smart seat and a server, wherein the smart seat is provided with a main control unit and a pressure sensor:

[0012] The pressure sensor is used to periodically collect the pressure signal of the target user based on a preset detection frequency, and whenever a pressure signal is collected, the pressure signal and the collection timestamp of the pressure signal are sent to the main control unit;

[0013] The main control unit is used to obtain the cushion sensing data based on the pressure signal conversion whenever receiving the pressure signal and the acquisition timestamp, obtain the account ID of the target user's currently logged-in account and the seat ID of the smart seat, create candidate data entries according to the account ID, the seat ID, the cushion sensing data and the acquisition timestamp, and store the candidate data entries in the time series database;

[0014] The main control unit is also used to receive a query instruction sent by the server, wherein the query instruction carries a plurality of preset time ranges of different lengths;

[0015] The main control unit is also used to match the collection timestamp with the preset time range among the multiple candidate data entries in the time series database, obtain the sample data entry corresponding to the preset time range, and send the sample data entry to the server;

[0016] The server is used to construct multiple seat cushion sensing data into sample time series feature data based on the collection timestamp, train the seat cushion hardness adjustment model based on the sample time series feature data, and store the trained seat cushion hardness adjustment model in association with the account ID and the seat ID.

[0017] On the other hand, an embodiment of the present disclosure further provides a computer-readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the data processing method of the above-mentioned smart seat.

[0018] The disclosed embodiments include at least the following beneficial effects: the pressure sensor of the smart seat periodically collects the pressure signal of the target user based on a preset detection frequency, and sends the pressure signal and the collection timestamp of the pressure signal to the main control unit whenever the pressure signal is collected; whenever the main control unit receives the pressure signal and the collection timestamp, it converts the pressure signal into cushion sensing data, obtains the account ID of the target user's currently logged-in account and the seat ID of the smart seat, creates candidate data entries based on the account ID, seat ID, cushion sensing data and collection timestamp, and stores the candidate data entries in a time series database; by periodically collecting pressure signals and recording corresponding collection timestamps, it is possible to obtain the pressure signal from the perspective of fine-grained timestamp granularity. The seat cushion sensing data is stored so that when the main control unit receives a query instruction sent by the server that carries a plurality of preset time ranges of different lengths, the main control unit can match the multiple candidate data entries in the time series database according to the preset time range and the collection timestamp, and obtain the sample data entries corresponding to the different preset time ranges, thereby constructing different training samples, increasing the number of samples, and then sending the sample data entries to the server. The server constructs the multiple seat cushion sensing data into sample time series feature data based on the collection timestamp, trains the seat cushion hardness adjustment model based on the sample time series feature data, and stores the trained seat cushion hardness adjustment model in association with the account ID and the seat ID, thereby effectively improving the training efficiency of the seat cushion hardness adjustment model.

[0019] Other features and advantages of the present disclosure will be set forth in the following description, and in part will be apparent from the description, or may be learned by practicing the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The accompanying drawings are used to provide further understanding of the technical solution of the present disclosure and constitute a part of the specification. Together with the embodiments of the present disclosure, they are used to explain the technical solution of the present disclosure and do not constitute a limitation on the technical solution of the present disclosure.

[0021] Figure 1 A schematic diagram of an optional implementation environment provided for an embodiment of the present disclosure;

[0022] Figure 2 An optional flow chart of the data processing method of the smart seat provided in the embodiment of the present disclosure;

[0023] Figure 3 An optional flow chart for constructing sample time series feature data for a server provided in an embodiment of the present disclosure;

[0024] Figure 4 An optional flow chart of a training cushion hardness adjustment model provided in an embodiment of the present disclosure;

[0025] Figure 5 Another optional flow chart of the data processing method of the smart seat provided in the embodiment of the present disclosure;

[0026] Figure 6 An optional flow chart for determining a target time range provided in an embodiment of the present disclosure;

[0027] Figure 7 An optional flow chart of the main control unit adjusting the gas volume of the airbag provided in the embodiment of the present disclosure;

[0028] Figure 8 An optional flow chart for preferentially adjusting the gas volume of the first target airbag provided in an embodiment of the present disclosure;

[0029] Fig. 9 A gas volume distribution diagram of an airbag of a smart seat provided in an embodiment of the present disclosure;

[0030] Fig.10 An optional flow chart for preferentially adjusting the gas volume of the second target airbag provided in an embodiment of the present disclosure;

[0031] Fig.11 An optional structural schematic diagram of the smart chair provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0032] In order to make the purpose, technical solution and advantages of the present disclosure more clear, the present disclosure is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present disclosure and are not used to limit the present disclosure.

[0033] It should be noted that in various specific embodiments of the present disclosure, when it comes to the need to perform relevant processing based on data related to the characteristics of the target object such as the target object attribute information or attribute information set, the permission or consent of the target object will be obtained first, and the collection, use and processing of these data will comply with relevant laws, regulations and standards. Among them, the target object can be a user. In addition, when the embodiment of the present disclosure needs to obtain the attribute information of the target object, the separate permission or separate consent of the target object will be obtained through a pop-up window or by jumping to a confirmation page. After clearly obtaining the separate permission or separate consent of the target object, the necessary target object-related data used to enable the normal operation of the embodiment of the present disclosure will be obtained.

[0034] In the embodiments of the present disclosure, the term "module" or "unit" refers to a computer program or a part of a computer program that has a predetermined function and works together with other related parts to achieve a predetermined goal, and can be implemented in whole or in part by using software, hardware (such as processing circuits or memories), or a combination thereof. Similarly, a processor (or multiple processors or memories) can be used to implement one or more modules or units. In addition, each module or unit can be part of an overall module or unit that includes the function of the module or unit.

[0035] To facilitate understanding of the technical solution provided by the embodiments of the present disclosure, some key terms used in the embodiments of the present disclosure are explained here:

[0036] A timestamp is a mark or data used to record the time when a specific event occurred, usually expressed in the format of date and time. In computer science and data processing, timestamps can be used to mark the time point when data is created, modified, or accessed, making it easier to track and manage the sequence of events. The common timestamp format is the Unix timestamp, for example: "1691957100" represents the number of seconds that have passed since 00:00:00 UTC on January 1, 1970.

[0037] Reference Figure 1 , Figure 1 A schematic diagram of an optional implementation environment provided for an embodiment of the present disclosure, the implementation environment includes a pressure sensor 101 of a smart seat, a main control unit 102 of the smart seat, and a server 103, wherein the main control unit 102 and the pressure sensor 101 are connected via a communication network, and the main control unit 102 and the server 103 are connected via a communication network.

[0038] Exemplarily, the interactive process of a data processing method for a smart seat is as follows: the pressure sensor 101 periodically collects the pressure signal of the target user based on a preset detection frequency, and each time a pressure signal is collected, the pressure signal and the collection timestamp of the pressure signal are sent to the main control unit 102. Then, each time the main control unit 102 receives the pressure signal and the collection timestamp, the seat cushion sensing data is obtained based on the pressure signal conversion, the account ID of the target user's currently logged-in account and the seat ID of the smart seat are obtained, and candidate data entries are created according to the account ID, seat ID, seat cushion sensing data and collection timestamp. The candidate data entries are stored in the time series database, and then the main control unit 102 receives the query instruction sent by the server 103; then, the main control unit 102 matches the multiple candidate data entries in the time series database according to the preset time range and the collection timestamp, obtains the sample data entry corresponding to the preset time range, and sends the sample data entry to the server 103. Finally, the server 103 constructs the multiple cushion sensing data into sample time series feature data based on the collection timestamp, trains the cushion hardness adjustment model based on the sample time series feature data, and stores the trained cushion hardness adjustment model in association with the account ID and the seat ID.

[0039] It should be noted that the data transmission formats of the embodiments of the present disclosure connected through the communication network include but are not limited to JSON format (JavaScript Object Notation), XML format (Extensible Markup Language), HTML format (HyperText Markup Language) and other data transmission formats.

[0040] On the one hand, the principle of the data processing method of the smart seat provided by the embodiment of the present disclosure is described in detail below.

[0041] It should be noted that the data processing method of the smart seat provided in the embodiment of the present disclosure is applied to a smart seat, which is provided with a main control unit, multiple pressure sensors, multiple matrix-distributed airbags and a gas quantity sensor. Each pressure sensor is respectively arranged in the seat cushion of the smart seat, and each airbag is also respectively arranged in the seat cushion of the smart seat. The gas quantity sensor is used to periodically collect the gas quantity distribution matrix of multiple airbags based on the detection frequency.

[0042] The number of the pressure sensors is the same as the number of the airbags, and the multiple pressure sensors are configured in a one-to-one correspondence with the multiple airbags.

[0043] Reference Figure 2 As shown, Figure 2This is an optional flow chart of a data processing method for a smart seat provided in an embodiment of the present disclosure. The data processing method for a smart seat includes but is not limited to the following steps 201 to 205.

[0044] Step 201: The pressure sensor periodically collects the pressure signal of the target user based on a preset detection frequency, and sends the pressure signal and the collection timestamp of the pressure signal to the main control unit whenever a pressure signal is collected.

[0045] Among them, the preset detection frequency can be 5 times / second, 1 time / second, or 10 times / minute, which is not specifically limited in this application.

[0046] Specifically, the pressure sensor periodically collects the pressure signal of the target user based on a preset detection frequency. Whenever a pressure signal is collected, the pressure sensor associates the pressure signal with the timestamp of the current collection, and then sends the pressure signal and the collection timestamp of the pressure signal to the main control unit. The time when the pressure signal was collected can be accurately recorded by the timestamp, providing precise time characteristics for subsequent data analysis, thereby improving the timeliness of pressure signal collection.

[0047] In a possible implementation, the pressure sensor periodically collects the pressure signal of the target user based on a detection frequency of 1 time / second. Whenever a pressure signal is collected (for example, the pressure signal collected at the collection time "1723500000" is "[1kPa, 1kPa, 1kPa]", the pressure signal collected at the collection time "1723500001" is "[1kPa, 1.5kPa, 1.5kPa]", the pressure signal collected at the collection time "1723500002" is "[2kPa, 2kPa, 2kPa]", the pressure signal collected at the collection time "1723500003" is "[3kPa, 2kPa, 2kPa]"), the pressure signal is “[1kPa,1kPa,1kPa]” and the acquisition timestamp of the pressure signal is “1723500000” and is sent to the main control unit, the pressure signal is “[2kPa,2kPa,2kPa]” and the acquisition timestamp of the pressure signal is “1723500001” and is sent to the main control unit, the pressure signal is “[2kPa,2kPa,2kPa]” and the acquisition timestamp of the pressure signal is “1723500002” and is sent to the main control unit, and the pressure signal is “[3kPa,2kPa,2kPa]” and the acquisition timestamp of the pressure signal is “1723500003” and is sent to the main control unit.

[0048] Step 202: Whenever the main control unit receives a pressure signal and a collection timestamp, it obtains cushion sensing data based on the pressure signal conversion, obtains the account ID of the target user's currently logged-in account and the seat ID of the smart seat, creates candidate data entries based on the account ID, seat ID, cushion sensing data and collection timestamp, and stores the candidate data entries in a time series database.

[0049] Among them, the seat cushion sensing data includes a sitting action identifier and a pressure distribution matrix. The sitting action identifier is used to characterize whether the target user is sitting on top of the seat cushion. The pressure distribution matrix is ​​a matrix composed of pressure signals transmitted by the pressure sensor. There are multiple pressure distribution matrices, and each pressure distribution matrix corresponds to a timestamp, which can be understood as a linear change of the pressure distribution matrix in the time dimension.

[0050] Furthermore, the sit-stand action flag can be represented by 0 / 1. When the sit-stand action flag is 1, it indicates that the target user is sitting on the cushion; when the sit-stand action flag is 0, it indicates that the target user is standing, and not sitting on the cushion.

[0051] It should be noted that the smart seat supports the target user to create and log in a user account. The target user can create a personalized user account on the touch screen corresponding to the smart seat, and then the main control unit will generate a new account ID in the user data table corresponding to the time series database, and associate the account ID with the newly created user account. Whenever the user uses the user account to log in, the main control unit can obtain the corresponding account ID based on the user account.

[0052] It should also be noted that the data structure definition of the candidate data entry can be {account ID: unique identification number of the account ID, seat ID: unique identification number of the seat ID, seat cushion sensing data: {sitting action ID: sitting action identification number, pressure distribution matrix: [pressure signals of each pressure sensor]}, collection timestamp: collection timestamp number}.

[0053] Specifically, whenever the main control unit receives a pressure signal and an acquisition timestamp, it obtains seat cushion sensing data based on the pressure signal conversion, obtains the account ID of the target user's currently logged-in account and the seat ID of the smart seat, and then creates candidate data entries according to the data structure definition of the candidate data entries based on the account ID, seat ID, seat cushion sensing data and acquisition timestamp. Then, the main control unit establishes a connection channel with the time series database. Finally, the candidate data entries are stored in the time series database through SQL statements. Therefore, from the perspective of fine-grained timestamp granularity, the seat cushion sensing data can be stored to facilitate subsequent data processing.

[0054] In a possible implementation, when the main control unit receives a pressure signal and an acquisition timestamp, and the acquisition timestamp is "1723500004" and the pressure signal is [ ]. Then, obtain the account ID of the target user's currently logged-in account as "4b91ea7d-5302-4532" and the seat ID of the smart seat "eccf19f8-26de-419e-b217", and then create a candidate data entry according to the data structure definition of the candidate data entry {Account ID: "4b91ea7d-5302-4532", Seat ID: "eccf19f8-26de-419e-b217", Cushion Sensing Data: {Sitting Action ID: "1", Pressure Distribution Matrix: [ ]}, collection timestamp: "1723500004"}, then, the main control unit establishes a connection channel with the time series database, and finally, stores the candidate data entry in the time series database through SQL statements.

[0055] Step 203: The main control unit receives the query instruction sent by the server.

[0056] The query instruction carries a plurality of preset time ranges of different lengths.

[0057] Specifically, the main control unit establishes a connection with the server through a network protocol, and receives the query instruction sent by the server through a specific coordination port, and then parses the data body of the query instruction to obtain data corresponding to the preset time ranges of various lengths carried by the query instruction. Then, the preset time ranges of various lengths carried by the query instruction can facilitate the subsequent main control unit to match sample data entries corresponding to different preset time ranges in the time series database, thereby improving the diversity of samples.

[0058] In a possible implementation, the main control unit establishes a connection with the server through the http network protocol, receives the query instruction sent by the server through a specific network port 8080, and then parses the data body of the query instruction to obtain the data body of the query instruction carrying {time range 1: timestamp "1723500000" ~ timestamp "1723500002", time range 2: timestamp "1723500000" ~ timestamp "1723500100", time range 3: timestamp "1723500000" ~ timestamp "1723501000"}.

[0059] Step 204: the main control unit matches the collection timestamp with the preset time range among the multiple candidate data entries in the time series database, obtains the sample data entry corresponding to the preset time range, and sends the sample data entry to the server.

[0060] Specifically, the main control unit establishes a connection with a time series database that stores data. The main control unit uses an SQL query statement to match multiple candidate data entries in the time series database with a collection timestamp according to a preset time range to obtain sample data entries corresponding to the preset time range. Finally, the sample data entries are formatted and sent to the server. The main control unit can then match multiple candidate data entries in the time series database with a collection timestamp according to a preset time range to obtain sample data entries corresponding to different preset time ranges, thereby enriching the number of samples.

[0061] In a possible implementation, the main control unit establishes a connection with a time series database storing data. The main control unit matches the acquisition timestamp according to time range 1 (timestamp "1723500000" ~ timestamp "1723500002") among multiple candidate data entries in the time series database through an SQL query statement to obtain sample data entries corresponding to the preset time range [sample data entry 1,...sample data entry 3. Finally, the sample data entries [sample data entry 1, sample data entry 2, sample data entry 3] are formatted and sent to the server.

[0062] In a possible implementation, the main control unit establishes a connection with a time series database that stores data. The main control unit matches the acquisition timestamp according to time range 3 (timestamp "1723500000" to timestamp "1723501000") among multiple candidate data entries in the time series database through an SQL query statement to obtain sample data entries [sample data entry 1, ... sample data entry 1001] corresponding to the preset time range. Finally, the sample data entries [sample data entry 1, ... sample data entry 1000] are formatted and sent to the server.

[0063] Step 205: The server constructs the multiple cushion sensing data into sample time series feature data based on the acquisition timestamp, trains a cushion hardness adjustment model based on the sample time series feature data, and associates the trained cushion hardness adjustment model with the account ID and the seat ID for storage.

[0064] Specifically, the server receives multiple sample data entries from the main control unit, and then constructs multiple seat cushion sensing data into sample time series feature data based on the collection timestamps in each sample data entry. Next, the seat cushion hardness adjustment model is trained based on the sample time series feature data, and the trained seat cushion hardness adjustment model is associated with the account ID and the seat ID and stored, thereby facilitating the subsequent flexible adjustment of the seat cushion hardness based on the seat cushion hardness adjustment model. This process does not require user operation, thereby achieving a user-imperceptible effect.

[0065] In one possible implementation, the server receives [sample data entry 1, sample data entry 2, sample data entry 3] from the main control unit, where sample data entry 1 is {account ID: "4b91ea7d-5302-4532", seat ID: "eccf19f8-26de-419e-b217", cushion sensing data: {sitting action ID: "1", pressure distribution matrix: [ ]}, collection timestamp: "1723500000"}, sample data entry 2 is {account ID: "4b91ea7d-5302-4532", seat ID: "eccf19f8-26de-419e-b217", cushion sensing data: {sitting action ID: "1", pressure distribution matrix: [ ]}, collection timestamp: "1723500001"}, sample data entry 3 is {account ID: "4b91ea7d-5302-4532", seat ID: "eccf19f8-26de-419e-b217", cushion sensing data: {sitting action ID: "1", pressure distribution matrix: [ ]}, collection timestamp: "1723500002"}.

[0066] Then, based on the collection timestamp "1723500000" in sample data entry 1, the collection timestamp "1723500001" in sample data entry 2, and the collection timestamp "1723500002" in sample data entry 3, multiple seat cushion sensing data are constructed into sample time series feature data. Next, the seat cushion hardness adjustment model is trained based on the sample time series feature data, and the trained seat cushion hardness adjustment model is associated with the account identifier "4b91ea7d-5302-4532" and the seat identifier "eccf19f8-26de-419e-b217" and stored.

[0067] It can be seen that the pressure sensor of the smart seat periodically collects the pressure signal of the target user based on the preset detection frequency, and sends the pressure signal and the collection timestamp of the pressure signal to the main control unit whenever the pressure signal is collected; whenever the main control unit receives the pressure signal and the collection timestamp, it converts the pressure signal to obtain the cushion sensing data, obtains the account ID of the target user's currently logged-in account and the seat ID of the smart seat, creates candidate data entries based on the account ID, seat ID, cushion sensing data and collection timestamp, and stores the candidate data entries in the time series database. By periodically collecting pressure signals and recording corresponding collection timestamps, it can detect the cushion sensing data from the perspective of fine-grained timestamp granularity. The data is stored so that when the main control unit receives a query instruction carrying a plurality of preset time ranges of different lengths sent by the server, the main control unit can match the multiple candidate data entries in the time series database according to the preset time range and the collection timestamp, and obtain the sample data entries corresponding to the different preset time ranges, thereby constructing different training samples, increasing the number of samples, and then sending the sample data entries to the server. The server constructs the multiple cushion sensing data into sample time series feature data based on the collection timestamp, trains the cushion hardness adjustment model based on the sample time series feature data, and stores the trained cushion hardness adjustment model in association with the account ID and the seat ID, thereby effectively improving the training efficiency of the cushion hardness adjustment model.

[0068] Further, refer to Figure 3 As shown, Figure 3 An optional flow chart of constructing sample time series feature data by a server provided in an embodiment of the present disclosure. The process of constructing sample time series feature data by the server includes but is not limited to the following steps 301 to 303.

[0069] Step 301: For each acquisition time stamp, fill the corresponding sit-stand action identifier with zeros to obtain a 1×m matrix element.

[0070] Specifically, for each acquisition time stamp, the corresponding sit-stand action identifier is padded with zeros to obtain a 1×m matrix element, thereby providing a data infrastructure for subsequently integrating the sit-stand action identifier into the pressure distribution matrix.

[0071] In one possible implementation, for the acquisition timestamp "1723500000" in the sample data entry 1, the corresponding sit action identifier 1 is padded with zeros to obtain a 1×m matrix element [1, 0, 0]; for the acquisition timestamp "1723500001" in the sample data entry 2, the corresponding sit action identifier 1 is padded with zeros to obtain a 1×m matrix element [1, 0, 0]; for the acquisition timestamp "1723500002" in the sample data entry 3, the corresponding sit action identifier 1 is padded with zeros to obtain a 1×m matrix element [1, 0, 0].

[0072] In one possible implementation, for the acquisition timestamp "1723500000" in the sample data entry 1, the corresponding sit action identifier 1 is padded with zeros to obtain a 1×m matrix element [0, 0, 1]; for the acquisition timestamp "1723500001" in the sample data entry 2, the corresponding sit action identifier 1 is padded with zeros to obtain a 1×m matrix element [0, 0, 1]; for the acquisition timestamp "1723500002" in the sample data entry 3, the corresponding sit action identifier 1 is padded with zeros to obtain a 1×m matrix element [0, 0, 1].

[0073] In one possible implementation, for the acquisition timestamp "1723500000" in the sample data entry 1, the corresponding sit action identifier 1 is padded with zeros to obtain a 1×m matrix element [0, 1, 0]; for the acquisition timestamp "1723500001" in the sample data entry 2, the corresponding sit action identifier 1 is padded with zeros to obtain a 1×m matrix element [0, 0, 1]; for the acquisition timestamp "1723500002" in the sample data entry 3, the corresponding sit action identifier 1 is padded with zeros to obtain a 1×m matrix element [0, 0, 1].

[0074] Step 302: splice the matrix elements into the pressure distribution matrix with the same acquisition time stamp to obtain the target matrix.

[0075] The dimension of the target matrix is ​​(n+1)×m.

[0076] Specifically, the matrix elements are spliced ​​into the pressure distribution matrix of the same acquisition timestamp to obtain the target matrix, which can provide multi-dimensional data support for the subsequent training of the seat cushion hardness adjustment model, thereby improving the prediction accuracy of the trained seat cushion hardness adjustment model.

[0077] In a possible implementation, the [1, 0, 0] corresponding to the acquisition timestamp “1723500000” in the sample data entry 1 is concatenated to the pressure distribution matrix [ ], and get the target matrix [ ], the [1, 0, 0] corresponding to the acquisition timestamp "1723500001" in sample data entry 2 is concatenated to the pressure distribution matrix [ ], and get the target matrix [ ], the [1, 0, 0] corresponding to the acquisition timestamp "1723500002" in sample data entry 3 is concatenated to the pressure distribution matrix [ ], and get the target matrix [ ].

[0078] Step 303: construct multiple target matrices into sample time series feature data according to the order of multiple acquisition time stamps.

[0079] The dimension of the sample time series feature data is (n+1)×(m×x), where x is the number of acquisition timestamps.

[0080] Specifically, multiple acquisition timestamps are sorted in sequence, and then multiple target matrices are constructed as sample time series feature data according to the sequence of each acquisition timestamp. Then, the sample time series feature data can retain the characteristics of linear dynamic changes in time through the order of timestamps, so that the subsequent seat cushion hardness adjustment model can continue to learn the influence of the law of time changes on the seat cushion hardness during the training process, thereby improving the prediction accuracy of the seat cushion hardness adjustment model for future seat cushion hardness adjustments. At the same time, since the sample time series feature data is a continuous time series feature, it also supports the prediction of seat cushion hardness at multiple continuous future time points, thereby improving the versatility of the seat cushion hardness adjustment model.

[0081] In a possible implementation, the acquisition timestamp "1723500000" in sample data entry 1, the acquisition timestamp "1723500001" in sample data entry 2, and the acquisition timestamp "1723500002" in sample data entry 3 are sequentially arranged to obtain a time series "1723500000-1723500001-1723500002", and then the target matrix [ ]×the target matrix of acquisition timestamp “1723500001”[ ]×the target matrix of acquisition timestamp “1723500002”[ ], get the sample time series feature data[ ].

[0082] Further, in combination with step 301 to step 303, first, for each acquisition timestamp, the corresponding sit-stand action identifier is padded with zeros to obtain a 1×m matrix element, and then the matrix elements are spliced ​​into the pressure distribution matrix of the same acquisition timestamp to obtain a target matrix. Finally, multiple target matrices are constructed as sample time series feature data in the order of multiple acquisition timestamps, and then a complete pressure distribution matrix is ​​formed by splicing the sit-stand action identifier and the pressure distribution matrix of the same timestamp together, which helps to integrate data from different sources or multiple sensors to avoid data loss or incompleteness. At the same time, the time series feature data is constructed in the order of timestamps to retain the time dependency of the original data.

[0083] Reference Figure 4 As shown, Figure 4 An optional flow chart of a seat cushion hardness adjustment model for training provided in an embodiment of the present disclosure includes but is not limited to the following steps 401 to 403 during the process of training the seat cushion hardness adjustment model.

[0084] Step 401: Perform position encoding on each target matrix to obtain a position vector.

[0085] Specifically, in the process of position encoding each target matrix, each value in the target matrix is ​​compared with a preset pressure threshold. When the value is greater than or equal to the preset pressure threshold, it is determined that the target user is sitting above the position corresponding to the value, and the position is set to a significant value (such as setting 1, and other positions are set to 0). Finally, the position vector corresponding to the target matrix is ​​obtained, and the specific position of the target user can be clearly indicated by obtaining the position vector. At the same time, by encoding the position information into a vector form, it can be conveniently used as part of the input of the seat cushion hardness adjustment model, so that the seat cushion hardness adjustment model can learn the complex relationship between position and data features.

[0086] In a possible implementation, assuming that when the value in the target matrix is ​​greater than or equal to 1.5 kpa, it is determined that the target user is sitting above the position corresponding to the value. First, the target matrix [ ], the position vector of the acquisition timestamp "1723500000" is [ ]; collect the target matrix with timestamp "1723500001"[ ], the position vector of the acquisition timestamp "1723500001" is [ ]; collect the target matrix with timestamp "1723500002"[ ], the position vector of the acquisition timestamp "1723500002" is [ ]; then [ ]、[ ]as well as[ ] Based on the time series features in the sample time series feature data, tensor data processing is performed to obtain the position vector [ ].

[0087] Step 402: The sample time series feature data is fused with the position vector and input into the seat cushion hardness adjustment model for regression processing to obtain a sample gas volume distribution matrix.

[0088] Among them, each matrix element in the sample gas volume distribution matrix is ​​the gas volume of each airbag.

[0089] Specifically, the sample time series feature data is fused with the position vector to obtain the tensor data of the input model, and then the tensor data is input into the seat cushion hardness adjustment model for nonlinear regression processing to obtain the sample gas volume distribution matrix, which can accurately measure the relative influence of position factors and pressure factors on the gas volume of the airbag, thereby improving the accuracy of the gas volume prediction of each airbag.

[0090] In a possible implementation, the sample time series feature data [ ] and position vector [ ] Fusion to obtain the tensor data of the input model ,Then, the data of the input model is input into the seat cushion hardness adjustment model for regression processing to obtain the sample gas volume distribution matrix [ .

[0091] Step 403: Obtain a labeled gas volume distribution matrix, determine a training loss according to the difference between the sample gas volume distribution matrix and the labeled gas volume distribution matrix, and train a seat cushion hardness adjustment model based on the training loss.

[0092] The labeled gas quantity distribution matrix is ​​a preset known gas quantity distribution matrix, and the values ​​in the labeled gas quantity distribution matrix are all manually labeled.

[0093] Specifically, the label gas volume distribution matrix is ​​obtained, and then in each iteration of training the cushion hardness adjustment model, the sample gas volume distribution matrix is ​​calculated through the cross entropy loss function (for example: [ ) and the labeled gas volume distribution matrix (for example: [ ) to determine the loss value of the current model iterative training, and reversely update the seat cushion hardness adjustment model based on the loss value of the current iterative training, so that the seat cushion hardness adjustment model continuously learns and improves the mapping relationship between the tensor data of the input model and the sample gas volume distribution matrix of the output model in each iterative training.

[0094] It should be noted that the calculation formula of the above cross entropy loss function is:

[0095]

[0096] Where N is the number of matrix elements of the sample gas volume distribution matrix, is the first The true label of the matrix element, whose label value is 0 or 1; is the first matrix of the sample gas volume distribution matrix for the cushion hardness adjustment model. The probability that the matrix element is predicted to be the positive class (label 1).

[0097] Further, in combination with step 401 to step 403, first, the position of each target matrix in the sample time series feature data is encoded to obtain a position vector, then, the sample time series feature data and the position vector are fused and input into the seat cushion hardness adjustment model for regression processing to obtain a sample gas volume distribution matrix, and finally, the label gas volume distribution matrix is ​​obtained, and the training loss is determined according to the difference between the sample gas volume distribution matrix and the label gas volume distribution matrix. The seat cushion hardness adjustment model is trained based on the training loss, and then through the fusion of the time series feature and the position vector, the seat cushion hardness adjustment model can take into account the influence of time change and spatial position at the same time when predicting, so as to more comprehensively understand the complex relationship behind the data and improve the accuracy and reliability of the prediction. In addition, the seat cushion hardness adjustment model adopts regression processing and can be applicable to scenarios of predicting continuous variables (such as continuous changes in gas volume distribution), thereby realizing accurate prediction of future continuous changes in gas volume distribution.

[0098] Further, it should be noted that the creation of candidate data entries according to the account identification, seat identification, cushion sensing data and collection timestamp in the above step 202 also includes but is not limited to the following step a1.

[0099] Step a1: Create candidate data entries based on the account ID, seat ID, cushion sensing data, gas volume distribution matrix, and collection timestamp.

[0100] It should be noted that the data structure definition of the candidate data entry also includes the data structure field of the gas quantity distribution matrix, that is, the candidate data entry can also be {account ID: unique identification number of the account ID, seat ID: unique identification number of the seat ID, seat cushion sensing data: {sitting action ID: sitting action identification number, pressure distribution matrix: [pressure signals of each pressure sensor]}, collection timestamp: collection timestamp number, gas quantity distribution matrix: [gas quantity of each airbag]}.

[0101] Specifically, candidate data entries are created based on the account ID, seat ID, seat cushion sensing data, gas volume distribution matrix and collection timestamp, and then unified storage management of data is achieved by increasing the data dimensions of the candidate data entries. There is no need to allocate additional space in the time series database to store the corresponding gas volume distribution matrix and to store the mapping relationship between the seat cushion sensing data and the gas volume distribution matrix.

[0102] In a possible implementation, candidate data entries are created based on the account ID, seat ID, cushion sensing data, gas volume distribution matrix, and acquisition timestamp, and sample data entry 1 is obtained as {account ID: "4b91ea7d-5302-4532", seat ID: "eccf19f8-26de-419e-b217", cushion sensing data: {sitting action ID: "1", pressure distribution matrix: [ ]}, acquisition timestamp: "1723500000", gas volume distribution matrix: [ }.

[0103] In a possible implementation, candidate data entries are created based on the account ID, seat ID, cushion sensing data, gas volume distribution matrix, and acquisition timestamp, and the sample data entry 2 is obtained as {account ID: "4b91ea7d-5302-4532", seat ID: "eccf19f8-26de-419e-b217", cushion sensing data: {sitting action ID: "1", pressure distribution matrix: [ ]}, acquisition timestamp: "1723500001", gas volume distribution matrix: [ }.

[0104] In a possible implementation, candidate data entries are created based on the account ID, seat ID, cushion sensing data, gas volume distribution matrix, and acquisition timestamp, and sample data entry 3 is obtained as {account ID: "4b91ea7d-5302-4532", seat ID: "eccf19f8-26de-419e-b217", cushion sensing data: {sitting action ID: "1", pressure distribution matrix: [ ]}, acquisition timestamp: "1723500002", gas volume distribution matrix: [ }.

[0105] Further, it should be noted that the process of acquiring the label gas quantity distribution matrix in the above step 403 also includes but is not limited to the following step b1.

[0106] Step b1: Obtain a sample data entry corresponding to a target timestamp in a time series database, and use a gas quantity distribution matrix in the sample data entry as a label gas quantity distribution matrix.

[0107] The target timestamp is the next acquisition timestamp adjacent to the last acquisition timestamp within the preset time range.

[0108] Specifically, the sample data entry corresponding to the target timestamp is obtained in the time series database, and the gas quantity distribution matrix in the sample data entry is used as the label gas quantity distribution matrix. By using the actual sample data in the time series database as the label data, the efficiency of label data reading can be improved during supervised training of the model, thereby improving the training efficiency of the seat cushion hardness adjustment model.

[0109] In one possible implementation, assuming that the current preset time range to be queried is from timestamp "1723500000" to timestamp "1723500002", the target timestamp is the next acquisition timestamp "1723500003" adjacent to the last acquisition timestamp "1723500002" in the preset time range. Then, the main control unit establishes a connection channel with the time series database, and then obtains the sample data entry 4 corresponding to the target timestamp "1723500003" through a SQL statement query, which is {Account ID: "4b91ea7d-5302-4532", seat ID: "eccf19f8-26de-419e-b217", cushion sensing data: {Sitting action ID: "1", pressure distribution matrix: [ ]}, acquisition timestamp: "1723500003", gas volume distribution matrix: [ }, and finally the gas volume distribution matrix [ As the label gas amount distribution matrix.

[0110] This application also improves the reasoning process of the seat cushion hardness adjustment model. Figure 5 As shown, Figure 5 This is another optional flow chart of the data processing method for the smart seat provided in the embodiment of the present disclosure. The data processing method for the smart seat includes but is not limited to the following steps 501 to 504.

[0111] Step 501: subsequently, when the pressure sensor collects the pressure signal of the target user for the first time within a unit time, the main control unit determines the target time range with the collection timestamp of the pressure signal collected for the first time as the starting point.

[0112] The unit time may be 24 hours, 12 hours, 2 hours, or 30 minutes, and this application does not make any specific limitation.

[0113] In a possible implementation, assuming that the unit time is 24 hours, and the pressure sensor first collects the pressure signal of the target user within 24 hours, and the collection timestamp corresponding to it is "1723600000", the main control unit determines the target time range as timestamp "1723600000" to timestamp "1723600002" with the collection timestamp "1723600000" of the first pressure signal collection as the starting point.

[0114] Step 502: The main control unit matches the target time range with the acquisition timestamp among multiple candidate data entries in the time series database, obtains the target data entry corresponding to the target time range, and sends the target data entry to the server.

[0115] In a possible implementation, the main control unit matches the target time range of timestamp "1723600000" to timestamp "1723600002" with the collection timestamp in each candidate data entry through an SQL query statement in multiple candidate data entries in the time series database, and obtains the target data entry corresponding to the target time range of timestamp "1723600000" to timestamp "1723600002" (for example: sample data entry a: {account ID: "4b91ea7d-5302-4532", seat ID: "eccf19f8-26de-419e-b217", cushion sensing data: {sitting action ID: "1", pressure distribution matrix: [ ]}, acquisition timestamp: "1723600000", gas volume distribution matrix: [ }; Sample data entry b: {Account ID: "4b91ea7d-5302-4532", Seat ID: "eccf19f8-26de-419e-b217", Cushion sensing data: {Sitting action ID: "1", Pressure distribution matrix: [ ]}, acquisition timestamp: "1723600001", gas volume distribution matrix: [ }; Sample data entry c: {Account ID: "4b91ea7d-5302-4532", Seat ID: "eccf19f8-26de-419e-b217", Cushion sensing data: {Sitting action ID: "1", Pressure distribution matrix: [ ]}, acquisition timestamp: "1723600002", gas volume distribution matrix: [ }), send sample data entry a, sample data entry b, and sample data entry c to the server.

[0116] Step 503: The server calls the corresponding seat cushion hardness adjustment model based on the account ID and the seat ID, constructs multiple seat cushion sensing data into target time series feature data based on the acquisition timestamp, inputs the target time series feature data into the currently called seat cushion hardness adjustment model for regression processing, obtains the target gas quantity distribution matrix, and sends the target gas quantity distribution matrix to the main control unit.

[0117] In a possible implementation, the server queries and calls the corresponding seat cushion hardness adjustment model based on the account identifier "4b91ea7d-5302-4532" and the seat identifier "eccf19f8-26de-419e-b217" in sample data entry 1 or sample data entry 2 or sample data entry 3, and then based on the sequence of the collection timestamp in sample data entry a, the collection timestamp in sample data entry b, and the collection timestamp in sample data entry c, multiple seat cushion sensing data (respectively, the seat cushion sensing data in sample data entry a {sitting action identifier: "1", pressure distribution matrix: [ ]}, the cushion sensing data in sample data entry b {sitting action flag: "1", pressure distribution matrix: [ ]}, the cushion sensing data in sample data entry c {sitting action flag: "1", pressure distribution matrix: [ ]}) is constructed as the target time series feature data[ Then the target time series characteristic data is input into the seat cushion hardness adjustment model for regression processing to obtain the target gas volume distribution matrix [ ], and finally the target gas quantity distribution matrix [ ] is sent to the master control unit.

[0118] Step 504: The main control unit adjusts the gas volume of the multiple airbags based on the target gas volume distribution matrix.

[0119] In a possible implementation, the main control unit receives the target gas quantity distribution matrix [ ], and then adjust the gas volume of multiple airbags based on the target gas volume distribution matrix, that is, adjust the gas volume of each airbag from the gas volume distribution matrix corresponding to the current acquisition timestamp "1723600002" [ Adjust to the gas volume distribution matrix corresponding to the next future timestamp "1723600003" [ ].

[0120] At the same time, in combination with step 501 to step 504, first, when the pressure sensor collects the pressure signal of the target user for the first time within a unit time, the main control unit determines the target time range with the collection timestamp of the first collection of the pressure signal as the starting point. Then, the main control unit matches the multiple candidate data entries in the time series database according to the target time range and the collection timestamp, obtains the target data entry corresponding to the target time range, and sends the target data entry to the server. Next, the server calls the corresponding seat cushion hardness adjustment model based on the account identifier and the seat identifier, constructs multiple seat cushion sensing data into target time series feature data based on the collection timestamp, inputs the target time series feature data into the currently called seat cushion hardness adjustment model for regression processing, obtains the target gas volume distribution matrix, and sends the target gas volume distribution matrix to the main control unit. Finally, the main control unit adjusts the gas volume of multiple airbags based on the target gas volume distribution matrix, and then can realize automatic and user-imperceptible seat airbag adjustment through the main control unit's automatic collection of pressure signals and matching of target data entry processing logic, the server's automatic prediction of the target gas volume distribution matrix processing logic, and the data interaction logic between the main control unit and the server.

[0121] Reference Figure 6 As shown, Figure 6 An optional flow chart for determining a target time range is provided in an embodiment of the present disclosure. The process of determining the target time range includes but is not limited to the following step 601.

[0122] Step 601: The main control unit periodically determines a target time range, taking the acquisition timestamp of the first acquisition of the pressure signal as the starting point.

[0123] In one possible implementation, the main control unit takes the acquisition timestamp "1723600000" when the pressure signal is first acquired as the starting point, and periodically determines the target time range (for example: the target time range is timestamp "1723600000" ~ timestamp "1723600002", the target time range is timestamp "1723600000" ~ timestamp "1723600100"; the target time range is timestamp "1723600000" ~ timestamp "1723601000", etc.), and then can continuously obtain different target time ranges to provide a diverse data base source for subsequent data processing.

[0124] Reference Figure 7 As shown, Figure 7 An optional flow chart of the main control unit adjusting the gas volume of the airbag provided in the embodiment of the present disclosure, wherein the process of the main control unit adjusting the gas volume of the airbag includes but is not limited to the following steps 701 to 702.

[0125] Step 701: The main control unit determines a target pressure sensor that collects pressure signals from a plurality of pressure sensors, and determines an airbag corresponding to the target pressure sensor as a first target airbag.

[0126] In a possible implementation, at the timestamp “1723600002”, the pressure distribution matrix obtained by the main control unit from the multiple pressure sensors is [ ] According to the pressure distribution matrix, all pressure sensors collect pressure signals, and then the main control unit determines all pressure sensors as target pressure sensors, and finally determines the airbags corresponding to all pressure sensors as the first target airbags.

[0127] In a possible implementation, at the timestamp “1723600005”, the pressure distribution matrix obtained by the main control unit from the multiple pressure sensors is [ ], according to the pressure distribution matrix, it can be known that the pressure sensor in the second row and second column, the pressure sensor in the second row and third column, the pressure sensor in the third row and first column, the pressure sensor in the third row and second column, and the pressure sensor in the third row and third column collect pressure signals, and then the main control unit determines the pressure sensor in the second row and second column, the pressure sensor in the second row and third column, the pressure sensor in the third row and first column, the pressure sensor in the third row and second column, and the pressure sensor in the third row and third column as target pressure sensors, and finally determines the airbag corresponding to the pressure sensor in the second row and second column as the first target airbag, the airbag corresponding to the pressure sensor in the second row and third column as the first target airbag, the airbag corresponding to the pressure sensor in the third row and first column as the first target airbag, the airbag corresponding to the pressure sensor in the third row and first column as the first target airbag, the airbag corresponding to the pressure sensor in the third row and second column as the first target airbag, and the airbag corresponding to the pressure sensor in the third row and third column as the first target airbag.

[0128] Step 702: The main control unit preferentially adjusts the gas volume of the first target airbag during the process of adjusting the gas volume of multiple airbags based on the target gas volume distribution matrix.

[0129] In one possible implementation, in the timestamp "1723600002", the main control unit preferentially adjusts the gas volume of the first target airbag in the process of adjusting the gas volume of multiple airbags based on the target gas volume distribution matrix, that is, preferentially adjusts the gas volume of the airbags corresponding to all pressure sensors.

[0130] In one possible implementation, in the timestamp "1723600005", the main control unit gives priority to adjusting the gas amount of the first target airbag in the process of adjusting the gas amount of multiple airbags based on the target gas amount distribution matrix, that is, giving priority to adjusting the gas amount of the airbag corresponding to the pressure sensor in the second row and second column, the gas amount of the airbag corresponding to the pressure sensor in the second row and third column, the gas amount of the airbag corresponding to the pressure sensor in the third row and first column, the gas amount of the airbag corresponding to the pressure sensor in the third row and second column, and the gas amount of the airbag corresponding to the pressure sensor in the third row and third column.

[0131] At the same time, in combination with step 701 to step 702, first, the main control unit determines the target pressure sensor that collects the pressure signal from multiple pressure sensors, and determines the airbag corresponding to the target pressure sensor as the first target airbag. Then, the main control unit gives priority to adjusting the gas volume of the first target airbag in the process of adjusting the gas volume of multiple airbags based on the target gas volume distribution matrix, thereby improving the adjustment efficiency of the seat cushion airbag by the main control unit, so that the seat cushion of the smart seat can be quickly adjusted to a hardness that the user feels comfortable in a short time, thereby improving the intelligence level of the smart seat.

[0132] Reference Figure 8 As shown, Figure 8 An optional flow chart for preferentially adjusting the gas volume of the first target airbag is provided in an embodiment of the present disclosure. The process of preferentially adjusting the gas volume of the first target airbag includes but is not limited to the following steps 801 to 803.

[0133] Step 801: When adjusting the gas volume of multiple airbags based on the target gas volume distribution matrix, the main control unit sorts multiple first target airbags according to the gas volume in each first target airbag to obtain a target adjustment order.

[0134] In a possible implementation, at the timestamp "1723600002", the main control unit adjusts the gas volume of multiple airbags based on the target gas volume distribution matrix according to the gas volume of the airbag corresponding to the pressure sensor in the first row and the first column. , the gas volume of the airbag corresponding to the pressure sensor in the first row and second column , the gas volume of the airbag corresponding to the pressure sensor in the first row and third column , the gas volume of the airbag corresponding to the pressure sensor in the second row and the first column , the gas volume of the airbag corresponding to the pressure sensor in the second row and second column , the gas volume of the airbag corresponding to the pressure sensor in the second row and third column , the gas volume of the airbag corresponding to the pressure sensor in the third row and the first column , the gas volume of the airbag corresponding to the pressure sensor in the third row and second column And the gas volume of the airbag corresponding to the pressure sensor in the third row and third column , sort the sizes of the first target airbags and get the target adjustment order [the first target airbag in the first row and the first column ( ), the first target airbag in the first row and second column ( ), the first target airbag in the first row and third column ( ), the first target airbag in the second row and the first column ( ), the first target airbag in the second row and second column ( ), the first target airbag in the third row and first column ( ), the first target airbag in the third row and second column ( ), the first target airbag in the second row and third column ( ), the first target airbag in the third row and third column ( )].

[0135] In a possible implementation, at the timestamp "1723600005", the main control unit adjusts the gas volume of the multiple airbags based on the target gas volume distribution matrix according to the gas volume of the airbag corresponding to the pressure sensor in the second row and the second column. , the gas volume of the airbag corresponding to the pressure sensor in the second row and third column , the gas volume of the airbag corresponding to the pressure sensor in the third row and first column , the gas volume of the airbag corresponding to the pressure sensor in the third row and second column And the gas volume of the airbag corresponding to the pressure sensor in the third row and third column , sort the sizes of the first target airbags and get the target adjustment order [the first target airbag in the third row and the first column ( ), the first target airbag in the second row and second column ( ), the first target airbag in the third row and second column ( ), the first target airbag in the second row and third column ( ), the first target airbag in the third row and third column ( )].

[0136] Step 802: Determine the first target airbag located between the first ranking and the second ranking in the target adjustment order as the second target airbag.

[0137] Among them, the first ranking is between the first ranking and the median ranking in the target adjustment order, and the second ranking is between the median ranking and the last ranking in the target adjustment order, which can be set according to actual needs.

[0138] In a possible implementation, in the timestamp "1723600002", according to the target adjustment order [the first target airbag in the first row and the first column ( ), the first target airbag in the first row and second column ( ), the first target airbag in the first row and third column ( ), the first target airbag in the second row and the first column ( ), the first target airbag in the second row and second column ( ), the first target airbag in the third row and first column ( ), the first target airbag in the third row and second column ( ), the first target airbag in the second row and third column ( ), the first target airbag in the third row and third column ( )] It can be seen that the first target airbag in the first ranking is the first target airbag in the first row and the first column, the first target airbag in the middle ranking is the first target airbag in the second row and the second column, and the first target airbag in the last ranking is the first target airbag in the third row and the third column. Therefore, the first target airbag in the first ranking is one of the first target airbag in the first row and the second column, the first target airbag in the first row and the third column, and the first target airbag in the second row and the first column; the first target airbag in the second ranking is one of the first target airbag in the third row and the first column, the first target airbag in the third row and the second column, and the first target airbag in the second row and the third column. That is, the first target airbag between the first ranking and the second ranking includes the first target airbag in the second row and the second column, the first target airbag in the first row and the third column, the first target airbag in the second row and the first column, the first target airbag in the third row and the first column, and then the first target airbag between the first ranking and the second ranking is determined as the second target airbag.

[0139] In a possible implementation, in the timestamp "1723600005", according to the target adjustment order [the first target airbag in the third row and the first column ( ), the first target airbag in the second row and second column ( ), the first target airbag in the third row and second column ( ), the first target airbag in the second row and third column ( ), the first target airbag in the third row and third column ( )] It can be seen that the first target airbag ranked first is the first target airbag in the third row and first column, the first target airbag ranked in the middle is the first target airbag in the third row and second column, and the first target airbag ranked last is the first target airbag in the third row and third column. Therefore, the first target airbag ranked first is the first target airbag in the second row and second column; the first target airbag ranked second is the first target airbag in the second row and third column, and the first target airbag between the first and second rankings is the first target airbag in the third row and second column. Then, the first target airbag between the first and second rankings is determined as the second target airbag.

[0140] Step 803: In the process of preferentially adjusting the gas volume of the first target airbag, preferentially adjusting the gas volume of the second target airbag.

[0141] In one possible implementation, in timestamp "1723600002", when the main control unit executes the step of preferentially adjusting the gas volume of the first target airbag, the gas volume of the second target airbag (i.e., the first target airbag in the second row and second column, or the first target airbag in the first row and third column, or the first target airbag in the second row and first column, or the first target airbag in the third row and first column, or the first target airbag in the third row and second column) is preferentially adjusted.

[0142] In a possible implementation, in the timestamp "1723600005", when the main control unit executes the step of preferentially adjusting the gas volume of the first target airbag, the gas volume of the second target airbag (ie, the first target airbag in the third row and second column) is preferentially adjusted.

[0143] At the same time, combined with step 801 to step 803, first, in the process of adjusting the gas volume of multiple airbags based on the target gas volume distribution matrix, the main control unit sorts the multiple first target airbags according to the size of the gas volume in each first target airbag to obtain a target adjustment order, and then, determines the first target airbag located between the first ranking and the second ranking in the target adjustment order as the second target airbag, and finally, in the process of preferentially adjusting the gas volume of the first target airbag, the gas volume of the second target airbag is preferentially adjusted, thereby improving the smoothness of the adjustment of the seat cushion airbag by the main control unit, so that the adjustment of the seat cushion airbag will not be too abrupt, thereby improving the intelligence level of the smart seat.

[0144] Take a practical example to illustrate, refer to Fig. 9 As shown, Fig. 9 The gas volume distribution diagram of the airbag of the smart seat provided in the embodiment of the present disclosure is as follows: Fig. 9 It can be seen that the plurality of airbags are respectively assigned airbag numbers (airbag No. 1, airbag No. 2, ..., airbag No. 20), wherein the airbag numbers are determined based on the positions of the airbags in the matrix distribution. Fig. 9It can be determined that the first target airbags include airbag No. 2, airbag No. 3, airbag No. 4, airbag No. 7, airbag No. 8, airbag No. 9, airbag No. 10, airbag No. 12, airbag No. 13, airbag No. 14, airbag No. 15, airbag No. 17, airbag No. 18, airbag No. 19, and airbag No. 20. Then, the target adjustment order obtained by sorting the gas volume in the first target airbags from large to small is [airbag No. 8, airbag No. 9] airbag, airbag No. 13, airbag No. 14, airbag No. 10, airbag No. 15, airbag No. 3, airbag No. 4, airbag No. 7, airbag No. 19, airbag No. 12, airbag No. 18, airbag No. 17, airbag No. 2, airbag No. 20], then the first target airbag ranked first is airbag No. 8, the first target airbag ranked last is airbag No. 20, the middle ranking is airbag No. 4, and the first ranked airbags are airbags No. 9 and No. 13. , No. 14 airbag, No. 10 airbag, No. 15 airbag, No. 3 airbag one of the six, the second ranking airbag includes No. 7 airbag, No. 19 airbag, No. 12 airbag, No. 18 airbag, No. 17 airbag, No. 2 airbag one of the six, then the second target airbag can include No. 3 airbag, No. 4 airbag and No. 7 airbag, can also include No. 15 airbag, No. 3 airbag, No. 4 airbag and No. 7 airbag, can also include No. 15 airbag , airbag No. 3, airbag No. 4, airbag No. 7, airbag No. 19 and airbag No. 12, and may also include airbag No. 10, airbag No. 15, airbag No. 3, airbag No. 4, airbag No. 7, airbag No. 19, airbag No. 12 and airbag No. 18, and may also include airbag No. 10, airbag No. 15, airbag No. 3, airbag No. 4, airbag No. 7, airbag No. 19, airbag No. 12, airbag No. 18, airbag No. 17 and airbag No. 2.

[0145] Reference Fig.10 As shown, Fig.10 An optional flow chart for preferentially adjusting the gas volume of the second target airbag is provided in an embodiment of the present disclosure. The process of preferentially adjusting the gas volume of the second target airbag includes but is not limited to the following steps 1001 to 1003.

[0146] Step 1001: Determine the first target airbag that is before the first ranking as the third target airbag.

[0147] Specifically, the above Fig. 9 Taking the gas volume distribution of the airbags of the smart seat as an example, the first target airbag in the first ranking is airbag No. 10, and the first target airbags before the first ranking include airbag No. 8, airbag No. 9, airbag No. 13 and airbag No. 14, that is, the third target airbag includes airbag No. 8, airbag No. 9, airbag No. 13 and airbag No. 14.

[0148] Step 1002: When the hardness of the seat cushion needs to be increased, multiple third target airbags are traversed in order of gas volume from large to small. For the third target airbag currently traversed, the second target airbag adjacent to the third target airbag is determined according to the airbag number, and the second target airbag adjacent to the third target airbag currently traversed is preferentially configured in order of distance from the backrest of the smart seat from large to small, until the second target airbags adjacent to each third target airbag are configured, and then the remaining second target airbags are configured.

[0149] Specifically, the above Fig. 9 Taking the gas volume distribution of the airbag of the smart seat as an example, when the cushion hardness needs to be improved, multiple third target airbags are traversed in the order of gas volume from large to small. For example, if the third target airbag currently traversed is airbag No. 8, the second target airbag adjacent to the third target airbag is determined according to the airbag number (airbag No. 3 and airbag No. 7. If the adjacent second target airbag is also the third target airbag, the adjacent second target airbag is removed, that is, airbag No. 13 and airbag No. 9 are removed). Then, airbag No. 3 and airbag No. 7 are placed in the order corresponding to the backrest of the smart seat. The second target airbag farthest from the backrest of the smart seat is configured first (i.e., airbag No. 7), and then the next second target airbag (i.e., airbag No. 3) is configured accordingly. Finally, the next third target airbag (i.e., airbag No. 9) is traversed and the above configuration is repeated, and so on, until the second target airbags adjacent to each third target airbag (i.e., airbag No. 8, airbag No. 9, airbag No. 13, and airbag No. 14) are configured, and then the remaining second target airbags (e.g., airbag No. 17, airbag No. 2) are configured.

[0150] Step 1003: When the hardness of the seat cushion needs to be lowered, multiple third target airbags are traversed in order of gas volume from small to large. For the third target airbag currently traversed, the second target airbag adjacent to the third target airbag is determined according to the airbag number, and the second target airbag adjacent to the third target airbag currently traversed is preferentially configured in order of distance from the backrest of the smart seat from large to small, until the second target airbags adjacent to each third target airbag are configured, and then the remaining second target airbags are configured.

[0151] Specifically, the above Fig. 9Taking the gas volume distribution of the airbag of the smart seat as an example, when the hardness of the seat cushion needs to be reduced, multiple third target airbags are traversed in order of gas volume from small to large. For example, if the third target airbag currently traversed is airbag No. 14, the second target airbags (airbags No. 15 and airbags No. 19) adjacent to the third target airbag are determined according to the airbag number, and then airbags No. 15 and airbags No. 19 are configured in order of the distance from the backrest of the smart seat from large to small, with the second target airbag (i.e., airbag No. 19) farthest from the backrest of the smart seat being configured first, and then the next second target airbag (i.e., airbag No. 15) is configured accordingly. Finally, the next third target airbag (airbag No. 13) is traversed and the above configuration is repeated, and so on, until the second target airbags adjacent to each third target airbag (i.e., airbag No. 14, airbag No. 13, airbag No. 9, and airbag No. 8) are configured, and then the remaining second target airbags (such as airbag No. 17 and airbag No. 2) are configured.

[0152] At the same time, in combination with step 1001 to step 1003, first, the first target airbag located before the first ranking is determined as the third target airbag. When the hardness of the seat cushion needs to be increased, multiple third target airbags are traversed in the order of gas volume from large to small. For the third target airbag currently traversed, the second target airbag adjacent to the third target airbag is determined according to the airbag number, and the second target airbag adjacent to the third target airbag currently traversed is preferentially configured in the order of the distance from the backrest of the smart seat from large to small. After the second target airbags adjacent to each third target airbag are configured, the remaining other second target airbags are configured. When the hardness of the seat cushion needs to be reduced, the second target airbag is configured in the order of gas volume from large to small. Multiple third target airbags are traversed in order from small to large in volume. For the third target airbag currently traversed, the second target airbag adjacent to the third target airbag is determined according to the airbag number, and the second target airbag adjacent to the third target airbag currently traversed is preferentially configured in order from large to small in distance from the backrest of the smart seat, until the second target airbags adjacent to each third target airbag are configured, and then the remaining second target airbags are configured, thereby improving the main control unit's ability to adjust the seat cushion airbag without perception, and further improving the smoothness of the main control unit's adjustment of the seat cushion airbag, so that the seat cushion airbag does not appear abrupt during the adjustment process, thereby improving the comfort of the smart seat.

[0153] It is to be understood that, although the steps in the above-mentioned various flow charts are shown in sequence according to the indication of the arrows, these steps are not necessarily performed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in the present embodiment, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, at least a portion of the steps in the above-mentioned flow charts can include a plurality of steps or a plurality of stages, and these steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of the steps or stages in other steps or other steps.

[0154] On the other hand, the present disclosure also provides a smart seat, referring to Fig.11 , Fig.11 This is an optional structural diagram of a data processing system of a smart seat provided in an embodiment of the present disclosure. The data processing system 1100 is provided with a smart seat 1110 and a server 1120. The smart seat 1110 is provided with a main control unit 1111 and a pressure sensor 1112:

[0155] The pressure sensor 1112 is used to periodically collect the pressure signal of the target user based on a preset detection frequency, and whenever a pressure signal is collected, the pressure signal and the collection timestamp of the pressure signal are sent to the main control unit 1111;

[0156] The main control unit 1111 is used to obtain cushion sensing data based on the pressure signal conversion whenever receiving the pressure signal and the acquisition timestamp, obtain the account ID of the target user's currently logged-in account and the seat ID of the smart seat 1110, create candidate data entries according to the account ID, the seat ID, the cushion sensing data and the acquisition timestamp, and store the candidate data entries in the time series database;

[0157] The main control unit 1111 is also used to receive a query instruction sent by the server 1120, wherein the query instruction carries a plurality of preset time ranges of different lengths;

[0158] The main control unit 1111 is also used to match the collection timestamp with the preset time range among the multiple candidate data entries in the time series database, obtain the sample data entry corresponding to the preset time range, and send the sample data entry to the server 1120;

[0159] Server 1120 is used to construct multiple seat cushion sensing data into sample time series feature data based on the collection timestamp, train a seat cushion hardness adjustment model based on the sample time series feature data, and store the trained seat cushion hardness adjustment model in association with the account identifier and the seat identifier.

[0160] The data processing system 1100 of the above-mentioned smart seat and the data processing method of the above-mentioned smart seat are based on the same inventive concept. The data processing process of the above-mentioned data processing system 1100 describes the data processing method of the smart seat of the embodiment of the present disclosure. The pressure sensor of the smart seat periodically collects the pressure signal of the target user based on a preset detection frequency. Whenever the pressure signal is collected, the pressure signal and the collection timestamp of the pressure signal are sent to the main control unit; whenever the main control unit receives the pressure signal and the collection timestamp, the cushion sensing data is obtained based on the pressure signal conversion, the account ID of the target user's currently logged-in account and the seat ID of the smart seat are obtained, and candidate data entries are created according to the account ID, seat ID, cushion sensing data and collection timestamp, and the candidate data entries are stored in the time series database, and the main control unit is periodically By collecting pressure signals and recording corresponding collection timestamps, it is possible to store cushion sensing data from the perspective of refined timestamp granularity, so that when the main control unit receives a query instruction sent by the server carrying a plurality of preset time ranges of different lengths, the main control unit can match the collection timestamps according to the preset time ranges in multiple candidate data entries in the time series database, and obtain sample data entries corresponding to different preset time ranges, thereby constructing different training samples, increasing the number of samples, and then sending the sample data entries to the server. The server constructs multiple cushion sensing data into sample time series feature data based on the collection timestamp, trains a cushion hardness adjustment model based on the sample time series feature data, and stores the trained cushion hardness adjustment model in association with the account ID and the seat ID, thereby effectively improving the training efficiency of the cushion hardness adjustment model.

[0161] The embodiments of the present disclosure also provide a computer-readable storage medium, which is used to store a computer program, and the computer program is used to execute the data processing method of the smart seat of each of the aforementioned embodiments.

[0162] The embodiment of the present disclosure also provides a computer program product, which includes a computer program stored in a computer-readable storage medium. A processor of a computer device reads the computer program from the computer-readable storage medium, and the processor executes the computer program, so that the computer device executes the data processing method for implementing the above-mentioned smart seat.

[0163] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate to describe the embodiments of the present disclosure, such as being able to be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0164] It should be understood that in the present disclosure, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0165] It should be understood that in the description of the embodiments of the present disclosure, the meaning of multiple (or multiple items) is more than two, greater than, less than, exceed, etc. are understood as not including the number, and above, below, within, etc. are understood as including the number.

[0166] In the several embodiments provided in the present disclosure, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0167] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0168] In addition, each functional unit in each embodiment of the present disclosure may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0169] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present disclosure is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the various embodiments of the present disclosure. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store program codes.

[0170] It should also be understood that the various implementations provided in the embodiments of the present disclosure can be combined arbitrarily to achieve different technical effects.

[0171] The above is a specific description of the preferred implementation of the present disclosure, but the present disclosure is not limited to the above-mentioned implementation mode. Technical personnel familiar with the field can also make various equivalent deformations or substitutions under the shared conditions without violating the spirit of the present disclosure. These equivalent deformations or substitutions are all included in the scope defined by the claims of the present disclosure.

Claims

1. A data processing method for a smart seat, characterized in that: The smart seat is provided with a main control unit and a pressure sensor, and the data processing method comprises: The pressure sensor periodically collects the pressure signal of the target user based on a preset detection frequency, and whenever the pressure signal is collected, sends the pressure signal and the collection timestamp of the pressure signal to the main control unit; Whenever the main control unit receives the pressure signal and the acquisition timestamp, the seat cushion sensing data is converted based on the pressure signal, the account identifier of the account currently logged in by the target user and the seat identifier of the smart seat are obtained, candidate data entries are created according to the account identifier, the seat identifier, the seat cushion sensing data and the acquisition timestamp, and the candidate data entries are stored in a time series database; The main control unit receives a query instruction sent by the server, wherein the query instruction carries a plurality of preset time ranges of different lengths; The main control unit matches the collection timestamp with the preset time range among the plurality of candidate data entries in the time series database, obtains a sample data entry corresponding to the preset time range, and sends the sample data entry to the server; The server constructs the plurality of seat cushion sensing data into sample time series feature data based on the acquisition timestamp, trains a seat cushion hardness adjustment model based on the sample time series feature data, and associates the trained seat cushion hardness adjustment model with the account identifier and the seat identifier for storage; The cushion sensing data includes a sitting action identifier and a pressure distribution matrix, the dimension of the pressure distribution matrix is ​​n×m, n and m are both positive integers, and n×m is the number of the pressure sensors. The server constructs the plurality of cushion sensing data into sample time series feature data based on the acquisition timestamp, including: For each of the acquisition timestamps, the corresponding sit-stand action identifier is padded with zeros to obtain a 1×m matrix element; splicing the matrix elements into the pressure distribution matrix of the same acquisition timestamp to obtain a target matrix, wherein the dimension of the target matrix is ​​(n+1)×m; Constructing the plurality of target matrices into sample time series feature data according to the order of the plurality of acquisition time stamps, wherein the dimension of the sample time series feature data is (n+1)×(m×x), wherein x is the number of the acquisition time stamps; The seat cushion of the smart seat is provided with a plurality of airbags distributed in a matrix, and the seat cushion hardness adjustment model is trained based on the sample time series feature data, including: Performing position encoding on each of the target matrices to obtain a position vector; The sample time series feature data is fused with the position vector and then input into the cushion hardness adjustment model for regression processing to obtain a sample gas volume distribution matrix, wherein each matrix element in the sample gas volume distribution matrix is ​​the gas volume of each airbag; A labeled gas volume distribution matrix is ​​obtained, a training loss is determined according to a difference between the sample gas volume distribution matrix and the labeled gas volume distribution matrix, and a seat cushion hardness adjustment model is trained based on the training loss.

2. The data processing method according to claim 1, characterized in that: The seat cushion of the smart seat is also provided with a gas volume sensor corresponding to each of the airbags, and the gas volume sensor is used to periodically collect the gas volume distribution matrix of the plurality of airbags based on the detection frequency; The step of creating a candidate data entry according to the account identifier, the seat identifier, the cushion sensing data, and the acquisition timestamp includes: Creating a candidate data entry according to the account identifier, the seat identifier, the seat cushion sensing data, the gas volume distribution matrix, and the acquisition timestamp; The step of obtaining a labeled gas quantity distribution matrix includes: The sample data entry corresponding to the target timestamp is obtained in the time series database, and the gas quantity distribution matrix in the sample data entry is used as the label gas quantity distribution matrix, wherein the target timestamp is the next acquisition timestamp adjacent to the last acquisition timestamp within the preset time range.

3. The data processing method according to claim 1, characterized in that: The seat cushion of the smart seat is provided with a plurality of airbags distributed in a matrix, and the data processing method further comprises: Subsequently, when the pressure sensor collects the pressure signal of the target user for the first time within a unit time, the main control unit periodically determines the target time range with the collection timestamp of the pressure signal collected for the first time as the starting point; The main control unit matches the multiple candidate data entries in the time series database with the acquisition timestamp according to the target time range, obtains the target data entry corresponding to the target time range, and sends the target data entry to the server; The server calls the corresponding seat cushion hardness adjustment model based on the account identifier and the seat identifier, constructs the plurality of seat cushion sensing data into target time series feature data based on the acquisition timestamp, inputs the target time series feature data into the currently called seat cushion hardness adjustment model for regression processing, obtains a target gas volume distribution matrix, and sends the target gas volume distribution matrix to the main control unit; The main control unit adjusts the gas amounts of the plurality of airbags based on the target gas amount distribution matrix.

4. The data processing method according to claim 3, characterized in that: The number of the pressure sensors is the same as the number of the airbags, the plurality of pressure sensors are configured in one-to-one correspondence with the plurality of airbags, and the main control unit adjusts the gas volume of the plurality of airbags based on the target gas volume distribution matrix, including: The main control unit determines a target pressure sensor that collects pressure signals from the multiple pressure sensors, and determines the airbag corresponding to the target pressure sensor as a first target airbag; The main control unit preferentially adjusts the gas volume of the first target airbag during the process of adjusting the gas volumes of the plurality of airbags based on the target gas volume distribution matrix.

5. The data processing method according to claim 4, characterized in that: The main control unit preferentially adjusts the gas volume of the first target airbag in a process of adjusting the gas volume of the plurality of airbags based on the target gas volume distribution matrix, including: In the process of adjusting the gas volume of the plurality of airbags based on the target gas volume distribution matrix, the main control unit sorts the plurality of first target airbags according to the gas volume in each of the first target airbags to obtain a target adjustment order; Determine the first target airbag located between a first ranking and a second ranking in the target adjustment sequence as a second target airbag, wherein the first ranking is located between a first ranking and a median ranking in the target adjustment sequence, and the second ranking is located between the median ranking and a last ranking in the target adjustment sequence; In the process of preferentially adjusting the gas volume of the first target airbag, preferentially adjusting the gas volume of the second target airbag.

6. The data processing method according to claim 5, characterized in that: The plurality of airbags are respectively configured with airbag numbers, the airbag numbers are determined based on the positions of the airbags in the matrix distribution, the target adjustment sequence is obtained by sorting the gas volume in the first target airbags in descending order, and the gas volume of the second target airbag is preferentially adjusted, including: determining the first target airbag located before the first ranking as the third target airbag; When the cushion hardness needs to be improved, multiple third target airbags are traversed in order of gas volume from large to small. For the third target airbag currently traversed, the second target airbag adjacent to the third target airbag is determined according to the airbag number, and the second target airbag adjacent to the third target airbag currently traversed is preferentially configured in order of distance from the backrest of the smart seat from large to small, until the second target airbags adjacent to each of the third target airbags are configured, and then the remaining second target airbags are configured; When the hardness of the seat cushion needs to be lowered, multiple third target airbags are traversed in order from small to large gas volume. For the third target airbag currently traversed, the second target airbag adjacent to the third target airbag is determined according to the airbag number, and the second target airbag adjacent to the third target airbag currently traversed is preferentially configured in order from large to small distance from the backrest of the smart seat, until the second target airbags adjacent to each of the third target airbags are configured, and then the remaining second target airbags are configured.

7. A data processing system for an intelligent seat, characterized in that: It includes a smart seat and a server, wherein the smart seat is provided with a main control unit and a pressure sensor; The pressure sensor is used to periodically collect the pressure signal of the target user based on a preset detection frequency, and whenever the pressure signal is collected, the pressure signal and the collection timestamp of the pressure signal are sent to the main control unit; The main control unit is used to convert the pressure signal into cushion sensing data based on the pressure signal whenever it receives the pressure signal and the acquisition timestamp, obtain the account identifier of the account currently logged in by the target user and the seat identifier of the smart seat, create candidate data entries according to the account identifier, the seat identifier, the cushion sensing data and the acquisition timestamp, and store the candidate data entries in a time series database; The main control unit is also used to receive a query instruction sent by the server, wherein the query instruction carries a plurality of preset time ranges of different lengths; The main control unit is further used to match the collection timestamp with the preset time range among the plurality of candidate data entries in the time series database, obtain a sample data entry corresponding to the preset time range, and send the sample data entry to the server; The server is used to construct the plurality of seat cushion sensing data into sample time series feature data based on the acquisition timestamp, train a seat cushion hardness adjustment model based on the sample time series feature data, and associate the trained seat cushion hardness adjustment model with the account identifier and the seat identifier for storage; The cushion sensing data includes a sitting action identifier and a pressure distribution matrix, the dimension of the pressure distribution matrix is ​​n×m, n and m are both positive integers, and n×m is the number of the pressure sensors. The server constructs the plurality of cushion sensing data into sample time series feature data based on the acquisition timestamp, including: For each of the acquisition timestamps, the corresponding sit-stand action identifier is padded with zeros to obtain a 1×m matrix element; splicing the matrix elements into the pressure distribution matrix of the same acquisition timestamp to obtain a target matrix, wherein the dimension of the target matrix is ​​(n+1)×m; Constructing the plurality of target matrices into sample time series feature data according to the order of the plurality of acquisition time stamps, wherein the dimension of the sample time series feature data is (n+1)×(m×x), wherein x is the number of the acquisition time stamps; The seat cushion of the smart seat is provided with a plurality of airbags distributed in a matrix, and the seat cushion hardness adjustment model is trained based on the sample time series feature data, including: Performing position encoding on each of the target matrices to obtain a position vector; The sample time series feature data is fused with the position vector and then input into the cushion hardness adjustment model for regression processing to obtain a sample gas volume distribution matrix, wherein each matrix element in the sample gas volume distribution matrix is ​​the gas volume of each airbag; A labeled gas volume distribution matrix is ​​obtained, a training loss is determined according to a difference between the sample gas volume distribution matrix and the labeled gas volume distribution matrix, and a seat cushion hardness adjustment model is trained based on the training loss.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the data processing method of the smart seat according to any one of claims 1 to 6 is implemented.

Citation Information

Patent Citations

  • Seat hardness adjusting system and seat hardness adjusting method

    CN115139866A

  • Hardness adjusting seat and seat hardness adjusting method

    CN117445770A