Human motion state recognition method, device, equipment and storage medium

Abstract

This application provides a method, apparatus, device, and storage medium for human motion state recognition. The method includes: acquiring multiple sample data sets and corresponding candidate motion state labels; obtaining sample partition pressure values ​​for each foot detection area according to preset pressure calculation rules and sample pressure data packets; determining sample feature values ​​for each foot detection area according to preset pressure thresholds and the pressure values ​​of each sample partition; determining the motion mode of any sample data set based on the corresponding sample feature values; acquiring a target data set; determining the target motion mode of the target data set within the motion modes based on the target data set; and determining the target motion state label of the target data set from the candidate motion state labels. This application can improve recognition efficiency, ensure recognition accuracy, and enhance the portability and practicality of wearable devices.

Description

Technical Field

[0001] This application relates to, but is not limited to, the field of motion state recognition technology, and in particular to a method, device, equipment, and storage medium for human motion state recognition. Background Technology

[0002] With the increasing intelligence of life, human motion state recognition is becoming more and more popular in the field of smart wearable devices.

[0003] Currently, trained machine learning models are typically used to identify human motion states. Images to be identified, collected by optical devices, or data to be identified from multiple sensors are input into the machine learning model to identify human motion states. However, machine learning models require a large amount of task-related datasets for training, which is time-consuming and has low recognition efficiency. Moreover, optical devices need to operate in a working environment that meets optical measurement standards. If the working environment does not meet optical measurement standards, the image acquisition effect of the optical devices will be poor, resulting in low accuracy in identifying human motion states. In addition, the more types and numbers of sensors in wearable devices, the more complex the structure of the wearable device, resulting in poor portability and practicality. Summary of the Invention

[0004] The following is an overview of the subject matter described in detail herein. This overview is not intended to limit the scope of the claims.

[0005] This application provides a method, apparatus, device, and storage medium for recognizing human motion states, which can improve recognition efficiency, ensure recognition accuracy, and enhance the portability and practicality of wearable devices.

[0006] To achieve the above objectives, a first aspect of this application proposes a method for human motion state recognition, comprising: acquiring multiple sample data sets and corresponding candidate motion state labels, wherein the sample data sets include multiple sample pressure data packets, the multiple sample pressure data packets in the sample data sets are arranged in chronological order, the sample pressure data packets include pressure data of multiple plantar detection areas, and the candidate motion state labels are used to characterize the human motion state category information of the corresponding sample data sets; obtaining sample partition pressure values ​​of each plantar detection area according to a preset pressure calculation rule and the sample pressure data packets; determining sample feature values ​​of each plantar detection area according to a preset pressure threshold and the pressure values ​​of each sample partition; determining the motion mode of any sample data set according to the sample feature values ​​corresponding to the sample data set; acquiring a target data set, determining the target motion mode of the target data set in the motion mode according to the target data set, and determining the target motion state label of the target data set in the candidate motion state labels.

[0007] In some embodiments, obtaining the sample partition pressure value of each corresponding plantar detection area according to the preset pressure calculation rule and the sample pressure data packet includes: determining the association between the pressure data and the plantar detection area according to the preset partitioning rule and the arrangement order of the pressure data in the sample pressure data packet; for any plantar detection area, determining the target pressure data associated with the plantar detection area from multiple pressure data according to the association, and obtaining the sample partition pressure value of the plantar detection area according to the preset pressure calculation rule and the target pressure data.

[0008] In some embodiments, determining the motion mode of a sample data set based on the sample feature values ​​corresponding to the sample data set for any given sample data set includes: for any given sample pressure data set, sequentially concatenating the sample feature values ​​of each foot detection region based on the arrangement order of the pressure data in the sample pressure data set to obtain sample motion features; for any given sample data set, sorting the sample motion features based on the sequential arrangement of the sample pressure data in the sample data set; concatenating the sorted sample motion features to obtain a motion sequence; and determining the motion mode of the sample data set based on the motion sequence.

[0009] In some embodiments, determining the motion pattern of the sample data set based on the motion sequence includes: determining a sample periodic sequence based on a preset periodic detection algorithm and the motion sequence; traversing each sample action feature in the sample periodic sequence and determining whether the current sample action feature is the same as the previous sample action feature; if the current sample action feature is the same as the previous sample action feature, deleting the current sample action feature from the sample periodic sequence to update the sample periodic sequence; obtaining a pattern sequence based on the updated sample periodic sequence; determining the sample action frequency based on the acquisition time of the sample pressure data packet corresponding to the head and tail of the pattern sequence; and determining the motion pattern of the sample data set based on the pattern sequence and the sample action frequency.

[0010] In some embodiments, the target data set includes multiple target pressure data sets; determining the target motion mode of the target data set in the motion mode according to the target data set includes: obtaining the target partition pressure value of each corresponding foot detection area according to the pressure calculation rule and the target pressure data set; determining the target feature value of each foot detection area according to the pressure threshold and the pressure value of each target partition; determining the target action feature of the target pressure data set according to the acquisition time sequence of the target pressure data set, taking the first determined target action feature as the head of the preset to be identified sequence, and judging whether the currently determined target action feature is consistent with the target action feature of the target pressure data set. The system checks whether the previously determined target action feature is the same; if the currently determined target action feature is different from the previously determined target action feature, it adds the currently determined target action feature to the end of the sequence to be identified to update the sequence to be identified; it matches the sequence to be identified with each of the pattern sequences, determines the initial screening motion pattern in each motion pattern, and determines the target sub-sequence that matches the initial screening motion pattern in the sequence to be identified; it determines the target action frequency based on the acquisition time of the target pressure data packet corresponding to the head and tail of the target sub-sequence; it matches the target action frequency with the sample action frequency corresponding to the initial screening motion pattern, and determines the target motion pattern in the initial screening motion pattern.

[0011] In some embodiments, the correlation refers to the correlation between the pressure data and the plantar detection area. The plantar detection area includes the left forefoot area, the left midfoot area, the left heel area, the right forefoot area, the right midfoot area, and the right heel area. Multiple evenly distributed pressure sensors are disposed within any of the plantar detection areas. The pressure sensors are used to acquire the corresponding pressure data. The distance between the pressure sensor and the center of the plantar detection area is negatively correlated with the corresponding correlation. For any given plantar detection area, based on the correlation, multiple pressure data... The process involves determining target pressure data associated with the plantar detection area, and obtaining sample partition pressure values ​​for the plantar detection area based on preset pressure calculation rules and the target pressure data. This includes: for any plantar detection area, determining target pressure data associated with the plantar detection area from multiple pressure data sets based on the correlation and a preset correlation threshold; obtaining target pressure values ​​for each target pressure data set based on preset pressure calculation rules and the target pressure data; and performing a weighted summation of the target pressure values ​​based on the correlation to obtain the sample partition pressure values ​​for the plantar detection area.

[0012] In some embodiments, the sample feature value includes a first feature value and a second feature value; determining the sample feature value of each plantar detection area based on a preset pressure threshold and the pressure value of each sample partition includes: when the pressure value of the sample partition is greater than or equal to the preset pressure threshold, determining the sample feature value of the corresponding plantar detection area as the first feature value; when the pressure value of the sample partition is less than the pressure threshold, determining the sample feature value of the corresponding plantar detection area as the second feature value.

[0013] To achieve the above objectives, a second aspect of this application proposes a human motion state recognition device, comprising: an acquisition unit, configured to acquire multiple sample data sets and corresponding candidate motion state labels, wherein the sample data sets include multiple sample pressure data packets, the multiple sample pressure data packets in the sample data sets are arranged in chronological order, the sample pressure data packets include pressure data of multiple foot detection areas, and the candidate motion state labels are used to characterize the human motion state category information of the corresponding sample data sets; a pressure calculation unit, configured to obtain sample partition pressure values ​​of each corresponding foot detection area according to a preset pressure calculation rule and the sample pressure data packets; a feature determination unit, configured to determine sample feature values ​​of each foot detection area according to a preset pressure threshold and the pressure values ​​of each sample partition; a pattern determination unit, configured to determine the motion pattern of any sample data set according to the sample feature values ​​corresponding to the sample data set; and a recognition unit, configured to acquire a target data set, determine the target motion pattern of the target data set in the motion patterns according to the target data set, and determine the target motion state label of the target data set in the candidate motion state labels.

[0014] To achieve the above objectives, a third aspect of this application provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the human motion state recognition method described in the first aspect.

[0015] To achieve the above objectives, a fourth aspect of the present application provides a storage medium, which is a computer-readable storage medium storing a computer program that, when executed by a processor, implements the human motion state recognition method described in the first aspect.

[0016] The human motion state recognition method, apparatus, device, and storage medium proposed in this application include, in embodiments of this application: acquiring multiple sample data sets and corresponding candidate motion state labels, wherein the sample data sets include multiple sample pressure data packets, the multiple sample pressure data packets in the sample data sets are arranged in chronological order, the sample pressure data packets include pressure data of multiple foot detection areas, and the candidate motion state labels are used to characterize the human motion state category information of the corresponding sample data sets; obtaining the sample partition pressure value of each corresponding foot detection area according to a preset pressure calculation rule and the sample pressure data packets; determining the sample feature value of each foot detection area according to a preset pressure threshold and the pressure value of each sample partition; determining the motion mode of the sample data set according to the sample feature value corresponding to any sample data set; acquiring a target data set, determining the target motion mode of the target data set in the motion mode according to the target data set, and determining the target motion state label of the target data set in the candidate motion state labels. According to the solution provided in the embodiments of this application, for any sample data set of human movement state category, pressure calculation is performed on the pressure data in the sample pressure data set to determine the sample partition pressure value of each foot detection area. Then, the pressure value is compared with the pressure threshold to determine the sample feature value of the foot detection area. The movement mode of the sample data set is then determined through the sample feature value. Subsequently, the target movement mode and target movement state label of the target data set can be determined based on the movement mode of each sample data set, thus realizing the effective identification of human movement state category. Compared with the time-consuming machine learning model training process, the process of determining the movement mode takes less time, improving the recognition efficiency. In addition, by using a wearable device equipped with a single type of pressure sensor, the pressure data of the foot detection area can be effectively acquired. The human movement state can be identified through the pressure data of the foot detection area without being affected by optical measurement standards, which can ensure the accuracy of recognition. Moreover, the structure of the wearable device is relatively simple, thereby improving the portability and practicality of the wearable device.

[0017] Other features and advantages of this application will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the application. The objectives and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the description, claims and drawings. Attached Figure Description

[0018] The accompanying drawings are used to provide a further understanding of the technical solutions of this application and constitute a part of the specification. They are used together with the embodiments of this application to explain the technical solutions of this application and do not constitute a limitation on the technical solutions of this application.

[0019] Figure 1 This is a flowchart of a human motion state recognition method provided in one embodiment of this application;

[0020] Figure 2 This is a flowchart of a method for obtaining sample partition pressure values ​​according to another embodiment of this application;

[0021] Figure 3 This is a flowchart of a method for determining a motion pattern provided in another embodiment of this application;

[0022] Figure 4 This is a flowchart illustrating a specific method for determining a motion pattern according to another embodiment of this application;

[0023] Figure 5 This is a flowchart of a method for determining a target motion pattern provided in another embodiment of this application;

[0024] Figure 6 This is a flowchart of another method for obtaining sample partition pressure values ​​provided in another embodiment of this application;

[0025] Figure 7 This is a flowchart of a method for determining sample feature values ​​provided in another embodiment of this application;

[0026] Figure 8 This is a schematic diagram of an optional pressure sensor and foot detection area distribution provided in another embodiment of this application;

[0027] Figure 9 This is a schematic diagram of the structure of a human motion state recognition device provided in another embodiment of this application;

[0028] Figure 10 This is a schematic diagram of the hardware structure of an electronic device provided in another embodiment of this application. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0030] In the description of this application, "several" means one or more, "multiple" means two or more, "greater than", "less than", "exceeding" etc. are understood to exclude the number itself, and "above", "below", "within" etc. are understood to include the number itself.

[0031] It should be noted that although functional modules are divided in the device schematic diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first," "second," etc., in the specification, claims, or the aforementioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0032] First, let's analyze some of the terms used in this application:

[0033] Artificial Intelligence (AI) is a new branch of computer science that studies, develops, and applies theories, methods, technologies, and systems to simulate, extend, and expand human intelligence. It aims to understand the essence of intelligence and produce intelligent machines that can react in a way similar to human intelligence. Research in this field includes robotics, speech recognition, image recognition, natural language processing, and expert systems. AI can simulate the information processes of human consciousness and thought. Furthermore, AI utilizes digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceiving the environment, acquiring knowledge, and using that knowledge to achieve optimal results.

[0034] Currently, trained machine learning models are typically used to identify human motion states. Images to be identified, collected by optical devices, or data to be identified from multiple sensors are input into the machine learning model to identify human motion states. However, machine learning models require a large amount of task-related datasets for training, which is time-consuming and has low recognition efficiency. Moreover, optical devices need to operate in a working environment that meets optical measurement standards. If the working environment does not meet optical measurement standards, the image acquisition effect of the optical devices will be poor, resulting in low accuracy in identifying human motion states. In addition, the more types and numbers of sensors in wearable devices, the more complex the structure of the wearable device, resulting in poor portability and practicality.

[0035] To address the problems of low recognition efficiency, low recognition accuracy, and poor portability, this application provides a method, apparatus, device, and storage medium for human motion state recognition. The method includes: acquiring multiple sample data sets and corresponding candidate motion state labels, wherein the sample data sets include multiple sample pressure data packets arranged in chronological order, and each sample pressure data packet includes pressure data from multiple foot detection areas; the candidate motion state labels are used to characterize the human motion state category information of the corresponding sample data set; obtaining the sample partition pressure value of each corresponding foot detection area according to a preset pressure calculation rule and the sample pressure data packets; determining the sample feature value of each foot detection area according to a preset pressure threshold and the pressure value of each sample partition; determining the motion mode of any sample data set based on the corresponding sample feature value; acquiring a target data set, determining the target motion mode of the target data set in the motion mode based on the target data set, and determining the target motion state label of the target data set from the candidate motion state labels. According to the solution provided in the embodiments of this application, for any sample data set of human movement state category, pressure calculation is performed on the pressure data in the sample pressure data set to determine the sample partition pressure value of each foot detection area. Then, the pressure value is compared with the pressure threshold to determine the sample feature value of the foot detection area. The movement mode of the sample data set is then determined through the sample feature value. Subsequently, the target movement mode and target movement state label of the target data set can be determined based on the movement mode of each sample data set, thus realizing the effective identification of human movement state category. Compared with the time-consuming machine learning model training process, the process of determining the movement mode takes less time, improving the recognition efficiency. In addition, by using a wearable device equipped with a single type of pressure sensor, the pressure data of the foot detection area can be effectively acquired. The human movement state can be identified through the pressure data of the foot detection area without being affected by optical measurement standards, which can ensure the accuracy of recognition. Moreover, the structure of the wearable device is relatively simple, thereby improving the portability and practicality of the wearable device.

[0036] The human motion state recognition method, device, equipment, and storage medium provided in this application are specifically described through the following embodiments. First, the human motion state recognition method in this application embodiment is described.

[0037] The human motion state recognition method provided in this application relates to the field of data processing technology. This method can be applied to a terminal, a server, or software running on either a terminal or a server. In some embodiments, the terminal can be a smartphone, tablet, laptop, desktop computer, etc.; the server can be configured as an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application implementing the human motion state recognition method, but is not limited to the above forms.

[0038] This application can be used in a wide variety of general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform specific tasks or implement specific abstract data types. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0039] It should be noted that in all specific embodiments of this application, when processing data related to user identity or characteristics, such as user information, user behavior data, user historical data, and user location information, user permission or consent is obtained first. Furthermore, the collection, use, and processing of this data comply with relevant laws, regulations, and standards. In addition, when embodiments of this application require access to sensitive personal information of users, separate permission or consent from the user is obtained through pop-ups or redirection to confirmation pages. Only after obtaining the user's separate permission or consent is the necessary user-related data required for the proper functioning of these embodiments acquired.

[0040] The embodiments of this application will be further described below with reference to the accompanying drawings.

[0041] like Figure 1 As shown, Figure 1This is a flowchart illustrating a human motion state recognition method according to an embodiment of this application. The human motion state recognition method can be executed by a server, a terminal, or a server in conjunction with a terminal. The human motion state recognition method includes, but is not limited to, the following steps S110 to S150:

[0042] Step S110: Obtain multiple sample data sets and corresponding candidate motion state labels. The sample data set includes multiple sample pressure data sets, which are arranged in chronological order. Each sample pressure data set includes pressure data from multiple foot detection areas. The candidate motion state labels are used to characterize the human motion state category information of the corresponding sample data set.

[0043] Step S120: Based on the preset pressure calculation rules and sample pressure data package, obtain the sample partition pressure values ​​of each foot detection area.

[0044] Step S130: Determine the sample feature values ​​of each foot detection area based on the preset pressure threshold and the pressure values ​​of each sample zone;

[0045] Step S140: For any sample data set, determine the motion mode of the sample data set based on the sample feature values ​​corresponding to the sample data set.

[0046] Step S150: Obtain the target data set; based on the target data set, determine the target motion mode of the target data set in the motion mode, and determine the target motion state label of the target data set in the candidate motion state labels.

[0047] Understandably, by setting up pressure sensors to acquire pressure data from the target user's soles, the pressure data exhibits different trends during different types of movement. The sample dataset contains pressure data within a sample time period. Within the sample dataset, sample pressure data packets correspond to sampling time points within the sample time period. Each sample pressure data packet includes pressure data from all sole detection areas at the corresponding sampling time point. By using pressure calculation rules to calculate regional pressure, the sample zone pressure values ​​for each sole detection area can be effectively calculated. Then, the sample feature values ​​for each sole detection area are determined, effectively reducing the amount of data and improving recognition efficiency. Furthermore, the movement pattern is determined by the sample feature values ​​of each sole detection area within the sample time period, enabling quick and accurate identification of the movement pattern within the sample dataset. The target dataset contains pressure data within a target time period. Through pattern matching, the target movement pattern corresponding to the target dataset can be accurately determined, thereby identifying the target movement state label. The target movement state label is then used for identification. The system identifies the human motion state category corresponding to the target data set. Based on this, for any sample data set of human motion state category, pressure calculations are performed on the pressure data within the sample pressure data set to determine the pressure values ​​of each foot detection area. Then, these values ​​are compared with pressure thresholds to determine the sample feature values ​​of the foot detection area. Subsequently, the motion pattern of the sample data set is determined based on the motion patterns of each sample data set. This enables effective identification of human motion state categories. Compared to the time-consuming machine learning model training process, the process of determining the motion pattern is shorter, improving recognition efficiency. Furthermore, using a wearable device equipped with a single type of pressure sensor can effectively acquire pressure data from the foot detection area. Identifying human motion states through the pressure data from the foot detection area is unaffected by optical measurement standards, ensuring accuracy. Moreover, the structure of the wearable device is relatively simple, thus improving its portability and practicality.

[0048] The foot detection area refers to the detection area in the wearable device, which corresponds to the user's foot. During the detection process, the user's foot is in contact with the detection area of ​​the wearable device, which can effectively obtain the user's foot pressure data.

[0049] It should be noted that after acquiring the target data set, the target partition pressure value of each foot detection area is calculated according to the pressure calculation rules and the target data set. Then, combined with the pressure threshold, the target feature value of each foot detection area is determined. Based on the target feature value, matching processing is performed on each movement mode to determine the target movement mode, thereby determining the target movement state label. In addition, the target data set can be updated according to the acquired real-time pressure data, and the target movement state label can be updated according to the processing result of the updated target data set.

[0050] It is worth noting that the sampling frequency of the sample pressure data packets in the sample dataset is fixed. For example, the sampling time interval between any two adjacent sample pressure data packets is 1 millisecond. In addition, the sampling frequency of the data packets in the target dataset is the same as that of the sample pressure data packets.

[0051] In practice, the target motion state label includes labels for normal motion states and labels for abnormal motion states. For example, normal motion states include, but are not limited to, walking, running, long jump, and rope skipping, while abnormal motion states include, but are not limited to, falling. After the terminal or server determines the target motion state label, it will send the data involved in the calculation process and the target motion state label to the cloud. Users can access the cloud through remote terminals to obtain the relevant data, effectively improving the convenience of information acquisition and enabling them to promptly learn about relevant abnormal information. In the remote terminal, a foot heat map of the motion process will be generated based on the corresponding target feature values, improving intuitiveness.

[0052] Additionally, refer to Figure 2 In one embodiment, Figure 1 Step S120 in the illustrated embodiment includes, but is not limited to, the following steps:

[0053] Step S210: Determine the correlation between pressure data and plantar detection area according to preset division rules and the arrangement order of each pressure data in sample pressure data package;

[0054] Step S220: For any plantar detection area, based on the correlation, determine the target pressure data associated with the plantar detection area from multiple pressure data, and obtain the sample partition pressure value of the plantar detection area according to the preset pressure calculation rules and the target pressure data.

[0055] It is understandable that for any two sample pressure data packets, the order of the pressure data at each detection location within the corresponding sample pressure data packet is fixed. For example, both the first and second sample pressure data packets contain pressure data at three detection locations. In the first sample pressure data packet, the pressure data at the first detection location is D11, the pressure data at the second detection location is D12, and the pressure data at the third detection location is D13, with the pressure data arranged in the order of D11, D12, D13. In the second sample pressure data packet, the pressure data at the first detection location is D21, the pressure data at the second detection location is D22, and the pressure data at the third detection location is D23, with the pressure data arranged in the order of D21, D22, D23. As can be seen, the arrangement order of pressure data at each detection location is fixed in the first and second sample pressure data packages. It should be noted that in order to increase the accuracy of human motion state recognition, pressure data from multiple detection locations needs to be acquired. The specific number of detection locations is set according to the actual situation and is not limited here. In addition, the foot detection area will cover multiple detection locations, and the detection locations covered by different foot detection areas will not overlap. Any foot detection area will be associated with multiple pressure data. Therefore, by setting the division rules, the correlation between pressure data and foot detection areas can be effectively determined. Then, by using the pressure calculation rules and the target pressure data associated with the foot detection areas, the sample partition pressure value of the foot detection area can be accurately calculated.Specifically, pressure data corresponding to each detection location in the plantar region is sequentially stitched together in a fixed direction to obtain a sample pressure data package. The division rules can be determined according to the detection location corresponding to the pressure data and the coverage area of ​​the plantar detection region. For example, the sample pressure data package contains pressure data corresponding to 30 detection locations, with a total of three detection regions. In the plantar region, the 30 detection locations occupy 6 rows, with each row containing 5 detection locations. The first detection region covers the detection locations in rows 1 and 2, the second detection region covers the detection locations in rows 3 and 4, and the third detection region covers the detection locations in rows 5 and 6. In the sample pressure data package, the pressure data listed first is the pressure data corresponding to the detection location located in the first row and on the leftmost side. According to the algorithm, for the first detection position, pressure data is sequentially pieced together from left to right. Then, the pressure data corresponding to the next detection position is sequentially pieced together until the last detection position. Therefore, the first detection area is associated with the pressure data arranged from the 1st to the 10th position, the second detection area with the pressure data arranged from the 11th to the 20th position, and the third detection area with the pressure data arranged from the 21st to the 30th position. The sample partition pressure value of the first detection area is calculated using the pressure data from the 1st to the 10th position; the sample partition pressure value of the second detection area is calculated using the pressure data from the 11th to the 20th position; and the sample partition pressure value of the third detection area is calculated using the pressure data from the 21st to the 30th position.

[0056] Additionally, refer to Figure 3 In one embodiment, Figure 1 Step S140 in the illustrated embodiment includes, but is not limited to, the following steps:

[0057] Step S310: For any sample pressure data packet, based on the arrangement order of each pressure data in the sample pressure data packet, the sample feature values ​​of each foot detection area are sequentially spliced ​​together to obtain the sample action features.

[0058] Step S320: For any sample dataset, sort the sample action features based on the sequential arrangement of the pressure data packets of each sample in the sample dataset.

[0059] Step S330: The sorted sample motion features are concatenated to obtain a motion sequence;

[0060] Step S340: Determine the motion pattern of the sample data set based on the motion sequence.

[0061] Understandably, in the sample pressure data packet, the order of the pressure data can determine the sorting of the plantar detection regions. That is, according to the sorting of the plantar detection regions, the sample feature values ​​of the plantar detection regions are sequentially concatenated to obtain the sample action features. For example, according to the order of the pressure data, the sorting of the six plantar detection regions can be determined as: left forefoot region, left midfoot region, left heel region, right forefoot region, right midfoot region, and right heel region. Assuming the sample feature value of the left forefoot region is 0, the sample feature value of the left midfoot region is 0, the sample feature value of the left heel region is 1, the sample feature value of the right forefoot region is 0, the sample feature value of the right midfoot region is 0, and the sample feature value of the right heel region is 0, the concatenation point... The sample action features are obtained as [001000]. In the sample data set, the sample pressure data packets are arranged in the order of acquisition time. For example, there are four sample pressure data packets in the sample data set. The sample action features of the first sample pressure data packet are [111111], the second sample pressure data packet are [001000], the third sample pressure data packet are [100100], and the fourth sample pressure data packet are [111111]. The motion sequences [111111], [001000], [100100], and [111111] are obtained by splicing. Then, the motion pattern is determined by the pattern features of the motion sequence.

[0062] Additionally, refer to Figure 4 In one embodiment, Figure 3 Step S340 in the illustrated embodiment includes, but is not limited to, the following steps:

[0063] Step S410: Determine the sample periodic sequence based on the preset periodic detection algorithm and motion sequence;

[0064] Step S420: Traverse the action features of each sample in the sample periodic sequence and determine whether the action features of the current sample are the same as the action features of the previous sample.

[0065] Step S430: If the current sample action feature is the same as the previous sample action feature, delete the current sample action feature in the sample periodic sequence to update the sample periodic sequence.

[0066] Step S440: Obtain the pattern sequence based on the updated sample periodic sequence;

[0067] Step S450: Determine the sample action frequency based on the acquisition time of the sample pressure data packets corresponding to the head and tail of the pattern sequence.

[0068] Step S460: Determine the motion pattern of the sample data set based on the pattern sequence and sample action frequency.

[0069] It is understandable that users will exhibit repetitive actions during continuous movement in the same type of motion state. Therefore, sample action features will appear periodically. By performing periodic analysis on the motion sequence using periodic detection algorithms, sample periodic sequences that characterize the motion pattern can be determined. For example, for the motion state of a long jump, the processed motion sequence is [111111], [001000], [100100], [111111], ..., and the sample periodic sequences can be determined as [111111], [001000], [100100]. Usually, the periodic sequence with the smallest period is used as the sample periodic sequence. In addition, since any... The occurrence of identical action features between two adjacent samples is determined by the frequency of the action. Higher frequency motion states have fewer consecutive and repetitive action features, while lower frequency motion states have more. Removing redundant consecutive and repetitive action features improves the efficiency of subsequent pattern matching and recognition. For example, for the walking motion state, the processed sample periodic sequence is [111111], [111111], [111001], [111001], [100111], [100111], [000111], [000111], [000111].

[11] The processed pattern sequences are [111111], [111001], [100111], and [000111]. For the running motion state, the processed sample periodic sequences are [111111], [111001], [100111], and [000111]. The processed pattern sequences are [111111], [111001], [100111], and [000111]. It can be seen that the walking pattern sequence and the running pattern sequence are the same. In order to distinguish different motion modes with the same pattern sequence, it is necessary to determine the sample by the acquisition time of the sample pressure data packet corresponding to the head and tail of the pattern sequence. Action frequency, for example, for the walking motion state, the sample pressure data packet acquisition time corresponding to the head of the pattern sequence is 12:00:00, and the sample pressure data packet acquisition time corresponding to the tail of the pattern sequence is 12:00:04. The calculated sample action frequency is 0.25Hz. For the running motion state, the sample pressure data packet acquisition time corresponding to the head of the pattern sequence is 13:00:00, and the sample pressure data packet acquisition time corresponding to the tail of the pattern sequence is 13:00:02. The calculated sample action frequency is 0.5Hz. Then, the motion pattern is determined by the pattern sequence and the sample action frequency, which can ensure the accuracy of the motion pattern.

[0070] Additionally, refer to Figure 5 In one embodiment, the target data set includes multiple target pressure data packets; Figure 1 Step S150 in the illustrated embodiment includes, but is not limited to, the following steps:

[0071] Step S510: Based on the pressure calculation rules and the target pressure data package, obtain the target zone pressure value for each corresponding plantar detection area;

[0072] Step S520: Determine the target feature value of each foot detection area based on the pressure threshold and the pressure value of each target zone;

[0073] Step S530: According to the order of acquisition time of the target pressure data packet, the target action features of the target pressure data packet are determined according to the corresponding target feature values. The first determined target action feature is used as the head of the preset sequence to be identified, and it is determined whether the currently determined target action feature is the same as the previous determined target action feature.

[0074] Step S540: If the currently determined target action feature is different from the previously determined target action feature, add the currently determined target action feature to the end of the sequence to be identified to update the sequence to be identified.

[0075] Step S550: Match the sequence to be identified with each pattern sequence, determine the initial screening motion pattern in each motion pattern, and determine the target sub-sequence that matches the initial screening motion pattern in the sequence to be identified.

[0076] Step S560: Determine the target action frequency based on the acquisition time of the target pressure data packets corresponding to the head and tail of the target subsequence;

[0077] Step S570: Match the target motion frequency with the sample motion frequency corresponding to the initial screening motion mode, and determine the target motion mode in the initial screening motion mode.

[0078] It is understood that the target pressure data packets within the target data set can be acquired in real time, or a certain number of target pressure data packets can be acquired at fixed time intervals; this is not limited here. Within the target data set, the target pressure data packets are arranged in the order of acquisition time. After acquiring the target data set, the target partition pressure value of each plantar detection area is calculated according to the pressure calculation rules and the target data set. Then, combined with the pressure threshold, the target feature value of each plantar detection area is determined. By splicing the target feature values ​​of each plantar detection area, the target action feature is obtained. For example, if there are 6 plantar detection areas, the target action feature obtained by splicing the target feature values ​​is [111111]. Then, according to the acquisition time sequence of the target pressure data packets, the target action features of the target pressure data packets are determined sequentially based on the corresponding target feature values. It is then determined whether the currently determined target action feature is the same as the previously determined target action feature. If they are different, the currently determined target action feature is added to the end of the sequence to be identified. For example, for the walking motion state, the first target action feature is determined first, and the first determined target action feature is [000111]. Therefore, the head of the sequence to be identified is [000111]. Through matching processing, it is determined that the sequence to be identified does not match any of the pattern sequences. Then, the second target action feature is determined, and the second determined target action feature is [111111], [11

[1111] is different from the previously determined [000111], so [111001] is added to the end of the sequence to be identified, and the sequence to be identified is updated to [000111] and [111111]. Through matching processing, it is determined that the sequence to be identified does not match any of the pattern sequences. Then, the third target action feature is determined, which is [111111]. [111111] is the same as the previously determined [111111], so the sequence to be identified is not updated. Then, the fourth target action feature is determined, which is [111001]. [111001] is different from the previously determined [111111], so [111001] is added to the end of the sequence to be identified. The sequence to be identified is added to the end of the sequence to be identified, and the sequence to be identified is updated to [000111], [111111], [111001]. Through matching processing, it is determined that the sequence to be identified does not match any of the pattern sequences. The above operation is repeated until the sequence to be identified is updated to [000111], [111111], [111001], [100111], [000111]. The sequence to be identified matches the pattern sequence of the walking state, and the sequence to be identified also matches the pattern sequence of the running state. Therefore, the movement patterns of the walking state and the running state are the initial screening movement patterns, and then the target subsequences [111111], [111001], [100111], [000111] are obtained.The target pressure data packet corresponding to the head [111111] of the target subsequence was acquired at 14:00:00, and the target pressure data packet corresponding to the tail [000111] of the target subsequence was acquired at 14:00:04. Therefore, the target motion frequency was determined to be 0.25Hz. This target motion frequency is the same as the motion frequency during running. Therefore, determining the target motion mode as the walking motion mode ensures the accuracy and efficiency of human motion state recognition.

[0079] It should be noted that during the data processing, the target action features of each target pressure data packet and each updated sequence to be identified will be stored in the database to ensure the integrity of data storage and facilitate subsequent data query and data reprocessing.

[0080] Additionally, refer to Figure 6 In one embodiment, the correlation refers to the correlation between pressure data and the foot detection area. The foot detection area includes the left forefoot area, the left midfoot area, the left heel area, the right forefoot area, the right midfoot area, and the right heel area. Multiple pressure sensors are evenly distributed in any one foot detection area. The pressure sensors are used to acquire the corresponding pressure data. The distance between the pressure sensor and the center of the foot detection area is negatively correlated with the corresponding correlation. Figure 2 Step S220 in the illustrated embodiment includes, but is not limited to, the following steps:

[0081] Step S610: For any plantar detection area, determine the target pressure data associated with the plantar detection area from multiple pressure data based on the correlation and a preset correlation threshold.

[0082] Step S620: Based on the preset pressure calculation rules and target pressure data, obtain the target pressure value for each target pressure data.

[0083] Step S630: Based on the correlation, the pressure values ​​of each target are weighted and summed to obtain the sample partition pressure values ​​of the plantar detection area.

[0084] Understandably, based on the preset division rules and the arrangement order of each pressure data in the sample pressure data package, the correlation between the pressure data and each foot detection area is determined. When the pressure data is not associated with the foot detection area, the correlation between the pressure data and the foot detection area is 0. After calculating the target pressure value of the target pressure data through the pressure calculation rules, for any foot detection area, each target pressure value is multiplied by the corresponding correlation and summed to obtain the sample partition pressure value of that foot detection area. In the foot detection area, the weight values ​​of different target pressure values ​​are different. The correlation is equivalent to the weight value of the target pressure value. By weighted summation, the accuracy of calculating the sample partition pressure value can be guaranteed, thereby improving the accuracy of human movement state recognition.

[0085] It is worth noting that the foot detection area can include 6 detection areas. The evenly distributed pressure sensors can ensure the accuracy of calculating the pressure values ​​of the sample areas, thereby improving the accuracy of human movement state recognition. For any foot detection area, the closer the detection position is to the center of the foot detection area, the greater the weight of the pressure data. Therefore, the distance between the pressure sensor and the center of the foot detection area is negatively correlated with the corresponding correlation.

[0086] It should be noted that, in addition to the six-area division method of left forefoot, left midfoot, left heel, right forefoot, right midfoot, and right heel, other area division methods can also be used during the testing process. The specific size and number of areas for foot testing are not limited here and can be adjusted according to the actual situation.

[0087] Additionally, refer to Figure 7 In one embodiment, the sample feature values ​​include a first feature value and a second feature value; Figure 1 Step S130 in the illustrated embodiment includes, but is not limited to, the following steps:

[0088] Step S710: If the pressure value of the sample partition is greater than or equal to the preset pressure threshold, determine the sample feature value of the corresponding plantar detection area as the first feature value.

[0089] Step S720: If the pressure value of the sample partition is less than the pressure threshold, determine the sample feature value of the corresponding foot detection area as the second feature value.

[0090] It is understandable that by setting a pressure threshold, the sample feature value of the foot detection area can be determined by comparison. For example, when the pressure value of the sample area is greater than or equal to the pressure threshold, the sample feature value of the foot detection area is 1, and when the pressure value of the sample area is less than the pressure threshold, the sample feature value of the foot detection area is 0.

[0091] It is worth noting that the pressure threshold can be determined through multiple experiments, which can ensure the accuracy of human motion state recognition.

[0092] In another embodiment, the number of pressure thresholds is greater than one; Figure 1 Step S130 in the illustrated embodiment includes, but is not limited to, the following steps:

[0093] Based on each pressure threshold, multiple pressure threshold ranges are determined, where any two pressure threshold ranges correspond to different pressure characteristic values.

[0094] The sample partition pressure value is matched with each pressure threshold range, the target pressure threshold range is determined among multiple pressure threshold ranges, and the sample feature value of the corresponding foot detection area is determined as the corresponding pressure feature value.

[0095] It is understandable that there can be one or more pressure thresholds. When there is only one pressure threshold, the sample feature value can be determined by judging the difference between the pressure value of the sample partition and the pressure threshold. When there are multiple pressure thresholds, it is necessary to first determine the pressure threshold range based on the pressure threshold, then determine the pressure threshold range that matches the pressure value of the sample partition, and finally use the pressure feature value corresponding to the target pressure threshold range as the sample feature value of the plantar detection area. For example, if there are two pressure thresholds, namely the first pressure threshold of 40 kPa and the second pressure threshold of 100 kPa, three pressure threshold ranges can be determined: the first pressure threshold range is less than 40 kPa, the second pressure threshold range is greater than 100 kPa, and so on. The pressure threshold is equal to 40 kPa and less than 100 kPa. The third pressure threshold range is greater than 100 kPa. The pressure feature value corresponding to the first pressure threshold range is 0, the pressure feature value corresponding to the second pressure threshold range is 0.5, and the pressure feature value corresponding to the third pressure threshold range is 1. Therefore, when the sample partition pressure value is within the first pressure threshold range, the corresponding sample feature value of the foot detection area is 0. When the sample partition pressure value is within the second pressure threshold range, the corresponding sample feature value of the foot detection area is 0.5. When the sample partition pressure value is within the third pressure threshold range, the corresponding sample feature value of the foot detection area is 1. This can improve the ability to recognize human movement status.

[0096] Additionally, refer to Figure 8 , Figure 8 This is a schematic diagram of an optional pressure sensor and foot detection area distribution provided in an embodiment of this application.

[0097] Understandably, in wearable devices, pressure sensors 810 are evenly distributed in the sole detection areas of the left and right feet, with the left and right sole detection areas being symmetrical. Figure 8As shown, the area can be divided into the left forefoot region 821, the left midfoot region 822, the left heel region 823, the right forefoot region 824, the right midfoot region 825, and the right heel region 826. By calculating the feature values ​​of each region, the recognition efficiency can be improved.

[0098] Additionally, refer to Figure 9 This application also provides a human motion state recognition device 800, comprising:

[0099] The acquisition unit 910 is used to acquire multiple sample data sets and corresponding candidate motion state labels. The sample data sets include multiple sample pressure data sets, which are arranged in chronological order. Each sample pressure data set includes pressure data from multiple foot detection areas. The candidate motion state labels are used to characterize the human motion state category information of the corresponding sample data set.

[0100] The pressure calculation unit 920 is used to obtain the sample partition pressure value of each foot detection area according to the preset pressure calculation rules and sample pressure data package.

[0101] The feature determination unit 930 is used to determine the sample feature values ​​of each foot detection area based on the preset pressure threshold and the pressure values ​​of each sample partition;

[0102] The pattern determination unit 940 is used to determine the motion pattern of any sample data set based on the sample feature values ​​corresponding to the sample data set.

[0103] The identification unit 950 is used to acquire a target data set, determine the target motion mode of the target data set in the motion mode according to the target data set, and determine the target motion state label of the target data set in the candidate motion state labels.

[0104] It is understood that the specific implementation of the human motion state recognition device 900 is basically the same as the specific implementation of the human motion state recognition method described above, and will not be repeated here.

[0105] Additionally, refer to Figure 10 , Figure 10 The hardware structure of an electronic device according to another embodiment is illustrated. The electronic device includes:

[0106] The processor 1001 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application.

[0107] The memory 1002 can be implemented as a read-only memory (ROM), static storage device, dynamic storage device, or random access memory (RAM). The memory 1002 can store the operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 1002 and called by the processor 1001 to execute the human motion state recognition method of the embodiments of this application, for example, executing the above-described... Figure 1 Method steps S110 to S150 Figure 2 Method steps S210 to S220, Figure 3 Method steps S310 to S340, Figure 4 Method steps S410 to S460 in the text Figure 5 Method steps S510 to S570 in the text Figure 6 Method steps S610 to S630 in the text Figure 7 Method steps S710 to S720;

[0108] Input / output interface 1003 is used to implement information input and output;

[0109] The communication interface 1004 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0110] Bus 1005 transmits information between various components of the device (e.g., processor 1001, memory 1002, input / output interface 1003, and communication interface 1004);

[0111] The processor 1001, memory 1002, input / output interface 1003 and communication interface 1004 are connected to each other within the device via bus 1005.

[0112] This application embodiment also provides a storage medium, which is a computer-readable storage medium for computer-readable storage. The storage medium stores one or more programs, which can be executed by one or more processors to implement the above-described human motion state recognition method, for example, executing the above-described... Figure 1 Method steps S110 to S150 Figure 2 Method steps S210 to S220, Figure 3 Method steps S310 to S340, Figure 4 Method steps S410 to S460 in the text Figure 5 Method steps S510 to S570 in the text Figure 6 Method steps S610 to S630 in the text Figure 7 Method steps S710 to S720.

[0113] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0114] The human motion state recognition method, apparatus, device, and storage medium provided in this application embodiment acquire multiple sample data sets and corresponding candidate motion state labels. Each sample data set includes multiple sample pressure data packets arranged chronologically. Each sample pressure data packet includes pressure data from multiple foot detection areas. Candidate motion state labels characterize the human motion state category information of the corresponding sample data set. Based on preset pressure calculation rules and the sample pressure data packets, sample partition pressure values ​​for each corresponding foot detection area are obtained. Based on preset pressure thresholds and the pressure values ​​of each sample partition, sample feature values ​​for each foot detection area are determined. For any sample data set, the motion mode of the sample data set is determined based on the corresponding sample feature values. A target data set is acquired, and based on the target data set, the target motion mode of the target data set is determined within the motion modes, and the target motion state label of the target data set is determined from the candidate motion state labels. Based on this, for any set of sample data for any human motion state category, pressure calculation is performed on the pressure data within the sample pressure data set to determine the pressure values ​​of sample partitions in each foot detection area. Then, by comparing these values ​​with pressure thresholds, the sample feature values ​​of the foot detection area are determined. Subsequently, the motion pattern of the sample data set is determined based on the motion patterns of each sample data set. This enables effective identification of human motion state categories. Compared to the time-consuming machine learning model training process, the process of determining the motion pattern is shorter, improving recognition efficiency. In addition, using a wearable device equipped with a single type of pressure sensor can effectively acquire pressure data from the foot detection area. Identifying human motion states through the pressure data from the foot detection area is not affected by optical measurement standards, ensuring recognition accuracy. Furthermore, the structure of the wearable device is relatively simple, thus improving its portability and practicality.

[0115] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0116] It will be understood by those skilled in the art that Figures 1 to 7 The technical solutions shown do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0117] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0118] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0119] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0120] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0121] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0122] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0123] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0124] If the integrated unit is implemented as 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 this application, in essence, 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. This computer software product is stored in a storage medium and includes multiple instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing programs, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0125] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.

Claims

1. A method of recognizing a human motion state, characterized by, The method comprises the following steps: obtaining a plurality of sample data sets and corresponding candidate motion state labels, wherein the sample data set comprises a plurality of sample pressure data packets, the plurality of sample pressure data packets in the sample data set are arranged in chronological order, the sample pressure data packet comprises pressure data of a plurality of plantar detection areas, and the candidate motion state label is used to represent the human motion state category information of the corresponding sample data set; determining the correlation between the pressure data and the plantar detection area according to a preset division rule and the arrangement order of each pressure data in the sample pressure data packet; for any plantar detection area, determining the target pressure data associated with the plantar detection area in the plurality of pressure data according to the correlation, and obtaining the sample partition pressure value of the plantar detection area according to a preset pressure calculation rule and the target pressure data; determining the sample feature value of each plantar detection area according to a preset pressure threshold and each sample partition pressure value; for any sample pressure data packet, sequentially splicing the sample feature values of each plantar detection area based on the arrangement order of each pressure data in the sample pressure data packet to obtain a sample action feature; for any sample data set, sorting the sample action feature based on the sequential arrangement of each sample pressure data packet in the sample data set; splicing the sorted sample action feature to obtain a motion sequence; determining the motion mode of the sample data set according to the motion sequence; obtaining a target data set, determining the target motion mode of the target data set in the motion mode according to the target data set, and determining the target motion state label of the target data set in the candidate motion state label; wherein the pressure threshold is one or more, and the determination of the sample feature value of each plantar detection area according to the preset pressure threshold and each sample partition pressure value comprises: determining a plurality of pressure threshold ranges according to each pressure threshold, wherein any two pressure threshold ranges correspond to different pressure feature values; matching the sample partition pressure value with each pressure threshold range to determine a target pressure threshold range in the plurality of pressure threshold ranges, and taking the pressure feature value corresponding to the target pressure threshold range as the sample feature value of the corresponding plantar detection area.

2. The method of claim 1, wherein, The determination of the motion mode of the sample data set according to the motion sequence comprises: determining a sample period sequence according to a preset periodic detection algorithm and the motion sequence; traversing each sample action feature in the sample period sequence to determine whether the current sample action feature and the previous sample action feature are the same; if the current sample action feature and the previous sample action feature are the same, deleting the current sample action feature in the sample period sequence to update the sample period sequence; obtaining a mode sequence according to the updated sample period sequence; According to the acquisition time of the sample pressure data packets corresponding to the head and tail of the mode sequence, a sample action frequency is determined; According to the mode sequence and the sample action frequency, a motion mode of the sample data set is determined.

3. The method of claim 2, wherein, The target data set includes a plurality of target pressure data packets; and determining, according to the target data set, a target motion mode of the target data set in the motion mode includes: According to the pressure calculation rule and the target pressure data packet, a target subarea pressure value corresponding to each of the plantar detection regions is obtained; According to the pressure threshold value and each of the target subarea pressure values, a target feature value of each of the plantar detection regions is determined; According to the acquisition time sequence of the target pressure data packets, each of the target feature values is used to determine a target action feature of the target pressure data packet, a first determined target action feature is taken as the head of a preset to-be-identified sequence, and it is determined whether the current determined target action feature is the same as the last determined target action feature; If the current determined target action feature is different from the last determined target action feature, the current determined target action feature is added to the tail of the to-be-identified sequence to update the to-be-identified sequence; The to-be-identified sequence and each of the mode sequences are matched, an initial screening motion mode is determined in each of the motion modes, and a target subsequence matching the initial screening motion mode is determined in the to-be-identified sequence; According to the acquisition time of the target pressure data packets corresponding to the head and tail of the target subsequence, a target action frequency is determined; The target action frequency and the sample action frequency corresponding to the initial screening motion mode are matched, and a target motion mode is determined in the initial screening motion mode.

4. The method of claim 1, wherein, The correlation is the correlation between the pressure data and the plantar detection region, the plantar detection region includes a left forefoot region, a left midfoot region, a left heel region, a right forefoot region, a right midfoot region, and a right heel region, a plurality of uniformly distributed pressure sensors are arranged in any of the plantar detection regions, the pressure sensors are used to acquire the corresponding pressure data, and the distance between the pressure sensors and the center of the plantar detection region is negatively correlated with the corresponding correlation; for any of the plantar detection regions, according to the correlation, target pressure data associated with the plantar detection region is determined from a plurality of pressure data, a sample subarea pressure value of the plantar detection region is obtained according to a preset pressure calculation rule and the target pressure data, including: For any of the plantar detection regions, according to the correlation and a preset correlation threshold value, target pressure data associated with the plantar detection region is determined from a plurality of pressure data; According to a preset pressure calculation rule and the target pressure data, a target pressure value of each of the target pressure data is obtained; Based on the correlation, each of the target pressure values is weighted and summed to obtain a sample subarea pressure value of the plantar detection region.

5. The method of claim 1, wherein, The sample feature value includes a first feature value and a second feature value; The sample feature value of each of the plantar detection areas is determined according to the preset pressure threshold and the sample subarea pressure value of each of the sample subareas, including: In the case that the sample subarea pressure value is greater than or equal to the preset pressure threshold, the sample feature value of the corresponding plantar detection area is determined as the first feature value; In the case that the sample subarea pressure value is less than the pressure threshold, the sample feature value of the corresponding plantar detection area is determined as the second feature value.

6. A human motion state recognition apparatus characterized by comprising: Including: An acquisition unit is configured to acquire a plurality of sample data sets and corresponding candidate motion state labels, wherein the sample data set includes a plurality of sample pressure data packets, the plurality of sample pressure data packets in the sample data set are arranged in chronological order, the sample pressure data packet includes pressure data of a plurality of plantar detection areas, and the candidate motion state label is used to represent human motion state category information corresponding to the sample data set; A pressure calculation unit is configured to determine an association between the pressure data and the plantar detection area according to a preset division rule and an arrangement order of each of the pressure data in the sample pressure data packet; for any plantar detection area, target pressure data associated with the plantar detection area is determined from the plurality of pressure data according to the association; and a sample subarea pressure value of the plantar detection area is obtained according to a preset pressure calculation rule and the target pressure data; A feature determination unit is configured to determine a sample feature value of each of the plantar detection areas according to a preset pressure threshold and each of the sample subarea pressure values; A mode determination unit is configured to sequentially splice the sample feature values of each of the plantar detection areas based on the arrangement order of each of the pressure data in the sample pressure data packet for any sample pressure data packet, to obtain a sample action feature; to sort the sample action feature based on the sequential arrangement of each of the sample pressure data packets in the sample data set for any sample data set; to splice the sorted sample action feature, to obtain a motion sequence; and to determine a motion mode of the sample data set according to the motion sequence; An identification unit is configured to acquire a target data set, to determine a target motion mode of the target data set in the motion mode according to the target data set, and to determine a target motion state label of the target data set in the candidate motion state label; The pressure threshold is one or more, and the sample feature value of each of the plantar detection areas is determined according to the preset pressure threshold and each of the sample subarea pressure values, including: A plurality of pressure threshold ranges are determined according to each of the pressure thresholds, wherein the pressure feature values corresponding to any two of the pressure threshold ranges are different; The sample partition pressure value is matched with each of the pressure threshold ranges, a target pressure threshold range is determined from the pressure threshold ranges, and the pressure feature value corresponding to the target pressure threshold range is taken as a sample feature value of the corresponding foot bottom detection area.

7. An electronic device, comprising: The electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the human motion state recognition method according to any one of claims 1 to 5 when executing the computer program.

8. A storage medium storing a computer program, characterized by The computer program is executed by the processor to implement the human motion state recognition method according to any one of claims 1 to 5.

Images