An ultra-wideband-based static human posture risk identification system and method

Abstract

This invention relates to the field of wireless sensing technology and discloses a static human posture risk recognition system and method based on ultra-wideband (UWB). The system includes: multiple UWB base stations deployed at preset base station locations in the space to be monitored, used to obtain the distance from each UWB tag to each UWB base station; multiple UWB tags set at preset key locations on the monitored human body, used to mark the position and posture of the monitored human body; and a data processing module, which is communicatively connected to the UWB base stations, used to identify the position and posture of the monitored human body based on the distance from each UWB tag to each UWB base station, and to determine the risk level based on the identification results. This invention relies solely on UWB base stations and tags to complete distance measurement, accurately capturing the static spatial geometric features of the human body, effectively solving the deficiency of traditional solutions in being unable to identify posture when the human body is stationary, reliably distinguishing the key states of the monitored human body, and quickly mapping the risk level and triggering early warnings.

Description

Technical Field

[0001] This invention relates to the field of wireless sensing technology, specifically to a static human posture risk recognition system and method based on ultra-wideband. Background Technology

[0002] With the increasing aging of the population, the safety and health monitoring of elderly people living alone has become an important social issue. Existing monitoring solutions using cameras pose serious privacy risks and are difficult to deploy in private spaces such as bedrooms and bathrooms. Although wearable solutions based on inertial measurement units (IMUs) can detect motion, accelerometers or gyroscopes rely on detecting the acceleration and angular velocity of motion and cannot effectively identify static postures. When the human body is completely still, such as after falling unconscious, sitting still, or lying still, the IMU data features disappear, making it impossible to distinguish key static postures such as "sitting in a chair" and "falling on the floor". Adding additional sensors would increase costs, power consumption, and wearing burden.

[0003] Existing monitoring solutions cannot accurately identify specific static human postures and automatically map them to risk levels, nor can they achieve continuous and intelligent assessment and early warning of users' safety status. Summary of the Invention

[0004] This invention provides a static human posture risk recognition system and method based on ultra-wideband to solve the problem of not being able to accurately identify specific static human postures and automatically map them to risk levels.

[0005] In a first aspect, the present invention provides a static human posture risk recognition system based on ultra-wideband (UWB). The system includes: multiple UWB base stations, multiple UWB tags, and a data processing module, wherein... Multiple ultra-wideband base stations are deployed at preset base station locations in the space to be monitored to obtain the distance from each ultra-wideband tag to each ultra-wideband base station; Multiple ultra-wideband tags are set at preset key locations on the monitored human body to mark the position and posture of the monitored human body; The data processing module, which communicates with the ultra-wideband base station, is used to identify the position and posture of the monitored human body based on the distance from each ultra-wideband tag to each ultra-wideband base station, and to determine the risk level based on the identification results.

[0006] The static human posture risk recognition system based on ultra-wideband provided by this invention relies solely on ultra-wideband base stations and tags to complete ranging, eliminating the need for cameras, inertial measurement units, and other equipment. This fundamentally avoids privacy leaks while reducing hardware costs, power consumption, and wearing burden. It accurately captures the static spatial geometric features of the human body, effectively solving the shortcomings of traditional solutions that cannot recognize posture when the human body is stationary. It can reliably distinguish the key states of the monitored human body and quickly map risk levels and trigger warnings, making it particularly suitable for scenarios with sensitive privacy and high reliability requirements, such as elderly care monitoring.

[0007] In one optional implementation, the number of ultra-wideband base stations is at least four, and the preset base station locations are multiple non-collinear points on the ceiling; The number of ultra-wideband tags is three or four, and the preset key parts include: head, chest, waist and legs.

[0008] The static human posture risk recognition system based on ultra-wideband provided by this invention uses at least four non-collinear ultra-wideband base stations deployed on the ceiling, and three or four ultra-wideband tags placed on key parts of the human body. This system can not only ensure stable and reliable positioning and posture geometric representation, but also simplify hardware deployment, reduce costs and wearing burden, accurately extract static posture features, and improve recognition robustness and accuracy.

[0009] In one alternative implementation, the system further includes clothing with ultra-wideband tags set on preset key areas of the clothing, including: head, chest, and waist, or head, chest, waist, and legs.

[0010] The present invention provides an ultra-wideband static human posture risk recognition system that integrates ultra-wideband tags into key parts of clothing. It is convenient to wear and does not feel foreign, which greatly reduces the burden and resistance of the elderly and other monitored subjects. The tag layout covers the head, chest, waist or legs, completely representing the static geometric structure of the human body and ensuring the accuracy of posture recognition. The integrated wearable design is more stable and reliable, suitable for daily long-term monitoring, and takes into account practicality, comfort and recognition effect.

[0011] Secondly, the present invention provides a static human posture risk recognition method based on ultra-wideband, the method being applied to any of the ultra-wideband static human posture risk recognition systems of the first aspect, the method comprising: Obtain distance data from each UWB tag to each UWB base station and construct a multi-dimensional feature vector; Based on multidimensional feature vectors, a pre-trained model is used to identify the posture of the monitored human body; Based on the posture of the monitored human body, the risk level of the monitored human body is determined by querying the preset posture-risk mapping rule base.

[0012] The present invention provides a static human posture risk recognition method based on ultra-wideband (UWB). It constructs a multi-dimensional feature vector based on pure UWB ranging data, combines it with a pre-trained model to achieve accurate recognition of static human posture, and then quickly maps the risk level through a preset rule base. The entire process does not require inertial sensors or vision devices, effectively solving the problems of static recognition failure and high privacy risks in traditional solutions. The process is simple, computationally efficient, accurate, robust, and can stably complete non-visual, all-weather human posture risk monitoring, making it suitable for high-privacy monitoring scenarios such as elderly care and health.

[0013] In one optional implementation, the training process of the pre-trained model includes: Obtain multiple multidimensional feature vectors and their corresponding human poses to construct training samples; The number of nodes in the fully connected deep neural network model is determined based on the number of ultra-wideband labels and ultra-wideband base stations, and the fully connected deep neural network model is trained based on the training samples to obtain the pre-trained model.

[0014] The static human posture risk recognition method based on ultra-wideband provided by this invention adaptively configures fully connected deep neural network nodes according to the number of ultra-wideband labels and base stations, resulting in high model structure matching and more thorough feature learning. It constructs sample training with pure ultra-wideband ranging feature vectors and corresponding postures, which makes the model training efficient, with better generalization and robustness, and can accurately learn spatial geometric features, ensuring stable and reliable static posture recognition.

[0015] In one optional implementation, based on the posture of the monitored human body, a preset posture-risk mapping rule base is queried to determine the risk level of the monitored human body, including: If the monitored person is lying on the floor or curled up on the floor, the risk level is high. If the monitored person's posture is bent over under the bed or the posture is uncertain, the risk level is medium risk. If the monitored person's posture is lying flat on the bed, lying on their side on the bed, sitting on the bed, standing or sitting off the bed, then the risk level is low.

[0016] The present invention provides a static human posture risk identification method based on ultra-wideband. By pre-setting clear posture-risk mapping rules, different static human postures are directly mapped to three levels of risk: high, medium, and low. The judgment logic is clear and the response is rapid. It accurately distinguishes between dangerous, warning, and normal states, realizes automatic risk level determination, avoids false alarms and missed alarms, improves the reliability and practicality of the monitoring system, and provides a stable basis for timely early warning and emergency response.

[0017] In one alternative implementation, the method further includes: If the risk level is high, an alarm message will be sent to the monitoring personnel and / or emergency center via wireless communication. If the risk level is medium, a notification message will be sent to the monitoring personnel via wireless communication.

[0018] The present invention provides a static human posture risk recognition method based on ultra-wideband, which triggers alarms and prompts according to risk level. High risk directly notifies the caregiver and emergency center, while medium risk only sends a reminder to the caregiver. The graded handling is more reasonable and efficient. It relies on wireless communication to realize remote real-time push, and the response is timely, without false alarms or omissions, which greatly improves the safety and emergency response efficiency of elderly care scenarios.

[0019] In some alternative implementations, the method further includes: Record the duration of the monitored human body in the same posture. If the duration exceeds a preset time threshold, send an alarm message to the monitoring personnel and / or emergency center via wireless communication.

[0020] The static human posture risk recognition method based on ultra-wideband provided by this invention breaks through the limitations of single posture risk judgment. By using the low-risk posture duration exceeding the limit mechanism, it accurately captures hidden dangers such as lying flat on the bed for too long, which may be suspected of being in a coma or incapacitated. It fills the blind spots of conventional recognition, takes into account both false alarm control and danger capture, and expands the dimensions of monitoring.

[0021] Thirdly, the present invention provides an electronic device, comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to perform the method described in the second aspect or any corresponding embodiment thereof.

[0022] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to perform the method described in the second aspect or any of its corresponding embodiments. Attached Figure Description

[0023] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0024] Figure 1 This is a structural block diagram of a static human posture risk recognition system based on ultra-wideband according to an embodiment of the present invention; Figure 2 This is a schematic diagram of a specific layout scenario for a static human posture risk recognition system based on ultra-wideband according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the first process of a static human posture risk recognition method based on ultra-wideband according to an embodiment of the present invention; Figure 4 This is a schematic diagram illustrating the construction principle of multidimensional feature vectors in the static human posture risk recognition method based on ultra-wideband according to an embodiment of the present invention. Figure 5 This is a schematic diagram of the second process of the static human posture risk recognition method based on ultra-wideband according to an embodiment of the present invention; Figure 6 This is a schematic diagram of the multilayer sensing mechanism in the static human posture risk recognition method based on ultra-wideband according to an embodiment of the present invention. Figure 7 This is a schematic diagram of risk identification in the static human posture risk identification method based on ultra-wideband according to an embodiment of the present invention; Figure 8 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] It is understood that before using the technical solutions disclosed in the various embodiments of the present invention, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in the present invention and their authorization should be obtained in accordance with relevant laws and regulations through appropriate means.

[0027] In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0028] This invention provides a static human posture risk recognition system and method based on ultra-wideband (UWB). By using UWB base stations and tag ranging, combined with a neural network model, the system identifies human postures to accurately identify static human postures and automatically map them to risk levels. This is particularly suitable for scenarios such as elderly care and health monitoring where privacy requirements are strict.

[0029] This embodiment provides a static human posture risk recognition system based on ultra-wideband, such as... Figure 1 As shown, the system includes: multiple ultra-wideband base stations, multiple ultra-wideband tags, and a data processing module.

[0030] like Figure 1 As shown, multiple ultra-wideband base stations are deployed at preset base station locations in the space to be monitored, and are used to obtain the distance from each ultra-wideband tag to each ultra-wideband base station.

[0031] Specifically, multiple Ultra Wide Band (UWB) base stations are deployed within the space to be monitored, forming a positioning coordinate system. Each UWB base station is deployed at a pre-defined base station location, such as the four corners of a rectangular space, or the four corners or four sides of a ceiling. Figure 2 The example shown uses the four corners of the ceiling (A1, A2, A3, A4) as an example only, but it is not a limitation. The selection of the preset base station location should be made to ensure that the ultra-wideband base station is not obstructed, so as to provide good data transmission effect.

[0032] like Figure 1 As shown, multiple ultra-wideband tags are set at preset key locations on the monitored human body to mark the position and posture of the monitored human body.

[0033] Specifically, multiple ultra-wideband tags are set at preset key locations on the monitored human body, and the position and posture of the monitored human body are marked by tags at different locations. Figure 2 Taking head (T1), chest (T2), waist (T3), and knee (T4) as examples, but not limited to these, the more UWB tags there are, the more accurate the pose recognition results. UWB base stations continuously measure and record the distance information from each UWB tag to each UWB base station by identifying the tag positions. Figure 2 In this context, d11, d12, d13, and d14 represent the distances from header label T1 to the four ultra-wideband base stations A1, A2, A3, and A4, respectively.

[0034] like Figure 1 As shown, the data processing module is connected to the ultra-wideband base station and is used to identify the position and posture of the monitored human body based on the distance from each ultra-wideband tag to each ultra-wideband base station, and to determine the risk level based on the identification results.

[0035] Specifically, the data processing module establishes a stable communication connection with the ultra-wideband base station, and receives the raw ranging data from each ultra-wideband tag to the corresponding base station in real time. Using pure ultra-wideband ranging information as the sole data source, without relying on inertial measurement units, cameras, or other auxiliary sensors, the module preprocesses the ranging data and constructs multi-dimensional feature vectors through built-in algorithms. Then, it calls a pre-trained deep fully connected neural network model to accurately analyze the static spatial geometric features of the human body and complete the location positioning and static posture classification and recognition of the monitored human body.

[0036] After recognition is completed, the system automatically queries the preset posture-risk mapping rule library based on the recognition results, and quickly maps the posture results to three levels of risk: high, medium and low. This provides an accurate and reliable basis for subsequent graded alarms and safety monitoring, ensuring the real-time performance, accuracy and robustness of static human posture risk recognition.

[0037] The ultra-wideband static human posture risk recognition system provided in this embodiment relies solely on ultra-wideband base stations and tags to complete ranging, eliminating the need for cameras, inertial measurement units, and other equipment. This fundamentally avoids privacy leaks while reducing hardware costs, power consumption, and wearing burden. It accurately captures the static spatial geometric features of the human body, effectively solving the shortcomings of traditional solutions that cannot recognize posture when the human body is stationary. It can reliably distinguish the key states of the monitored human body and quickly map risk levels and trigger warnings, making it particularly suitable for scenarios with sensitive privacy and high reliability requirements, such as elderly care monitoring.

[0038] In some alternative implementations, the number of ultra-wideband (UWB) base stations is at least four, with the preset base station locations being multiple non-collinear points on the ceiling. The number of UWB tags is three or four, with preset key body parts including: head, chest, waist, and legs.

[0039] Specifically, the number of ultra-wideband base stations is at least four, and the preset base station locations are multiple non-collinear points on the ceiling. This four-point non-collinear deployment constructs a stable three-dimensional positioning coordinate system, avoiding ranging blind spots and geometric degradation problems caused by collinear layouts. This meets the hardware requirements for complete acquisition of static human spatial geometric features. Figure 2 As shown, four ultra-broadband base stations can be deployed at the four corners of the ceiling in the bedroom.

[0040] The number of ultra-wideband tags is set to three or four, with key areas pre-defined to cover the head, chest, waist, and legs. The tag positions are aligned with the core posture representation nodes of the torso, which can fully reflect the spatial distribution differences of static postures such as standing, sitting, lying down, falling, and curling up. Three to four tags can ensure recognition accuracy while taking into account hardware cost, ease of wear, and system power consumption, making them suitable for long-term wearable scenarios such as elderly care and monitoring.

[0041] The ultra-wideband-based static human posture risk recognition system provided in this embodiment uses at least four non-collinear ultra-wideband base stations deployed on the ceiling, along with three or four ultra-wideband tags placed on key parts of the human body. This not only ensures stable and reliable positioning and posture geometric representation, but also simplifies hardware deployment, reduces costs and wearing burden, accurately extracts static posture features, and improves recognition robustness and accuracy.

[0042] In some alternative implementations, the system also includes clothing with ultra-wideband tags set on preset key areas of the clothing, including: head, chest, waist, or head, chest, waist, and legs.

[0043] Specifically, ultra-wideband tags are fixedly placed in key pre-defined areas of clothing, making them easy for monitored individuals to wear and use daily without the need for additional adhesive or handheld attachment, thus improving ease of use and wearing comfort. For example, an elderly person being monitored might wear a lightweight vest with four built-in UWB tags, located on the shoulder (replacing the head), chest, lower back, and upper thigh on one side. To further reduce costs and wearing burden, three tags (e.g., head, chest, and waist) can also be used to achieve basic posture recognition.

[0044] The ultra-wideband static human posture risk recognition system provided in this embodiment integrates ultra-wideband tags into key parts of clothing, making it convenient to wear without any foreign body sensation. This significantly reduces the burden and resistance of monitoring subjects such as the elderly. The tag layout covers the head, chest, waist, or legs, fully representing the static geometric structure of the human body and ensuring posture recognition accuracy. The integrated wearable design is more stable and reliable, suitable for daily long-term monitoring, and takes into account practicality, comfort, and recognition effect.

[0045] According to an embodiment of the present invention, an embodiment of a static human posture risk recognition method based on ultra-wideband is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0046] This embodiment provides a static human posture risk recognition method based on ultra-wideband, which can be used in the aforementioned computing system or the ultra-wideband static human posture risk recognition system. Figure 3 This is a flowchart of a static human posture risk recognition method based on ultra-wideband according to an embodiment of the present invention, such as... Figure 3 As shown, the process includes the following steps: Step S201: Obtain the distance data from each UWB tag to each UWB base station and construct a multi-dimensional feature vector.

[0047] Specifically, distance data from each UWB tag to each UWB base station is acquired synchronously in real time. For each sampling time t, the distance data from all tags to all base stations are combined sequentially to form a multidimensional feature vector D_t. This multidimensional feature vector fully represents the geometric structure information of the monitored human body in the monitored space at time t. For example, if... Figure 2In the scenario shown, there are four base stations and four tags. The multidimensional feature vector is: D_t = [d_T1-A1, d_T1-A2, ..., d_T1-A4, d_T2-A1, ..., d_T4-A4].

[0048] like Figure 4 The diagram illustrates the construction principle of a multi-dimensional feature vector. The collected distance data is transmitted to the feature vector construction engine. Within the engine, each distance data point is converted into a standardized feature vector. Then, all distance data are arranged in a fixed order to ensure complete consistency in input dimension and structure between the training and inference phases. After processing by the engine, a multi-dimensional feature vector is finally generated. The dimension of the multi-dimensional feature vector is equal to the number of UWB base stations multiplied by the number of UWB tags, with each element corresponding to a single base station-tag ranging result.

[0049] Step S202: Based on multi-dimensional feature vectors, the posture of the monitored human body is identified using a pre-trained model.

[0050] Specifically, based on deep fully connected neural networks, the static geometric structure of the human body in space can be analyzed, and the posture of the human body in a completely static state can be accurately identified.

[0051] A multidimensional feature vector D_t is input into a pre-trained deep fully connected neural network model. The output layer nodes of the pre-trained model correspond to specific static pose categories, and the probability distribution of each pose is output through a normalized exponential (Softmax) function. The model transforms the input raw distance data layer by layer into high-level semantic features through forward propagation. This process does not rely on IMU sensor data; it achieves static recognition solely by mining the spatial structure information contained in the UWB distance data. Finally, the pose category with the highest probability is taken as the final identified static pose of the monitored human body.

[0052] UWB signals have strong anti-interference capabilities and stable ranging; neural networks can learn complex and nonlinear spatial geometric relationships, have better adaptability to different body shapes and environmental layouts, and have higher recognition accuracy.

[0053] Step S203: Based on the posture of the monitored human body, query the preset posture-risk mapping rule base to determine the risk level of the monitored human body.

[0054] Specifically, a mapping rule base from attitude semantics to risk level is established. Based on the final identified attitude result, the preset attitude-risk rule base is queried to determine the corresponding risk level.

[0055] A new end-to-end recognition paradigm that directly translates UWB distance data into attitude skips the traditional approaches of "first locating coordinates, then analyzing the trajectory" or "fusing IMU data," resulting in a simple and efficient technical path.

[0056] The static human posture risk recognition method based on ultra-wideband provided in this embodiment relies on pure ultra-wideband ranging data to construct multi-dimensional feature vectors, and combines them with a pre-trained model to achieve accurate recognition of static human posture. Then, it quickly maps risk levels through a preset rule base. The entire process does not require inertial sensors or vision devices, effectively solving the problems of static recognition failure and high privacy risks in traditional solutions. The process is simple, computationally efficient, accurate, robust, and can stably complete non-visual, all-weather human posture risk monitoring, making it suitable for high-privacy monitoring scenarios such as elderly care and health.

[0057] This embodiment provides a static human posture risk recognition method based on ultra-wideband, which can be used in the aforementioned computer system. Figure 5 This is a flowchart of a static human posture risk recognition method based on ultra-wideband according to an embodiment of the present invention, such as... Figure 5 As shown, the process includes the following steps: Step S301: Obtain the distance data from each UWB tag to each UWB base station, and construct a multi-dimensional feature vector. For details, please refer to [link to relevant documentation]. Figure 3 Step S201 of the illustrated embodiment will not be described again here.

[0058] Step S302: Based on multi-dimensional feature vectors, the posture of the monitored human body is identified using a pre-trained model.

[0059] In some alternative implementations, the training process of the pre-trained model includes: Step a1: Obtain multiple multidimensional feature vectors and their corresponding human poses to construct training samples.

[0060] Step a2: Determine the number of nodes in the fully connected deep neural network model based on the number of ultra-wideband labels and ultra-wideband base stations, and train the fully connected deep neural network model based on the training samples to obtain the pre-trained model.

[0061] Specifically, a large amount of multidimensional feature vector data corresponding to different human postures is collected. Human postures cover a variety of static postures, including but not limited to: lying on the floor under the bed, curled up in bed, bending over under the bed, uncertain posture, lying flat on the bed, lying on the side in bed, sitting in bed, standing under the bed, and sitting under the bed. Then, a three-layer fully connected neural network is trained using the collected pure UWB distance feature vectors and the corresponding posture labels.

[0062] Fully connected deep neural networks employ a multilayer perceptron architecture. The specific structure is an optimized solution determined after extensive comparative experiments, achieving the best balance between accuracy, robustness, and computational efficiency. For example... Figure 6 As shown, the structure of the multilayer sensing mechanism includes: (1) Input layer: The number of nodes strictly corresponds to the dimension of the feature vector D_t. For example, when 4 tags and 4 base stations are deployed, the number of nodes in the input layer is 16.

[0063] (2) Hidden Layer: Contains multiple consecutive fully connected layers, with the number of nodes decreasing layer by layer (128→64→32) to achieve hierarchical abstraction and efficient compression of UWB spatial geometric features. Experiments show that this configuration can most effectively learn the complex mapping from distance data to pose. Specifically: The first hidden layer has 128 nodes, uses the ReLU activation function, and employs a dropout rate of 0.3 to force the network to learn more robust features and prevent overfitting.

[0064] The second hidden layer has 64 nodes and uses the ReLU activation function. It also employs a dropout rate of 0.3 to further enhance the model's generalization ability.

[0065] The third hidden layer has 32 nodes and uses the ReLU activation function. This layer does not use Dropout, in order to provide a stable and deterministic high-level feature representation for the output layer, ensuring the reliability of the final classification decision.

[0066] (3) Output layer: The number of nodes corresponds to the number of core static pose categories to be identified, and the normalized probability distribution is output using the Softmax activation function.

[0067] Step S303: Based on the posture of the monitored human body, query the preset posture-risk mapping rule base to determine the risk level of the monitored human body.

[0068] Specifically, step S303 includes: Step S3031: If the monitored human body is lying on the ground or curled up under the bed, the risk level is high risk.

[0069] Step S3032: If the posture of the monitored human body is bent over under the bed or the posture is uncertain, the risk level is medium risk.

[0070] Step S3033: If the posture of the monitored human body is lying flat on the bed, lying on the side on the bed, sitting on the bed, standing or sitting off the bed, then the risk level is low risk.

[0071] Specifically, the preset posture-risk mapping rule base is shown in Table 1. The risk levels are divided into three levels: high risk (requiring immediate intervention), medium risk (requiring attention or alerts), and low risk (normal state).

[0072] Table 1 Preset Attitude-Risk Mapping Rule Base

[0073] Based on the posture of the monitored human body, a preset posture-risk mapping rule base is queried to determine the risk level of the monitored human body. For example... Figure 7 As shown, when the posture recognition result is "lying flat on the bed", the risk assessment result is low risk and safe; when the posture recognition result is "falling off the bed", the risk assessment result is high risk and dangerous. This is just an example and is not limited to this.

[0074] The ultra-wideband static human posture risk identification method provided in this embodiment directly corresponds different static human postures to three levels of risk: high, medium, and low, through a pre-defined clear posture-risk mapping rule. The judgment logic is clear and the response is rapid, accurately distinguishing between dangerous, warning, and normal states, realizing automatic risk level determination, avoiding false alarms and missed alarms, improving the reliability and practicality of the monitoring system, and providing a stable basis for timely early warning and emergency response.

[0075] The static human posture risk recognition method based on ultra-wideband provided in this embodiment adaptively configures fully connected deep neural network nodes according to the number of ultra-wideband labels and base stations, resulting in high model structure matching and more thorough feature learning. It constructs sample training with pure ultra-wideband ranging feature vectors and corresponding postures, which makes the model training efficient, with better generalization and robustness. It can accurately learn spatial geometric features and ensure stable and reliable static posture recognition.

[0076] In some alternative implementations, the method further includes: If the risk level is high, an alarm message will be sent to the monitoring personnel and / or emergency center via wireless communication.

[0077] Specifically, when the system detects that the monitored person is in a high-risk posture (such as falling off the bed or curling up in bed, which may endanger life), it will immediately trigger the highest level of alarm procedure. Simultaneously, alarm information will be sent to the monitoring personnel's terminal (mobile phone / monitoring platform) and / or the emergency center system, ensuring that multiple parties receive emergency notifications at the same time. Wireless communication methods (such as cellular networks, NB-IoT, Wi-Fi, etc.) will be used to ensure remote, real-time push notifications regardless of geographical location. The alarm content may include the monitored person's location information, posture recognition results, timestamps, and pre-set annotations (such as "suspected fall"), providing crucial decision-making information for monitoring personnel and emergency centers.

[0078] When the model identifies a high-risk event with high confidence, an alarm message is immediately sent to the children's mobile phones via Wi-Fi. The message could be something like, "The elderly person may have fallen off the bed. Please check on them immediately." This is just an example and is not a limitation.

[0079] If the risk level is medium, a notification message will be sent to the monitoring personnel via wireless communication.

[0080] Specifically, when the system detects that the monitored person is in a medium-risk posture (such as bending over under the bed, uncertain posture, etc., indicating potential risks or abnormalities, but not yet posing an immediate threat to life), a notification process will be triggered. The notification message will only be sent to the monitoring personnel's terminal, not directly to the emergency center, to avoid unnecessary strain on public emergency resources. Wireless communication will also be used, allowing the system to reach the monitoring personnel via SMS, app push notifications, WeChat messages, etc. The notification content may include the posture recognition result (e.g., "suspected bending over"), location, and timestamp, reminding the monitoring personnel to pay attention to the monitored person's condition and promptly conduct manual verification or intervention.

[0081] The ultra-wideband static human posture risk recognition method provided in this embodiment triggers alarms and prompts according to risk level. High risk directly notifies the caregiver and emergency center, while medium risk only sends a reminder to the caregiver. The graded handling is more reasonable and efficient. It relies on wireless communication to realize remote real-time push, which responds in a timely manner and does not false alarm or miss alarm, greatly improving the safety and emergency response efficiency of elderly care scenarios.

[0082] In some alternative implementations, the method further includes: Record the duration of the monitored human body in the same posture. If the duration exceeds a preset time threshold, send an alarm message to the monitoring personnel and / or emergency center via wireless communication.

[0083] Specifically, after completing posture recognition and risk level determination, the system will continuously track the duration of posture maintenance. If the risk level based on posture recognition is low risk, but the duration of the same posture (e.g., lying flat on a bed) exceeds a preset time threshold (e.g., 10 hours), the monitored person may be unconscious. At this time, an alarm message will also be sent to the monitoring personnel and / or emergency center via wireless communication.

[0084] Building upon the basic posture-risk mapping logic, a mechanism for determining the duration of low-risk postures exceeding limits is introduced. This addresses potential risks that cannot be covered by single posture recognition (such as coma or disability caused by prolonged bed rest). The specific implementation process is as follows: The basic posture recognition and risk assessment system first performs posture recognition using pure UWB data and then determines the risk level based on a preset rule base. For example, when the "lying flat on the bed" posture is recognized, it is judged as a low-risk state according to conventional rules. The duration of maintaining a low-risk posture is continuously timed, recording the duration of the same posture from the first recognition to the current moment. For example, the duration for which the monitored elderly person maintains the "lying flat on the bed" posture is accumulated and statistically analyzed in real time. The duration is compared with a preset low-risk posture time threshold (e.g., 10 hours, which can be flexibly configured according to the monitoring scenario and the user's health status). If the duration does not exceed the threshold, it indicates a normal sleep or rest state, and no additional actions are triggered; if the duration exceeds the threshold, even if the posture itself belongs to the low-risk category, the system will judge that there is an abnormal risk (e.g., the elderly person has not turned over for a long time, is suspected of being unconscious, or is unable to move due to a sudden illness).

[0085] When a low-risk posture lasts for an extended period, the system will break from the usual logic of not triggering alarms for low-risk situations and will proactively send alarm information to the monitoring personnel and / or emergency center via wireless communication to indicate that there may be an abnormality in the monitored person and to promptly remind them to check and intervene.

[0086] The ultra-wideband static human posture risk recognition method provided in this embodiment breaks through the limitations of single posture risk judgment. By using the low-risk posture duration exceeding the limit mechanism, it accurately captures hidden dangers such as prolonged lying flat on the bed, which may indicate coma or disability. This fills the blind spots of conventional recognition, takes into account both false alarm control and danger capture, and expands the dimensions of monitoring.

[0087] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0088] The following is a detailed reference. Figure 8 This diagram illustrates a suitable structural schematic for implementing an electronic device according to embodiments of the present invention. The electronic device may include a processor (e.g., a central processing unit, graphics processor, etc.) 801, which can perform various appropriate actions and processes based on a program stored in read-only memory (ROM) 802 or a program loaded from memory 808 into random access memory (RAM) 803. The RAM 803 also stores various programs and data required for the operation of the electronic device. The processor 801, ROM 802, and RAM 803 are interconnected via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0089] Typically, the following devices can be connected to I / O interface 805: input devices 806 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 807 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 808 including, for example, magnetic tapes, hard disks, etc.; and communication devices 809. Communication device 809 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 8 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.

[0090] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 809, or installed from a memory 808, or installed from a ROM 802. When the computer program is executed by the processor 801, it performs the functions defined in the ultra-wideband-based static human posture risk recognition method of the embodiments of the present invention.

[0091] Figure 8 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of the present invention.

[0092] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the ultra-wideband-based static human posture risk recognition method shown in the above embodiments is implemented.

[0093] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A static human posture risk recognition system based on ultra-wideband, characterized in that, The system includes: multiple ultra-wideband base stations, multiple ultra-wideband tags, and a data processing module, wherein... Multiple ultra-wideband base stations are deployed at preset base station locations in the space to be monitored to obtain the distance from each ultra-wideband tag to each ultra-wideband base station; Multiple ultra-wideband tags are set at preset key locations on the monitored human body to mark the position and posture of the monitored human body; The data processing module, which is communicatively connected to the ultra-wideband base station, is used to identify the position and posture of the monitored human body based on the distance from each ultra-wideband tag to each ultra-wideband base station, and to determine the risk level based on the identification results.

2. The system according to claim 1, characterized in that, The number of ultra-wideband base stations is at least four, and the preset base station locations are multiple non-collinear points on the ceiling; The number of ultra-wideband tags is three or four, and the preset key parts include: head, chest, waist and legs.

3. The system according to claim 2, characterized in that, The system also includes clothing with ultra-wideband tags set on preset key parts of the clothing. The preset key parts include: head, chest, and waist, or head, chest, waist, and legs.

4. A static human posture risk recognition method based on ultra-wideband, characterized in that, The method is applied to the ultra-wideband-based static human posture risk recognition system according to any one of claims 1-3, and the method includes: Obtain distance data from each UWB tag to each UWB base station and construct a multi-dimensional feature vector; Based on the multidimensional feature vector, the posture of the monitored human body is identified using a pre-trained model; Based on the posture of the monitored human body, the risk level of the monitored human body is determined by querying the preset posture-risk mapping rule base.

5. The method according to claim 4, characterized in that, The training process of the pre-trained model includes: Obtain multiple multidimensional feature vectors and their corresponding human poses to construct training samples; The number of nodes in the fully connected deep neural network model is determined based on the number of ultra-wideband labels and ultra-wideband base stations, and the fully connected deep neural network model is trained based on the training samples to obtain a pre-trained model.

6. The method according to claim 4, characterized in that, Based on the posture of the monitored human body, a preset posture-risk mapping rule base is queried to determine the risk level of the monitored human body, including: If the monitored person is lying on the floor or curled up on the floor, the risk level is high. If the monitored person's posture is bent over under the bed or the posture is uncertain, the risk level is medium risk. If the monitored person's posture is lying flat on the bed, lying on their side on the bed, sitting on the bed, standing or sitting off the bed, then the risk level is low.

7. The method according to claim 6, characterized in that, The method further includes: If the risk level is high, an alarm message will be sent to the monitoring personnel and / or emergency center via wireless communication. If the risk level is medium, a notification message will be sent to the monitoring personnel via wireless communication.

8. The method according to claim 4, characterized in that, The method further includes: The system records the duration of the monitored human body in the same posture. If the duration exceeds a preset time threshold, an alarm message is sent to the monitoring personnel and / or emergency center via wireless communication.

9. An electronic device, characterized in that, include: A memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, the processor executing the computer instructions to perform the method of any one of claims 4 to 8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing a computer to perform the method of any one of claims 4 to 8.

Images