Indoor personnel position and typical attitude monitoring method based on electrostatic detection technology

Abstract

The invention discloses an indoor personnel position and typical attitude monitoring method based on an electrostatic detection technology, and the method comprises the steps: collecting an electrostatic simulation signal generated by the movement of indoor personnel through an electrostatic sensor array, and carrying out the signal conditioning and analog-to-digital conversion of the electrostatic simulation signal, and obtaining an electrostatic digital signal; constructing an electrostatic signal data set based on the electrostatic digital signal, training a time sequence classification model based on the electrostatic signal data set, and establishing a mapping relationship between the electrostatic signal and the personnel position and attitude type; and using the trained time sequence classification model to discriminate the current position and the typical posture type of the indoor personnel according to the static test signal acquired in real time. Under the non-contact condition, the electrostatic sensor is used for detecting charges carried by a human body, and the position and the typical posture type of an indoor person are judged according to signals of the electrode array.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of indoor personnel position and posture monitoring, and particularly relates to an indoor personnel position and typical posture monitoring method based on electrostatic detection technology. BACKGROUND

[0002] With the deep penetration of Internet of Things technology, the intelligent demand of human-computer interaction continues to upgrade, and human posture recognition as the core technology connecting the physical world and digital system has increasingly highlighted its application value in health monitoring, public safety, smart home and other fields. For example, in the medical field, remote monitoring equipment integrated with posture recognition can significantly improve the response efficiency of medical staff to abnormal postures of bedridden patients (such as pre-falling signs and limb stiffness), and reliable posture monitoring can further reduce the incidence of safety accidents such as falling and getting lost in special scenarios such as nursing homes and kindergartens.

[0003] The current mainstream human posture monitoring system has significant technical limitations: first, the optical monitoring system based on visual cameras, whose recognition accuracy is highly dependent on lighting conditions, has a high error rate in strong light, backlight or occlusion scenarios, and there is a significant privacy leakage risk in real-time imaging, such as in private scenarios such as bedrooms and nursing rooms, which has low user acceptance; second, wearable monitoring devices, which directly collect joint motion data through mechanical sensors, but need to be charged frequently and are easily affected by clothing friction and motion amplitude when worn, which has significant application limitations; third, millimeter wave monitoring, which has high energy consumption as an active measurement method, requires frequent power replacement or continuous power supply for long-term deployment, has certain limitations on monitoring distance, and signals are easily disturbed by metal obstacles, such as household appliances and metal furniture, which can reflect signals and cause measurement deviation; in addition, the recognition resolution of small movements is limited, making it difficult to accurately capture detailed motion data, and the measurement stability and detail recognition ability in complex scenarios need to be improved.

[0004] Currently, electrostatic sensing technology mainly has the following problems in the actual application of personnel position and posture recognition: first, the electrostatic signal quickly decays as the measurement target moves away, and the signal-to-noise ratio and measurement accuracy are greatly affected by the distance, making it difficult to achieve accurate quantitative analysis; second, the indoor environment adaptability problem, environmental medium parameters fluctuation (such as air humidity, furniture material) will change the electric field distribution characteristics, and make the human posture recognition error expand. SUMMARY

[0005] To solve the above technical problems in the prior art, the present application provides an indoor personnel position and typical posture monitoring method based on electrostatic detection technology to meet the human posture recognition needs in indoor environments.

[0006] In one aspect, to achieve the above object, the present application provides an indoor personnel position and typical posture monitoring method based on electrostatic detection technology, comprising:

[0007] An electrostatic analog signal generated by indoor personnel activity is collected by an electrostatic sensor array, and signal conditioning and analog-to-digital conversion are performed on the electrostatic analog signal to obtain an electrostatic digital signal;

[0008] An electrostatic signal dataset is constructed based on the electrostatic digital signal, and a time series classification model is trained based on the electrostatic signal dataset to establish a mapping relationship between the electrostatic signal and the personnel position and posture type;

[0009] The trained time series classification model is used to determine the current position and typical posture type of the indoor personnel according to the real-time collected electrostatic test signal.

[0010] Preferably, the signal conditioning includes signal amplification and filtering processing.

[0011] Preferably, the electrostatic signal dataset is constructed by:

[0012] Under different environmental parameter conditions, the electrostatic signals of the indoor personnel in different positions and different typical postures are repeatedly collected, and multi-dimensional features of the electrostatic signals are extracted, and all test data and corresponding features are integrated to construct the electrostatic signal dataset.

[0013] Preferably, the electrostatic sensor array includes at least three electrostatic electrodes arranged in a preset geometric shape on the top of the indoor space for detecting the electric field change caused by the body charge.

[0014] Preferably, the electrostatic electrodes are arranged in an equilateral triangle, and the coordinates of each electrostatic electrode in a three-dimensional coordinate system are (x i , y i , H), where i is the electrode identifier of the electrostatic electrode, and H is the vertical height of the electrostatic sensor from the indoor floor.

[0015] The coordinates of the human target, i.e., the indoor personnel, in the three-dimensional coordinate system are (x, y, h), where h represents the height parameter of the human posture.

[0016] The straight-line distance between the human target and each electrostatic sensor is d i .

[0017] Wherein, the coordinates of the electrostatic electrodes in the three-dimensional coordinate system, the coordinates of the human target in the three-dimensional coordinate system, and the straight-line distance between the human target and each electrostatic sensor satisfy the condition: (x-x i )² + (y-y i )² + (H-h)²= d i ².

[0018] Preferably, the mapping function relationship between the output signal intensity and the straight-line distance between the human target and each electrostatic sensor is:

[0019] S i = f (d i );

[0020] In the formula, S i is the output signal intensity of the electric sensor, and f (・) is a mapping function determined by calibration.

[0021] Preferably, the time series classification model is a time convolution network model, which is used to process time series data of electrostatic signals and output a classification result.

[0022] In another aspect, to achieve the above-mentioned purpose, the application also provides an indoor personnel position and typical posture monitoring system based on electrostatic detection technology, comprising:

[0023] An electrostatic sensor array is arranged on the top of the indoor space and used to collect electrostatic analog signals generated by the activities of indoor personnel;

[0024] A signal conditioning circuit is connected to the electrostatic sensor array and used to amplify and filter the electrostatic analog signals;

[0025] A microprocessor unit is connected to the signal conditioning circuit and used to convert the conditioned analog signals into digital signals;

[0026] A storage unit is used to store electrostatic signal data sets and a trained time series classification model;

[0027] A processing unit is used to call the time series classification model and determine the position and typical posture type of the indoor personnel according to real-time electrostatic digital signals.

[0028] Compared with the prior art, the application has the following advantages and technical effects:

[0029] (1) The application uses the electrostatic induction principle to sense the electric field changes caused by the electric charges naturally carried by the human body, without the need for camera to collect optical images or personnel to wear any equipment, which fundamentally eliminates the visual privacy leakage problem and eliminates the bondage and inconvenience brought by wearable devices, and realizes non-contact, high-privacy and safe precise monitoring.

[0030] (2) The application not only uses the powerful feature learning ability of the model, but also introduces physical prior knowledge as a constraint, effectively overcoming the defects of insufficient generalization ability of pure data-driven methods under signal attenuation and environmental fluctuations, thereby realizing more accurate and reliable joint calculation of personnel position (including two-dimensional plane coordinates and height) and typical posture. BRIEF DESCRIPTION OF DRAWINGS

[0031] The accompanying drawings, which form a part of the present application, are intended to provide further understanding of the present application and are incorporated herein in conjunction with the description. The accompanying drawings of the present application, together with the description, explain the principles of the present application and the application but are not meant to limit the present application.

[0032] Fig. 1 A position and posture discrimination schematic diagram of an embodiment of the present application;

[0033] Fig. 2 An indoor activity area, electrostatic sensor coordinate position and structure schematic diagram of an embodiment of the present application;

[0034] Fig. 3 An indoor personnel position and typical posture monitoring method flowchart based on electrostatic detection technology of an embodiment of the present application;

[0035] Among them, 1, indoor area division, 2, electrostatic electrode array basic structure. DETAILED DESCRIPTION

[0036] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict. The present application will be described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.

[0037] It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a group of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from here.

[0038] As Figs. 1-3 , the present embodiment proposes an indoor personnel position and typical posture monitoring method based on electrostatic detection technology, comprising:

[0039] An electrostatic analog signal generated by indoor personnel activity is collected by an electrostatic sensor array, and signal conditioning and analog-to-digital conversion are performed on the electrostatic analog signal to obtain an electrostatic digital signal;

[0040] An electrostatic signal dataset is constructed based on the electrostatic digital signal, and a time series classification model is trained based on the electrostatic signal dataset to establish a mapping relationship between the electrostatic signal and the personnel position and posture type;

[0041] The trained time series classification model is used to determine the current position and typical posture type of the indoor personnel according to the real-time collected electrostatic test signal.

[0042] Further, the signal conditioning includes signal amplification and filtering processing.

[0043] Specifically, the electrostatic signals generated by the activities of the indoor person on the electrode array are collected by the electrostatic sensor, and after being processed by the amplification and filtering module, the analog signals are converted into digital signals by the built-in ADC of the microprocessor.

[0044] Further, the electrostatic signal dataset is constructed, including:

[0045] Under different environmental parameter conditions, the electrostatic signals of the indoor person located at different positions and in different typical postures are repeatedly collected, and the multi-dimensional features of the electrostatic signals are extracted, and all the test data and corresponding features are integrated to construct the electrostatic signal dataset.

[0046] Specifically, the electrostatic sensor signals of the indoor person located at different spatial positions, in different typical postures and under different environmental parameter conditions within the measurement range of the sensor are collected by three electrostatic sensors; for each collection situation, multiple tests are repeatedly performed, the multi-dimensional features of the electrostatic signals in each test are extracted, and all the test data and corresponding features are integrated to construct the electrostatic signal dataset covering multiple scenes.

[0047] Further, the classification model constructed by the time convolution network is trained by the dataset to establish the mapping between the electrostatic signals and the positions and typical posture types of the indoor person.

[0048] Specifically, based on the physical characteristics of the electrostatic field, the output signal strength of the electrostatic sensor is only related to the straight-line distance of the human target relative to the electrostatic sensor: three electrostatic electrodes arranged in an equilateral triangle in the indoor ceiling lamp are denoted as electrodes A, B and C, and the side length of the equilateral triangle is l; a three-dimensional coordinate system is preset in the room, and the coordinates of the three electrostatic electrodes in the three-dimensional coordinate system are (x i , y i , H), where i is A, B and C, and H represents the vertical height of the electrostatic sensor from the indoor floor; the coordinates of the human target in the three-dimensional coordinate system are (x, y, h), where h represents the height parameter of the human posture; the output signal strength of each electrostatic sensor is denoted as S i , and the straight-line distance between the human target and each electrostatic sensor is denoted as d i , which satisfies: (x-x i )² + (y-y i )² + (H-h)² = d i ² for any sensor.

[0049] Further, by simultaneously solving the above three equations (about electrodes A, B and C) and combining the mapping function relationship, the spatial position parameters (x, y) and the height parameter h of the human target can be solved by the classification model constructed by the time convolution network.

[0050] Furthermore, the mapping function relationship between the output signal strength and the straight-line distance between the human target and each electrostatic sensor is as follows:

[0051] S i = f (d i );

[0052] In the formula, S i Let f(・) be the output signal strength of the electrical sensor, and f(・) be the calibrated mapping function.

[0053] Specifically, the classification model based on a temporal convolutional network is trained using the constructed electrostatic signal sample dataset. During training, d... i and mapping function S i The calculated spatial position parameters (x, y) and height parameter h of the human target enable the classification model to learn the correlation between electrostatic signal features and typical posture types, ultimately resulting in a trained temporal convolutional network classification model that maps the output signal of the electrostatic sensor to the position and typical posture types of people indoors.

[0054] Furthermore, the attitude information is obtained through parsing in the following manner:

[0055] Based on real-time electrostatic signals of indoor occupants collected by electrostatic sensors, and combined with a pre-constructed electrostatic signal sample dataset containing different typical postures of indoor occupants, multi-dimensional features are first extracted from the real-time signals. Then, the extracted features are matched with the baseline signal features corresponding to each typical posture in the sample dataset. Simultaneously, the human height parameter h obtained through sensor signals is input into a temporal convolutional network classification model trained on the sample dataset for prediction. Finally, the current posture type and posture information of the indoor occupants are obtained, realizing the model mapping from electrostatic signal data to posture information.

[0056] Furthermore, electrostatic test signals caused by typical postures of personnel at any indoor location are collected using electrostatic sensors. A trained time-series classification model is then invoked to determine the personnel's position and posture type. The time-series classification model is a temporal convolutional network model, which processes the time-series data of the electrostatic signals and outputs the classification results.

[0057] This embodiment also provides an indoor personnel position and typical posture monitoring system based on electrostatic detection technology, including:

[0058] An electrostatic sensor array, installed on the indoor ceiling, is used to collect simulated electrostatic signals generated by indoor human activities;

[0059] A signal conditioning circuit, connected to the electrostatic sensor array, is used to amplify and filter the electrostatic analog signal;

[0060] The microprocessor unit, connected to the signal conditioning circuit, is used to convert the conditioned analog signal into a digital signal;

[0061] Storage unit, used to store electrostatic signal dataset and trained time series classification model;

[0062] The processing unit is used to call the time series classification model to determine the location and typical posture type of indoor personnel based on real-time electrostatic digital signals.

[0063] Specifically, the indoor personnel location and typical posture monitoring system based on electrostatic detection technology is implemented in a specific embodiment as follows:

[0064] (1) Obtain electrostatic digital signals through an electrostatic sensor array, signal conditioning circuit and microprocessor unit.

[0065] (2) Construct an electrostatic signal dataset containing electrostatic signals of people in different positions and train a temporal convolutional network classification model.

[0066] (3) Measure the electrostatic signal of the posture of the person in the room at any position in the room, and call the temporal convolutional network classification model to determine the position and posture type of the person in the room.

[0067] The system acquires electrostatic digital signals from the electrode array via an electrostatic sensor array, a signal conditioning circuit, and a microprocessor unit. First, a power supply and power conversion circuit are connected to power the sensor system. Then, the signal conditioning circuit acquires the electrostatic analog signals from the electrode array, and the microprocessor unit receives the analog signals and converts them into digital signals using its built-in ADC.

[0068] Construct an electrostatic signal dataset containing electrostatic signals of people in typical postures when in different indoor positions, and train a temporal convolutional network classification model: For any room, fix an electrostatic sensor in the center of the roof, and control the indoor temperature and humidity through air conditioning, humidifier and dehumidifier.

[0069] Electrostatic sensor signals of personnel in different indoor positions and typical postures were collected. The data included various personnel positions and typical postures. Electrostatic signals were collected repeatedly for each situation to form a training dataset and train a temporal convolutional network classification model.

[0070] When measuring the electrostatic signals of indoor personnel's posture, a classification model is called to determine the personnel's position and posture type: When personnel and sensors change position and posture, electrostatic sensor signals different from the training dataset are collected through electrostatic sensors, and the trained classification model is called to predict the indoor personnel's position and posture type.

[0071] In this embodiment, the electrostatic sensor is a circular array of three electrodes, but other structures can also be used; the time series classification model uses a temporal convolutional network, but other time series models can also be used; the personnel position and posture discrimination uses a classification model, but regression and mechanistic models can also be used.

[0072] To address the problem of recognizing typical postures of people in different indoor locations, this embodiment collects electrostatic signals indoors using an electrode array, constructs an electrostatic signal dataset, trains a temporal convolutional network classification model, and determines the position and typical posture type of the person.

[0073] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for monitoring the position and typical posture of indoor personnel based on electrostatic detection technology, characterized in that, include: The electrostatic analog signals generated by indoor human activities are collected by an electrostatic sensor array, and the electrostatic analog signals are conditioned and converted from analog to digital to obtain electrostatic digital signals. An electrostatic signal dataset is constructed based on the electrostatic digital signal, and a time series classification model is trained based on the electrostatic signal dataset to establish a mapping relationship between electrostatic signals and personnel position and posture type. Using a trained time-series classification model, the current location and typical posture type of indoor personnel are determined based on real-time electrostatic test signals.

2. The indoor personnel position and typical posture monitoring method based on electrostatic detection technology according to claim 1, characterized in that, The signal conditioning includes signal amplification and filtering.

3. The indoor personnel position and typical posture monitoring method based on electrostatic detection technology according to claim 1, characterized in that, Constructing the electrostatic signal dataset includes: Under different environmental parameter conditions, electrostatic signals were repeatedly collected from indoor personnel in different positions and typical postures. Multi-dimensional features of the electrostatic signals were extracted, and all test data and corresponding features were integrated to construct the electrostatic signal dataset.

4. The indoor personnel position and typical posture monitoring method based on electrostatic detection technology according to claim 1, characterized in that, The electrostatic sensor array includes at least three electrostatic electrodes arranged in a preset geometry on the ceiling of the room to detect changes in the electric field caused by human body charge.

5. The indoor personnel position and typical posture monitoring method based on electrostatic detection technology according to claim 4, characterized in that, The electrostatic electrodes are arranged in an equilateral triangle, and the coordinates of each electrostatic electrode in the three-dimensional coordinate system are (x, y, y) and (x, y, y). i y i ,H), where i is the electrode identifier of the electrostatic electrode, and H is the vertical height of the electrostatic sensor from the indoor ground; The coordinates of an indoor person as a human target in the three-dimensional coordinate system are (x, y, h), where h represents the height parameter of the human posture; The straight-line distance between the human target and each electrostatic sensor is d. i ; The conditions that the coordinates of the electrostatic electrodes in the three-dimensional coordinate system, the coordinates of the human target in the three-dimensional coordinate system, and the straight-line distances between the human target and each electrostatic sensor satisfy are: (x - x i )² + (y-y i )² + (H-h)² =d i ².

6. The indoor personnel position and typical posture monitoring method based on electrostatic detection technology according to claim 5, characterized in that, The mapping function relationship between the output signal strength and the straight-line distance between the human target and each electrostatic sensor is as follows: S i = f (d i ); In the formula, S i Let f(・) be the output signal strength of the electrical sensor, and f(・) be the calibrated mapping function.

7. The indoor personnel position and typical posture monitoring method based on electrostatic detection technology according to claim 1, characterized in that, The time series classification model is a temporal convolutional network model, which is used to process time series data of electrostatic signals and output classification results.

8. An indoor personnel position and typical posture monitoring system based on electrostatic detection technology for implementing the indoor personnel position and typical posture monitoring method based on electrostatic detection technology according to any one of claims 1-7, characterized in that, include: An electrostatic sensor array, installed on the indoor ceiling, is used to collect simulated electrostatic signals generated by indoor human activities; A signal conditioning circuit, connected to the electrostatic sensor array, is used to amplify and filter the electrostatic analog signal; The microprocessor unit, connected to the signal conditioning circuit, is used to convert the conditioned analog signal into a digital signal; Storage unit, used to store electrostatic signal dataset and trained time series classification model; The processing unit is used to call the time series classification model to determine the location and typical posture type of indoor personnel based on real-time electrostatic digital signals.

Images