Fall detection method and system and fall alarm method

Abstract

The invention discloses a tumble detection method, a tumble detection system and a tumble alarm method. The tumble detection method comprises the following steps: acquiring X-axis, Y-axis and Z-axis acceleration and X-axis, Y-axis and Z-axis angular velocity data of an inertial sensor; calculating an attitude angle change rate according to the XYZ three-axis acceleration and the XYZ three-axis angular velocity; calculating a three-axis acceleration variance and a resultant acceleration peak value according to the XYZ three-axis acceleration; according to the calculated three-axis acceleration variance, the attitude angle change rate and the resultant acceleration peak value, whether falling occurs is judged. According to the invention, through the multi-mode sensor, the fall detection precision in a complex scene can be significantly improved.

Description

Technical Field

[0001] This invention relates to the field of fall detection technology, specifically to a fall detection method, system, and fall alarm method. Background Technology

[0002] In daily life, helmets play a crucial role in ensuring personal safety during cycling. With the increasing popularity of outdoor sports, fall detection technology has become a research hotspot in the field of wearable devices. Currently, mainstream solutions include:

[0003] Detection based on a single sensor: such as relying solely on an accelerometer and determining a fall by setting an acceleration threshold, but this is easily affected by daily activities (such as sitting down quickly or jumping), resulting in a high false alarm rate.

[0004] Visual assistance solutions utilize cameras or infrared sensors, but are limited by lighting conditions, privacy concerns, and equipment deployment costs, making them difficult to apply in mobile scenarios (such as cycling and mountain climbing).

[0005] Smart bracelet / watch solution: It monitors wrist movements, but the connection between hand movements and body center of gravity is weak, and it is not sensitive to postures such as side falls and slips, resulting in a high rate of missed detection. Summary of the Invention

[0006] The purpose of this invention is to overcome at least one defect in the prior art and to provide a fall detection method, system and fall alarm method.

[0007] In a first aspect, the present invention discloses a fall detection method, comprising the following steps:

[0008] Acquire XYZ triaxial acceleration and XYZ triaxial angular velocity data from the inertial sensor;

[0009] Calculate the rate of change of attitude angle based on the XYZ triaxial acceleration and XYZ triaxial angular velocity;

[0010] Calculate the variance of triaxial acceleration and the peak value of the resultant acceleration based on the XYZ triaxial acceleration;

[0011] The determination of whether a fall has occurred is based on the calculated variance of triaxial acceleration, rate of change of attitude angle, and peak resultant acceleration.

[0012] In some embodiments, after acquiring the XYZ triaxial acceleration and XYZ triaxial angular velocity data from the inertial sensor, the method further includes a data preprocessing step for the acquired XYZ triaxial acceleration and XYZ triaxial angular velocity data.

[0013] Data preprocessing steps include filtering and noise reduction.

[0014] In some embodiments, determining whether a fall has occurred based on the calculated three-axis acceleration variance, attitude angle change rate, and resultant acceleration peak value specifically includes: comparing the calculated three-axis acceleration variance, attitude angle change rate, and resultant acceleration peak value with their respective set thresholds; if all three exceed the corresponding thresholds, a fall is determined.

[0015] In some embodiments, the attitude angle change rate is calculated based on the XYZ triaxial acceleration and XYZ triaxial angular velocity, specifically including:

[0016] Calculate the static initial attitude angle using triaxial acceleration;

[0017] Gyroscope angular velocity zero bias processing;

[0018] Perform Kalman filtering to correct the attitude angle and the dynamic zero bias of the gyroscope;

[0019] Based on the updated total zero bias estimate after Kalman filtering, the original angular velocity of the gyroscope is fully compensated to obtain a high-precision attitude angle change rate.

[0020] And / or,

[0021] The variance of triaxial acceleration is calculated based on the XYZ triaxial acceleration, specifically including:

[0022] Using a set sliding window length, the acceleration variance of each axis within the current sliding window is calculated iteratively point by point to obtain the single-axis acceleration variance of the X, Y, and Z axes;

[0023] Vector synthesis is performed on the triaxial acceleration data to obtain the triaxial resultant acceleration;

[0024] The variance of the triaxial resultant acceleration is calculated using the sliding window sample variance formula.

[0025] In some embodiments, the calculation of the peak resultant acceleration based on the XYZ triaxial acceleration specifically includes:

[0026] Preprocess the triaxial acceleration data;

[0027] Calculate the instantaneous value of the resultant acceleration at each time step: Based on the standardized triaxial data, calculate the resultant acceleration at each sampling point according to the vector magnitude formula;

[0028] Peak value extraction of the resultant acceleration sequence: After obtaining the instantaneous values ​​of the resultant acceleration at all sampling points, the maximum value in the sequence is extracted as the peak value.

[0029] Secondly, this invention discloses a fall alarm method, comprising the following steps:

[0030] Fall detection is performed using the fall detection method described in the first aspect;

[0031] An alarm is triggered when a fall is detected and the helmet is being worn.

[0032] In some embodiments, the fall alarm method of the present invention further includes the following steps: acquiring distance data from a ranging sensor and pressure data from a pressure sensor, and determining whether the helmet is being worn based on the distance data and pressure data;

[0033] The method for determining whether a helmet is being worn is based on distance and pressure data. Specifically, the method involves comparing the collected distance and pressure data with set distance and pressure thresholds, respectively. If the pressure data is greater than the set pressure threshold and the distance data is less than the set distance threshold, then the helmet is determined to be being worn.

[0034] In some embodiments, triggering an alarm procedure specifically includes: sending an alarm signal to a mobile app, so that the mobile app can start a countdown after receiving the alarm signal, and if no instruction to cancel the alarm is received within a set time, an alarm is sent to a preset contact.

[0035] Thirdly, the present invention discloses a fall detection system, including a terminal device and a smart helmet. The smart helmet is wirelessly connected to the terminal device. The smart helmet is equipped with a fall detection circuit, which includes a power supply circuit, a main control module, an inertial sensor, a pressure sensor, a distance sensor, and a wireless communication module. The power supply circuit is used to supply power to the entire fall detection circuit. The inertial sensor, pressure sensor, and distance sensor are electrically connected to the main control module, and the main control module is electrically connected to the wireless communication module.

[0036] The main control module is used to implement the fall detection method described in any of the above embodiments.

[0037] In some embodiments, the ranging sensor is a laser ranging sensor, which is electrically connected to the main control module via an I2C bus;

[0038] The SCL pin of the laser rangefinder is electrically connected to the first electrode of the first transistor, the second electrode of the first transistor is electrically connected to the SCL pin of the main control module, the control electrode of the first transistor is connected to the second voltage, the SDA pin of the laser rangefinder is electrically connected to the first electrode of the second transistor, the second electrode of the second transistor is electrically connected to the SDA pin of the main control module, and the control electrode of the second transistor is electrically connected to the second voltage.

[0039] In some embodiments, the SCL pin of the laser rangefinder is electrically connected to one end of resistor R4, and the other end of resistor R4 is electrically connected to a second voltage; the SDA pin of the laser rangefinder is electrically connected to one end of resistor R3, and the other end of resistor R3 is electrically connected to a second voltage.

[0040] And / or,

[0041] The second terminal of the first transistor is electrically connected to one end of resistor R2, and the other end of resistor R2 is electrically connected to the first voltage. The second terminal of the second transistor is electrically connected to one end of resistor R1, and the other end of resistor R1 is electrically connected to the first voltage.

[0042] And / or,

[0043] It also includes a first diode and a second diode. The positive terminal of the first diode is electrically connected to the first terminal of the first transistor, and the negative terminal of the first diode is electrically connected to the second terminal of the first transistor. The positive terminal of the second diode is electrically connected to the first terminal of the second transistor, and the negative terminal of the second diode is electrically connected to the second terminal of the second transistor.

[0044] In some embodiments, the output terminal of the pressure sensor is electrically connected to the ADC input terminal of the main control module, or the output terminal of the pressure sensor is electrically connected to the input terminal of the ADC module, and the output terminal of the ADC module is electrically connected to the input terminal of the main control module.

[0045] In some embodiments, the power supply circuit includes an LDO module, an input filter capacitor, and an output filter capacitor. The input terminal of the LDO module is electrically connected to a first voltage, and the output terminal of the LDO module is used to output a second voltage. The input terminal of the LDO module is electrically connected to one end of the input filter capacitor, and the other end of the input filter capacitor is grounded. The output terminal of the LDO module is electrically connected to one end of the output filter capacitor, and the other end of the output filter capacitor is grounded.

[0046] In some embodiments, the first voltage is provided by a battery.

[0047] In some embodiments, the wireless communication module includes at least one of a 4G communication module, a Bluetooth module, and a WIFI module.

[0048] In some embodiments, the fall detection system of the present invention further includes an alarm module, which is electrically connected to the main control module.

[0049] The beneficial effects of this invention are as follows: The fall detection method of this invention includes: acquiring XYZ triaxial acceleration and XYZ triaxial angular velocity data from an inertial sensor; calculating the attitude angle change rate based on the XYZ triaxial acceleration and XYZ triaxial angular velocity; calculating the triaxial acceleration variance and the peak value of the resultant acceleration based on the XYZ triaxial acceleration; and determining whether a fall has occurred based on the calculated triaxial acceleration variance, attitude angle change rate, and peak value of the resultant acceleration. This invention outputs a high-precision attitude angle change rate and feeds this high-precision change rate back to the Kalman filter prediction step (replacing the "preliminary compensated angular velocity" in the Kalman filter prediction step with this high-precision change rate), forming a closed-loop iteration (returning to execute the Kalman filter prediction step to achieve real-time closed-loop iteration). This invention can significantly improve the fall detection accuracy in complex scenarios. Attached Figure Description

[0050] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0051] Figure 1 This is a schematic diagram of a fall detection system provided in one embodiment of the present invention;

[0052] Figure 2 A circuit diagram of a ranging sensor provided in one embodiment of the present invention;

[0053] Figure 3 A circuit diagram of a pressure sensor provided in one embodiment of the present invention;

[0054] Figure 4 A circuit diagram of an inertial sensor provided in one embodiment of the present invention;

[0055] Figure 5 A circuit diagram of a power supply circuit provided in one embodiment of the present invention.

[0056] Figure 6 A flowchart of a fall alarm method provided in one embodiment of the present invention;

[0057] Figure 7 This is a schematic diagram of an interface for adding emergency callers in an APP according to an embodiment of the present invention;

[0058] Figure 8 This is a schematic diagram of a manual alarm cancellation interface provided in one embodiment of the present invention. Detailed Implementation

[0059] 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, and 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.

[0060] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature; in the description of this invention, unless otherwise stated, "a plurality of" or "several" means two or more.

[0061] See Figure 6 This invention discloses a fall detection method, comprising the following steps:

[0062] Acquire XYZ triaxial acceleration and XYZ triaxial angular velocity data from the inertial sensor;

[0063] Calculate the rate of change of attitude angle based on the XYZ triaxial acceleration and XYZ triaxial angular velocity;

[0064] Calculate the variance of triaxial acceleration and the peak value of the resultant acceleration based on the XYZ triaxial acceleration;

[0065] The determination of whether a fall has occurred is based on the calculated variance of triaxial acceleration, rate of change of attitude angle, and peak resultant acceleration.

[0066] In some embodiments, after acquiring the XYZ triaxial acceleration and XYZ triaxial angular velocity data from the inertial sensor, the method further includes a data preprocessing step for the acquired XYZ triaxial acceleration and XYZ triaxial angular velocity data.

[0067] In some embodiments, the data preprocessing steps include filtering and noise reduction.

[0068] In some embodiments, the data preprocessing step includes: performing sliding window filtering on the original acceleration signal and angular velocity signal to filter out high-frequency noise.

[0069] In some embodiments, the data preprocessing step includes: removing outliers, such as signals that are outside a reasonable range, to avoid extreme data affecting calculation accuracy.

[0070] In some embodiments, determining whether a fall has occurred based on the calculated three-axis acceleration variance, attitude angle change rate, and resultant acceleration peak value specifically includes: comparing the calculated three-axis acceleration variance, attitude angle change rate, and resultant acceleration peak value with their respective set thresholds; if all three exceed the corresponding thresholds, a fall is determined.

[0071] In some embodiments, a fall is determined only if all three conditions are met simultaneously: "variance exceeds the standard", "attitude angle change rate exceeds the standard", and "resultant acceleration peak exceeds the standard", and the duration is greater than or equal to the minimum judgment time (e.g., 100ms).

[0072] Calculate the variance of triaxial acceleration within a set time window, such as the mean of triaxial acceleration within the window, and calculate the variance of each axis. You can take the maximum value or the mean.

[0073] In some embodiments, the maximum value of the three-axis variance is taken as the judgment parameter and compared with the variance threshold: if the maximum value of the three-axis variance > the variance threshold, then it is marked as "variance exceeds the standard".

[0074] In some embodiments, the rate of change threshold comparison is performed: if the maximum value of the attitude angle rate of change is greater than the rate of change threshold, then "attitude angle rate of change exceeds the limit" is marked.

[0075] In some embodiments, the peak value of the resultant acceleration within the time window is calculated, and the maximum value of the resultant acceleration within the window is taken as the determination parameter.

[0076] Peak threshold comparison: If the peak resultant acceleration is greater than the peak threshold, then mark "Peak resultant acceleration exceeds the limit".

[0077] Feature data for fall detection (three-axis acceleration variance, attitude angle change rate, peak resultant acceleration, etc.) are extracted using an ICM20608 six-axis sensor. Initial thresholds are set: angular velocity > 180 ° / s and acceleration variance > 10 m. 2 / s 4 The peak resultant acceleration is > 20 m / s² 2 If the feature data exceeds the threshold, it is judged as a fall. The threshold can be dynamically adjusted based on the user's historical data.

[0078] In some embodiments, the attitude angle change rate is calculated based on the XYZ triaxial acceleration and XYZ triaxial angular velocity, specifically including:

[0079] Calculating the initial angle using triaxial acceleration, i.e., calculating the static initial attitude angle using triaxial acceleration;

[0080] Zero bias processing is used to remove the angular velocity bias of a gyroscope.

[0081] Perform Kalman filtering to correct the attitude angle and the dynamic zero bias of the gyroscope;

[0082] Calculate the rate of change of angular velocity.

[0083] Perform Kalman filtering to correct the attitude angle and gyroscope dynamic zero bias. Specifically, this includes using "attitude angle + gyroscope dynamic zero bias" as the state vector, the calculated initial attitude angle as the observation value, fusing the zero bias angular velocity, and correcting the gyroscope's cumulative error.

[0084] The calculation of the rate of change of angular velocity specifically includes: based on the total zero bias estimate updated by Kalman filtering, performing complete compensation on the original angular velocity of the gyroscope to obtain a high-precision rate of change of attitude angle.

[0085] Gyroscope angular velocity zero bias processing: Gyroscope static zero bias calibration and preliminary compensation, continuously collect raw angular velocity data (X / Y / Z axes) of the gyroscope in the static stage, calculate the arithmetic mean of the data of each axis as the static zero bias.

[0086] Initial static zero bias compensation: Subtract the corresponding axis static zero bias from the original angular velocity to obtain the initial debiasing angular velocity.

[0087] The statistical mean method eliminates fixed systematic errors. Under static conditions, the ideal output of the gyroscope should be 0, and the measured mean is the fixed zero bias.

[0088] Accelerometer calculation of static initial attitude angles (Kalman filter initialization preparation) provides initial observations (roll / pitch angle, pitch angle) for Kalman filtering. Yaw angle is initialized to 0 if there are no accelerometer observations.

[0089] The specific steps for calculating the static initial attitude angle using accelerometers are as follows: Based on the read static accelerometer data, normalization processing is performed (to eliminate amplitude errors):

[0090] The initial attitude angles are calculated based on gravity decomposition: roll, pitch, and yaw. In a static state, the accelerometer is only subjected to gravity, and the projection of the gravity vector onto the carrier coordinate system can uniquely determine roll / pitch.

[0091] Perform Kalman filtering, including Kalman filter initialization, Kalman filter prediction step, and Kalman filter update step.

[0092] Kalman filter initialization, including:

[0093] Define the state vector (simultaneously estimating attitude angles and gyroscope dynamic bias):

[0094] Initialize the state covariance matrix P: Initially, it is an identity matrix (1s on the diagonal and 0s off-diagonal).

[0095] Initialize the state transition matrix F;

[0096] Initialize the input matrix G (correlated with gyroscope input and state transition), the observation matrix H (observe only roll / pitch), the process noise Q (the degree of trust in the gyroscope, the smaller the value, the greater the trust), and the observation noise R (the degree of trust in the accelerometer, the smaller the value, the greater the trust).

[0097] Kalman filtering state-space modeling uses "attitude angle + dynamic zero bias" as the joint state to achieve multivariate collaborative estimation.

[0098] Kalman filter initialization: Specifically initializes the state vector and matrix of the Kalman filter, laying the foundation for dynamic zero bias estimation and attitude angle correction.

[0099] The Kalman filter prediction step (high-precision rate of change for attitude angle prediction) specifically includes: using the obtained initial compensated angular velocity to predict the attitude angle and dynamic zero bias at the next moment (corresponding to the "output high-precision attitude angle rate of change → for attitude angle prediction" step).

[0100] State prediction (based on angular velocity after preliminary compensation):

[0101] in, This is the state estimate from the previous time step. This is the angular velocity after initial compensation.

[0102] Kalman filter update step (updating zero bias and attitude angle by combining accelerometer observations): Correct the prediction error with the real-time attitude angle (observation value) calculated by the accelerometer, and update the attitude angle and gyroscope dynamic zero bias (corresponding to the step of "updating zero bias and attitude angle by combining accelerometer observations").

[0103] This invention integrates static and dynamic zero bias to fully compensate for the original angular velocity, resulting in a high-precision attitude angle change rate (corresponding to the "Kalman filter dynamic zero bias estimation → total zero bias full compensation" step).

[0104] This invention reads the original angular velocity via SPI → preliminary static zero bias compensation → dynamic zero bias estimation via Kalman filtering ([Kalman prediction step: high-precision rate of change predicts attitude angle] → [Kalman update step: accelerometer observation updates attitude angle + dynamic zero bias]) → complete total zero bias compensation (complete total zero bias (static + dynamic) compensation step) → outputs high-precision attitude angle rate of change step → feeds back to the Kalman prediction step (closed loop) for attitude angle prediction, and updates the zero bias and attitude angle in combination with accelerometer observations.

[0105] This invention simultaneously estimates the attitude angle and gyroscope dynamic zero bias using Kalman filtering, and compensates the original gyroscope data with the total zero bias to obtain a high-precision attitude angle change rate. This rate is then used for attitude angle prediction and updating, ultimately achieving attitude angle correction and cumulative error suppression. Total zero bias = static zero bias + dynamic zero bias. This invention incorporates the gyroscope dynamic zero bias into the Kalman state vector, using a two-dimensional compensation approach of "static zero bias as a baseline + dynamic zero bias real-time tracking," combined with accelerometer observation constraints. This eliminates both fixed hardware errors and suppresses cumulative errors caused by time-varying drift, ultimately outputting a high-precision, highly stable attitude angle change rate.

[0106] In some embodiments, calculating the variance of triaxial acceleration based on the XYZ triaxial acceleration specifically includes:

[0107] Data preprocessing is performed on continuous sampling data based on a triaxial accelerometer (e.g., 100 samples per second at 100Hz).

[0108] By setting a sliding window length and using the sample variance formula and incremental update algorithm, the acceleration variance of each axis within the current sliding window is calculated point by point to obtain the single-axis acceleration variance of the X-axis, Y-axis and Z-axis;

[0109] Vector synthesis is performed on the triaxial acceleration data to obtain the triaxial resultant acceleration;

[0110] The variance of the triaxial combined acceleration, i.e., the triaxial combined variance, is calculated using the sliding window sample variance formula.

[0111] In some embodiments, the calculation of the peak resultant acceleration based on the XYZ triaxial acceleration specifically includes:

[0112] Preprocess the triaxial acceleration data;

[0113] Calculate the instantaneous value of the resultant acceleration at each time step: Based on the standardized triaxial data, calculate the resultant acceleration at each sampling point according to the vector magnitude formula;

[0114] Peak value extraction of the resultant acceleration sequence: After obtaining the instantaneous values ​​of the resultant acceleration at all sampling points, the maximum value in the sequence is extracted as the peak value.

[0115] Based on the same inventive concept, this invention discloses a fall alarm method, comprising the following steps:

[0116] Fall detection is performed using the fall detection method described in any of the above embodiments;

[0117] An alarm is triggered when a fall is detected and the helmet is being worn.

[0118] This invention simultaneously satisfies the requirements of being worn and triggering an alarm system upon a fall. This system sends an alarm signal to a mobile app via Bluetooth / 4G.

[0119] In some embodiments, the fall alarm method of the present invention further includes the following steps: acquiring distance data from a ranging sensor and pressure data from a pressure sensor, and determining whether the helmet is being worn based on the distance data and pressure data;

[0120] The method for determining whether a helmet is being worn is based on distance and pressure data. Specifically, the method involves comparing the collected distance and pressure data with set distance and pressure thresholds, respectively. If the pressure data is greater than the set pressure threshold and the distance data is less than the set distance threshold, then the helmet is determined to be being worn.

[0121] In some embodiments, triggering an alarm procedure specifically includes: sending an alarm signal to a mobile app, so that the mobile app can start a countdown after receiving the alarm signal, and if no instruction to cancel the alarm is received within a set time, an alarm is sent to a preset contact.

[0122] This invention allows users to manually confirm the alarm (users can manually cancel the alarm). Figure 8 This is a schematic diagram of a manual alarm cancellation interface provided in one embodiment of the present invention. Users can manually cancel the alarm via the APP within 30 seconds (the time is configurable) after sending the alarm signal. After cancellation, the system automatically returns to the data acquisition phase. If there is no response within 30 seconds, a multi-level alarm mechanism is triggered, including: a buzzer alarm; sending an alarm and locating emergency contacts; and detecting the alarm and locating the emergency center. Figure 7 This is a schematic diagram of the emergency caller addition interface provided in one embodiment of the present invention; the present invention can also realize system linkage: GPS positioning data is synchronized to the cloud platform and automatically pushed to emergency contacts and hospitals or emergency centers.

[0123] See Figure 1 This invention discloses a fall detection system, including a terminal device and a smart helmet. The smart helmet is wirelessly connected to the terminal device. The smart helmet is equipped with a fall detection circuit, which includes a power supply circuit, a main control module, an inertial sensor, a pressure sensor, a distance sensor, and a wireless communication module. The power supply circuit supplies power to the entire fall detection circuit. The inertial sensor, pressure sensor, and distance sensor are electrically connected to the main control module, and the main control module is electrically connected to the wireless communication module.

[0124] The main control module is used to implement the fall detection method described in any of the above embodiments.

[0125] This invention employs a multi-sensor fusion detection scheme, which significantly improves detection accuracy and call reliability in complex scenarios through hardware-layer redundancy design and multi-modal communication assurance mechanism.

[0126] In some embodiments, the ranging sensor is a laser ranging sensor, which is electrically connected to the main control module via an I2C bus.

[0127] In some embodiments, the inertial sensor is a six-axis sensor.

[0128] In some embodiments, the main control module employs a processor.

[0129] In some embodiments, see Figure 2 The SCL pin of the laser rangefinder is electrically connected to the first electrode of the first transistor Q1A, the second electrode of the first transistor Q1A is electrically connected to the SCL pin of the main control module, the control electrode of the first transistor Q1A is connected to the second voltage, the SDA pin of the laser rangefinder is electrically connected to the first electrode of the second transistor Q1B, the second electrode of the second transistor Q1B is electrically connected to the SDA pin of the main control module, and the control electrode of the second transistor Q1B is connected to the second voltage.

[0130] In some embodiments, see Figure 2 The SCL pin of the laser rangefinder is electrically connected to one end of resistor R4, and the other end of resistor R4 is electrically connected to the second voltage. The SDA pin of the laser rangefinder is electrically connected to one end of resistor R3, and the other end of resistor R3 is electrically connected to the second voltage.

[0131] In some embodiments, see Figure 2 The second terminal of the first transistor is electrically connected to one end of resistor R2, and the other end of resistor R2 is electrically connected to the first voltage. The second terminal of the second transistor is electrically connected to one end of resistor R1, and the other end of resistor R1 is electrically connected to the first voltage.

[0132] In some embodiments, see Figure 2 The fall detection circuit of the present invention further includes a first diode and a second diode. The positive terminal of the first diode is electrically connected to the first terminal of the first transistor, the negative terminal of the first diode is electrically connected to the second terminal of the first transistor, the positive terminal of the second diode is electrically connected to the first terminal of the second transistor, and the negative terminal of the second diode is electrically connected to the second terminal of the second transistor.

[0133] See Figure 2The laser rangefinder sensor's shutdown control pin is electrically connected to one end of resistor R6, and the other end of resistor R6 is electrically connected to the second voltage. The laser rangefinder sensor's general-purpose input / output pin is electrically connected to one end of resistor R5, and the other end of resistor R5 is electrically connected to the second voltage. The laser rangefinder sensor's XSHUT (shutdown control pin) needs to be connected to a GPIO output pin of the MCU: the MCU controls the sensor to enter sleep (low level) or wake-up (high level) state by outputting a high or low level to this pin, thus achieving power consumption management. The laser rangefinder sensor's GPIO1 (general-purpose input / output pin) is usually connected to the MCU's GPIO input (or interrupt input) pin: taking the VL53LXX sensor as an example, this pin is often configured as a "range completion interrupt" function. After the sensor completes a range measurement, it outputs a level signal through GPIO1 to trigger the MCU's interrupt response, allowing the MCU to read the measurement data promptly.

[0134] In some embodiments, see Figure 3 The output terminal of the pressure sensor is electrically connected to the ADC input terminal of the main control module, or the output terminal of the pressure sensor is electrically connected to the input terminal of the ADC module, and the output terminal of the ADC module is electrically connected to the input terminal of the main control module. The first output terminal of the pressure sensor is connected to a third voltage (e.g., +5V), and the second output terminal of the pressure sensor is grounded via a resistor. The second output terminal of the pressure sensor is electrically connected to the ADC input terminal of the main control module, or the second output terminal of the pressure sensor is electrically connected to the input terminal of the ADC module, and the output terminal of the ADC module is electrically connected to the input terminal of the main control module.

[0135] In some embodiments, the main control module uses a main control chip of model RV1126 and an inertial sensor of model ICM20608. Figure 4 The circuit diagram shows an inertial sensor provided in one embodiment of the present invention. It employs an FSR402 thin-film pressure sensor and a VL53L0X laser rangefinder. The fall detection circuit of the present invention uses I2C, SPI, and ADC protocols as its main hardware communication protocols.

[0136] In some embodiments, see Figure 5 The power supply circuit includes an LDO module, an input filter capacitor, and an output filter capacitor. The input terminal of the LDO module is electrically connected to a first voltage, and the output terminal of the LDO module is used to output a second voltage. The input terminal of the LDO module is electrically connected to one end of the input filter capacitor, and the other end of the input filter capacitor is grounded. The output terminal of the LDO module is electrically connected to one end of the output filter capacitor, and the other end of the output filter capacitor is grounded.

[0137] In some embodiments, the first voltage is provided by a battery.

[0138] In some embodiments, the fall detection circuit of the present invention further includes an alarm module, which is electrically connected to the main control module. The alarm module can be an audible alarm module or a visual alarm module. Preferably, the alarm module is an audible alarm module, such as a buzzer.

[0139] In some embodiments, the wireless communication module includes at least one of a 4G communication module, a Bluetooth module, and a WIFI module.

[0140] The terminal device can be a mobile phone.

[0141] In some embodiments, distance sensors are installed at multiple (e.g., 3) key points (top, left and right sides) in the helmet liner, with the probes facing the head (≤3cm when wearing, ≥5cm when not wearing); pressure sensors are installed at pressure points (forehead, back of the head, and cheekbones on both sides) in the helmet liner, such as 3-4 in total, to ensure that at least 2 sensors can be squeezed by the head when wearing (to avoid missing detection at a single pressure point).

[0142] This invention uses multiple sensors to collect data, and then the main control module performs noise reduction and preprocessing on the data. It acquires XYZ triaxial acceleration and XYZ triaxial angular velocity data from the ICM20608 inertial sensor via the SPI protocol. The initial angle is calculated using the triaxial acceleration, the gyroscope angular velocity is zero-biased, Kalman filtering is performed, and the rate of change of angular velocity is calculated. The peak value of the triaxial combined acceleration is calculated from the triaxial acceleration data. Continuous sampling data is acquired using the triaxial accelerometer, such as 100 samples per second at 100Hz, to calculate the single-axis variance and the triaxial combined variance. The VL53L0X laser rangefinder is connected to a 3.3V power supply pin and communicates via the I2C protocol. After SPAD (offset correction) calibration, the distance data is acquired. This invention uses a 10-bit ADC to convert the pressure magnitude of the FSR402 thin-film pressure sensor into voltage data, calculates the voltage-pressure linear curve, and obtains the pressure data. This invention determines the helmet wearing status (e.g., voltage <0.5V, distance <2cm) based on distance and pressure data collected by sensors, setting initial thresholds of angular velocity > 180° / s and acceleration variance > 10 m. 2 / s 4 The peak resultant acceleration is >20 m / s². 2 If the characteristic data (variance of three-axis acceleration, rate of change of attitude angle, peak resultant acceleration, etc.) exceeds the threshold, it is judged as a fall. The threshold will be dynamically adjusted based on the user's historical data.

[0143] This invention proposes a system based on a multi-sensor fusion detection and dynamic adaptive algorithm for smart helmets. Through hardware-level redundancy design, software-level false alarm suppression strategy, and multi-modal communication guarantee mechanism, it significantly improves the detection accuracy and emergency call reliability in complex scenarios. The smart helmet system achieves high-precision fall detection and automatically triggers graded emergency calls through multi-modal sensor data fusion.

[0144] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A fall detection method, characterized in that, Includes the following steps: Acquire XYZ triaxial acceleration and XYZ triaxial angular velocity data from the inertial sensor; Calculate the rate of change of attitude angle based on the XYZ triaxial acceleration and XYZ triaxial angular velocity; Calculate the variance of triaxial acceleration and the peak value of the resultant acceleration based on the XYZ triaxial acceleration; The determination of whether a fall has occurred is based on the calculated variance of triaxial acceleration, rate of change of attitude angle, and peak resultant acceleration.

2. The fall detection method as described in claim 1, characterized in that: After acquiring the XYZ triaxial acceleration and XYZ triaxial angular velocity data from the inertial sensor, the process also includes a data preprocessing step for the acquired XYZ triaxial acceleration and XYZ triaxial angular velocity data.

3. The fall detection method as described in claim 1, characterized in that: The determination of whether a fall has occurred is based on the calculated three-axis acceleration variance, attitude angle change rate, and resultant acceleration peak value. Specifically, the calculated three-axis acceleration variance, attitude angle change rate, and resultant acceleration peak value are compared with their respective set thresholds. If all three exceed the corresponding thresholds, a fall is determined.

4. The fall detection method as described in claim 1 or 3, characterized in that: The attitude angle change rate is calculated based on the XYZ triaxial acceleration and XYZ triaxial angular velocity, specifically including: Calculate the static initial attitude angle using triaxial acceleration; Gyroscope angular velocity zero bias processing; Perform Kalman filtering to correct the attitude angle and the dynamic zero bias of the gyroscope; Based on the updated total zero bias estimate after Kalman filtering, the original angular velocity of the gyroscope is fully compensated to obtain a high-precision attitude angle change rate. And / or, The variance of triaxial acceleration is calculated based on the XYZ triaxial acceleration, specifically including: Using a set sliding window length, the acceleration variance of each axis within the current sliding window is calculated iteratively point by point to obtain the single-axis acceleration variance of the X, Y, and Z axes; Vector synthesis is performed on the triaxial acceleration data to obtain the triaxial resultant acceleration; The variance of the triaxial resultant acceleration is calculated using the sliding window sample variance formula.

5. A fall alarm method, characterized in that, Includes the following steps: Fall detection is performed using the fall detection method as described in any one of claims 1 to 4; An alarm is triggered when a fall is detected and the helmet is being worn.

6. The fall alarm method as described in claim 5, characterized in that: It also includes the following steps: acquiring distance data from the range sensor and pressure data from the pressure sensor, and determining whether the helmet is being worn based on the distance data and pressure data; The method for determining whether a helmet is being worn is based on distance and pressure data. Specifically, the method involves comparing the collected distance and pressure data with set distance and pressure thresholds, respectively. If the pressure data is greater than the set pressure threshold and the distance data is less than the set distance threshold, then the helmet is determined to be being worn.

7. The fall alarm method as described in claim 5, characterized in that: The alarm trigger procedure specifically includes: sending an alarm signal to the mobile app, so that the mobile app can start a countdown after receiving the alarm signal. If no instruction to cancel the alarm is received within the set time, an alarm is sent to the preset contact.

8. A fall detection system, characterized in that: The device includes a terminal device and a smart helmet. The smart helmet is wirelessly connected to the terminal device. The smart helmet is equipped with a fall detection circuit, which includes a power supply circuit, a main control module, an inertial sensor, a pressure sensor, a distance sensor, and a wireless communication module. The power supply circuit supplies power to the entire fall detection circuit. The inertial sensor, pressure sensor, and distance sensor are electrically connected to the main control module, and the main control module is electrically connected to the wireless communication module.

9. The fall detection system as described in claim 8, characterized in that: The ranging sensor is a laser ranging sensor, which is electrically connected to the main control module via an I2C bus; The SCL pin of the laser rangefinder is electrically connected to the first electrode of the first transistor, the second electrode of the first transistor is electrically connected to the SCL pin of the main control module, the control electrode of the first transistor is connected to the second voltage, the SDA pin of the laser rangefinder is electrically connected to the first electrode of the second transistor, the second electrode of the second transistor is electrically connected to the SDA pin of the main control module, and the control electrode of the second transistor is electrically connected to the second voltage.

10. The fall detection system as described in claim 9, characterized in that: The SCL pin of the laser rangefinder is electrically connected to one end of resistor R4, and the other end of resistor R4 is electrically connected to the second voltage; the SDA pin of the laser rangefinder is electrically connected to one end of resistor R3, and the other end of resistor R3 is electrically connected to the second voltage. And / or, The second terminal of the first transistor is electrically connected to one end of resistor R2, and the other end of resistor R2 is electrically connected to the first voltage. The second terminal of the second transistor is electrically connected to one end of resistor R1, and the other end of resistor R1 is electrically connected to the first voltage. And / or, It also includes a first diode and a second diode. The positive terminal of the first diode is electrically connected to the first terminal of the first transistor, and the negative terminal of the first diode is electrically connected to the second terminal of the first transistor. The positive terminal of the second diode is electrically connected to the first terminal of the second transistor, and the negative terminal of the second diode is electrically connected to the second terminal of the second transistor.

Images