Mobile monitoring equipment and method for triggering abnormal event

Abstract

A mobile monitoring device and a method of triggering an abnormal event, the mobile monitoring device being integrated with a motion sensor (20) for collecting motion data of a target object, further identifying an activity pattern of the target object according to the motion data, automatically recording an abnormal event of a physiological parameter if the activity pattern belongs to an abnormal activity pattern, abnormal events can be used as a basis for subsequent health state analysis. Compared with manual active operation of keys to trigger and record abnormal events, the mobile monitoring equipment is more suitable for animal scenes, and the problem that the animals cannot actively record the abnormal events in the mobile monitoring field is solved.

Description

Mobile monitoring device and method for triggering abnormal events Technical Field

[0001] The present application relates to the field of mobile monitoring technology, more precisely, to a mobile monitoring device and a method for triggering abnormal events. Background Art

[0002] A wearable ECG recorder is a portable device that can be worn on a subject to record their heart activity over a long period of time. The use of wearable ECG recorders for animals differs from that for humans: When a human uses a wearable ECG recorder, they can automatically record ECG events by pressing a button if they detect any abnormalities in their health.

[0003] However, wearable ECG recorders for animals are mainly used on animals, and the above-mentioned scheme for triggering ECG events cannot be applied to wearable ECG recorders for animals.

[0004] Summary of the Invention

[0005] In view of this, the present application provides a mobile monitoring device and a method for triggering abnormal events, where the device automatically records abnormal events of the physiological parameters of the monitored object to solve the problem that the monitored object cannot trigger abnormal events independently.

[0006] In a first aspect, an embodiment of the present application provides a mobile monitoring device, comprising:

[0007] a physiological parameter acquisition module, configured to acquire parameter data of target physiological parameters of a target subject, the mobile monitoring device being worn by the target subject;

[0008] A motion sensor, configured to collect motion data of the target object;

[0009] A processing device is communicatively connected to the physiological parameter acquisition module and the motion sensor, and is configured to: obtain motion data of the target object and determine activity characteristics of the target object based on the motion data; input the activity characteristics of the target object into a pre-trained activity pattern recognition model, and have the activity pattern recognition model recognize the activity pattern of the target object based on the activity characteristics; if the activity pattern belongs to a preset abnormal pattern, generate an abnormal event of the target physiological parameter based on the occurrence time of the activity pattern.

[0010] In a second aspect, an embodiment of the present application provides a method for triggering an abnormal event, including:

[0011] obtaining motion data of the target animal, and determining activity characteristics of the target animal based on the motion data;

[0012] Inputting the activity characteristics of the target animal into a pre-trained activity pattern recognition model, and allowing the activity pattern recognition model to recognize the activity pattern of the target animal based on the activity characteristics;

[0013] If the activity pattern belongs to a preset abnormal pattern, an abnormal event of the target physiological parameter is generated according to the occurrence time of the activity pattern.

[0014] As can be seen from the above technical solutions, the mobile monitoring device and method for triggering abnormal events provided in the embodiments of the present application can collect motion data of a target subject, identify the target subject's activity pattern based on the motion data, and automatically record abnormal physiological parameter events when the activity pattern is abnormal. Abnormal events can serve as a basis for subsequent health status analysis. Compared with manually triggering abnormal event recording by keystroke, this solution is more suitable for animal use scenarios and solves the problem that animals cannot actively record abnormal events. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0016] FIG1 is a schematic structural diagram of a mobile monitoring device provided in an embodiment of the present application;

[0017] FIG2 is a diagram illustrating an example of the placement of a motion sensor according to an embodiment of the present application;

[0018] FIG3 is a flow chart of a method for triggering an abnormal event according to an embodiment of the present application;

[0019] FIG4 is a schematic diagram of a topological structure of a neural network model provided in an embodiment of the present application;

[0020] FIG5 is a schematic diagram of an operation process of a mobile monitoring device provided in an embodiment of the present application;

[0021] FIG6 is a schematic diagram of a process for triggering an ECG event according to an embodiment of the present application;

[0022] FIG7 is a schematic diagram of an interface showing ECG events and activity patterns according to an embodiment of the present application;

[0023] FIG8 is a schematic diagram of an interface for displaying electrocardiogram events provided in an embodiment of the present application. DETAILED DESCRIPTION

[0024] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0025] During their research on wearable ECG recorders for animals, the inventors discovered that while the product form factor of these devices is similar to that of human wearable ECG recorders, some of the features of these devices cannot be easily and directly applied to animal devices. For example, animals cannot actively interact with the devices, which can result in ECG events not being recorded, thus hindering the subsequent use of ECG data to diagnose the animal's condition.

[0026] Of course, the inventors also note that this problem is not limited to ECG recorders, but can also occur with other medical devices used to monitor animal health. Furthermore, the abnormal events that need to be recorded are not limited to ECG events, but can also include abnormal events of any other physiological parameters. Therefore, the present invention provides a mobile monitoring device that automatically records abnormal events.

[0027] As shown in Figure 1, the mobile monitoring device provided in the embodiment of the present application includes: a physiological parameter acquisition module 10, a motion sensor 20, and a processing device 30. It is understood that the mobile monitoring device may also include conventional structures such as a communication module, a storage unit, a power supply, or other units or structures desired by those skilled in the art.

[0028] The mobile monitoring device is worn by a target subject, which can be a human or an animal. The physiological parameter acquisition module 10 is used to collect parameter data of target physiological parameters of the target subject. The target physiological parameters may include: electrocardiogram (ECG) parameters, respiratory parameters, blood oxygen parameters, blood pressure parameters, body temperature parameters, etc. Taking ECG signals as an example, the mobile monitoring device may include a wearable vest, and the physiological parameter acquisition module 10 may include ECG lead electrodes. The animal wears the vest to carry the mobile monitoring device, and the ECG lead electrodes are placed on designated parts of the animal's body to collect the animal's ECG signals in real time. The ECG signals are processed to obtain ECG waveform data and heart rate values.

[0029] The motion sensor 20 may include one or more sensors for collecting motion data of the target object. For example, the motion sensor 20 may also be set on the animal's body by wearing a vest, and the motion data of the monitored object is collected by the motion sensor 20. In some embodiments, the motion sensor 20 may include an accelerometer and a gyroscope to collect acceleration data and / or angular velocity data of the animal. Taking Figure 2 as an example, three motion sensors a, b and c can be placed on the puppy's neck, back and abdomen respectively to collect motion data of different body parts of the puppy. Of course, the placement and number of the motion sensors in the figure are merely examples, and the embodiments of the present application do not specifically limit them.

[0030] The processing device 30 is communicatively connected to the physiological parameter acquisition module 10 and the motion sensor 20 and is configured to execute the abnormal event triggering method shown in FIG3 , which specifically includes steps S301-S303. The processing device 30 can be a unit with data processing capabilities, such as a processor or controller, and can be electrically connected to the physiological parameter acquisition module 10 and the motion sensor 20 to enable data communication between them.

[0031] S301: Obtain motion data of a target object, and determine activity characteristics of the target object based on the motion data.

[0032] After the motion sensor 20 collects the motion data of the target object, it can send the motion data to the processing device 30, which analyzes the motion data and determines the activity characteristics of the target object. The activity characteristics represent the activity (vitality or movement) of the target object.

[0033] In one specific implementation, the activity features include posture features and / or activity features. Posture features can reflect the target object's body posture, such as standing, lying down, sitting, walking, etc.; activity features can reflect the target object's motion state, such as walking, running, hitting, squatting, etc. The motion data collected by the motion sensor may include acceleration data and / or angular velocity data. Thus, the target object's posture features can be determined based on the target object's acceleration data and / or angular velocity data; and the target object's activity features can be determined based on the target object's acceleration data.

[0034] The specific calculation process for posture features: A motion sensor can periodically collect motion data of a target object at a certain acquisition frequency. The motion data is arranged into a set of motion data according to the acquisition time. As the number of acquisitions increases, motion data is continuously added to the motion data set. If the motion data includes multiple types, a set of motion data corresponding to each type can be obtained. Taking acceleration and angular velocity as examples, a set of acceleration data and a set of angular velocity data can be obtained, and the two sets of motion data have a corresponding relationship in terms of acquisition time. The posture features of the target object can be calculated periodically. Within a calculation cycle, an acquisition window can be used to extract a certain amount of motion data from each set of motion data, such as 512 angular velocity data and 512 acceleration data. The extracted motion data are then transformed into angular velocity coordinates to obtain the posture features for the current cycle. Of course, when calculating the posture features for the next cycle, the acquisition window can be slid backward to extract a new set of motion data. The motion data extracted between the two acquisition windows may or may not include repeated motion data. The posture features can be expressed as attitude angles. For example, the posture features include at least one of the following: heading angle, pitch angle, and roll angle. It should be noted that the posture features may be obtained in other ways that can be thought of by those skilled in the art, and the embodiments of the present application do not specifically limit this.

[0035] The specific calculation process of activity characteristics: The motion sensor can collect the vertical acceleration and horizontal acceleration of the target object. After processing a set of vertical accelerations, multiple vertical acceleration measurement results can be obtained. After processing a set of horizontal accelerations, multiple horizontal acceleration measurement results can be obtained. After comprehensively processing the multiple vertical acceleration measurement results and the multiple horizontal acceleration measurement results, multiple combined acceleration measurement results can be obtained. For example, the activity characteristics include at least one of the following: the average maximum combined acceleration, the standard deviation of the combined acceleration, the average combined acceleration, the average maximum vertical acceleration, the standard deviation of the vertical acceleration, the average vertical acceleration, the average maximum horizontal acceleration, the standard deviation of the horizontal acceleration, and the average horizontal acceleration.

[0036] S302: Inputting the activity features of the target object into a pre-trained activity pattern recognition model, and allowing the activity pattern recognition model to recognize the activity pattern of the target object based on the activity features.

[0037] Specifically, the embodiments of the present application can use multiple groups of training samples to pre-train an activity pattern recognition model. The activity pattern recognition model can be a machine learning model, and the machine learning model includes but is not limited to a neural network model. The training samples are sample activity features, and each group of sample activity features can have an activity pattern label. The trained activity pattern recognition model has the ability to recognize activity patterns based on the activity features of any object. After the activity features of the target object are input into the activity pattern recognition model, the activity pattern recognition model can output the activity pattern corresponding to the activity features based on the activity features of the target object. It should be noted that in some embodiments, the training samples of the activity pattern recognition model may not have activity pattern labels, and the activity pattern recognition model learns the differences between various activity patterns. In some embodiments, activity patterns may include but are not limited to fainting, falling, fright, and collision.

[0038] It's understandable that animals behave differently in different environments, such as familiar or unfamiliar ones. To ensure the accuracy of activity patterns, different activity pattern recognition models can be trained for different environments. For example, indoors can be considered a safe zone, while outdoors can be considered a dangerous zone. Therefore, separate activity pattern recognition models can be trained for safe and dangerous zones.

[0039] Different activity pattern recognition models can have the same structure but different model parameters. Taking a neural network model as an example, different transfer parameter matrices are trained for different environments. The transfer parameter matrix can include two types of parameters: correlation coefficients Wij and bias components bj. Correlation coefficients represent the influence of model parameters in the previous layer of the neural network model on the computational results of the next layer, while bias components represent the extent to which a model parameter in the same layer of the neural network model influences the computational results of that layer. Taking the neural network model shown in Figure 4 as an example, the correlation coefficient of input layer i can represent its influence on hidden layer j. Similarly, the correlation coefficient of hidden layer j can represent its influence on output layer k. A bias component in input layer i, such as the second bias component from the top, can represent its influence on the computational results of the entire input layer i. In practical applications, activity pattern recognition models can be constructed by selecting appropriate model parameters based on the target object's environment. Specifically, within the model parameters pre-trained based on a machine learning model, target model parameters corresponding to the target object's activity area are determined; these target model parameters are then applied to the machine learning model to obtain an activity pattern recognition model.

[0040] To accurately determine the target subject's activity zone, as shown in FIG1 , the mobile monitoring device may further include a positioning module 40, such as a Global Positioning System (GPS). The positioning module 40 may be worn by the target subject to determine the target subject's real-time location. The processing device 30 is in communication with the positioning module 40 and can obtain the target subject's location data collected by the positioning module 40, thereby determining the target subject's activity zone based on the location data.

[0041] Specifically, the target object's activity area can be pre-set, such as an indoor area or an outdoor area, or further divided into a feeding area, a play area, a rest area, etc. The processing device 30 can pre-record the area information of each activity area, such as the specific range and area size. After obtaining the real-time location data collected by the positioning module 40, the target object's current activity area is determined based on the real-time location data, such as an outdoor area. The model parameters corresponding to the activity area are then obtained from the pre-trained model parameters, and the model parameters are then used to construct an activity pattern recognition model.

[0042] After the activity features obtained in step S302 are input into the activity pattern recognition model, the activity pattern recognition model outputs the recognition result of the activity pattern. For example, the above three posture features and nine activity degree features (a total of 12 feature parameters) are used as x1-x 12 The neural network model shown in Figure 4 is input and outputs the recognition probabilities y1-y5. The activity pattern corresponding to the maximum recognition probability is used as the target object's activity pattern. The activity patterns y1-y5 represent: impact, fainting, fright, fall, and normal, respectively.

[0043] In actual applications, if an animal is in an abnormal situation such as being frightened, it will often make special calls, or some abnormal sounds may appear in the external environment to frighten the animal, causing it to behave abnormally. Therefore, in order to further improve the accuracy of the activity pattern recognition results, environmental sound data can also be added for identification together with the activity features. Therefore, as shown in Figure 1, the mobile monitoring device can also include a sound collection module 50. The sound collection module 50 is carried on the target object and is used to collect environmental sound data of the environment in which the target object is located. The processing device 30 is also connected to the sound collection module 50 for communication, obtains environmental sound data, and extracts sound features based on the environmental sound data, such as sound features including at least one of the following: Mel-frequency cepstral coefficients, zero-crossing rate; the sound features are input into the activity pattern recognition model, and the activity pattern recognition model identifies the activity pattern of the target object based on the activity features and sound features. As shown in Figure 4, the two sound features are combined with the above-mentioned three posture features and 9 activity features (a total of 14 feature parameters) as x1-x 14Input a neural network model to recognize the activity patterns of the target object.

[0044] S303: If the activity pattern belongs to a preset abnormal pattern, an abnormal event of the target physiological parameter is generated according to the occurrence time of the activity pattern.

[0045] The target subject's activity pattern can be divided into two categories: normal mode and abnormal mode. Abnormal modes include shock, fainting, collision, and falling. Abnormal modes may be accompanied by abnormalities in target physiological parameters, such as the occurrence of ECG events in ECG parameters. To record abnormal events, if it is determined that the activity pattern belongs to a preset abnormal mode, an abnormal event for the target physiological parameter can be automatically generated. It should be noted that the time of occurrence of the abnormal event plays an important role in the subsequent diagnostic analysis of the target subject. Therefore, when generating abnormal events for the target physiological parameter, the time of occurrence of the abnormal event is mainly recorded. Of course, the time of occurrence of the abnormal event is the time of occurrence of the activity pattern. For example, if the target subject experiences a fainting activity pattern at 15:27, an ECG event for the ECG parameters is generated at 15:27. One way to generate an ECG event is to generate a time tag on the ECG parameter to mark the time at which an ECG event occurred. Of course, abnormal events of physiological parameters other than ECG parameters can also be generated, such as an apnea event for respiratory parameters, etc. This embodiment of the application does not specifically limit this.

[0046] To facilitate more accurate analysis of abnormal conditions in target subjects by medical personnel, abnormal events of target physiological parameters can be associated and stored with activity patterns. Therefore, the mobile monitoring device can further include a storage module, with the processing device 30 in communication with the storage module. The abnormal events and activity patterns are associated and stored based on the time of occurrence of the activity pattern, for example: 10:15: impact - ECG event; 10:17: syncope - apnea event. This allows medical personnel to refer to the activity pattern information when analyzing abnormal events for a more comprehensive analysis.

[0047] As can be seen from the above technical solutions, the embodiments of the present application provide a mobile monitoring device with an integrated motion sensor for collecting motion data of a target subject. The device then identifies the target subject's activity pattern based on the motion data. If the activity pattern is abnormal, the device automatically records abnormal physiological parameter events as a basis for subsequent health status analysis. Compared to manually triggering the recording of abnormal events with keystrokes, this mobile monitoring device is more suitable for animal use, resolving the dilemma of animals being unable to actively record abnormal events.

[0048] In practical applications, after powering on a mobile monitoring device, it can perform routine testing procedures, such as a self-test. If the self-test fails, a self-test fault alarm is issued. If the self-test passes, further monitoring can be performed. As shown in Figure 5, this can monitor whether the power button is activated; if so, the device shuts down. It can also monitor battery life; if not, a low-battery alarm is issued. It can also monitor whether the Bluetooth connection is lost; if so, a Bluetooth connection prompt is issued. It can also monitor whether a lead is disconnected; if so, a lead disconnection alarm is issued. If not, ECG data collection is initiated, and both ECG data collection and ECG event triggering are initiated. Figure 6 shows a flow chart of the ECG event triggering process. Angular velocity and acceleration data are collected, and after data preprocessing, the posture angle is calculated to obtain posture features. Acceleration data is collected, and after data preprocessing, the vectorial synthesis of accelerations (total acceleration, vertical acceleration, and horizontal acceleration) is calculated to obtain activity features. Ambient sound data is collected, and after data preprocessing, the features of the ambient sound are extracted to obtain sound features. In addition, positioning data is collected and pre-processed to determine whether it is within a preset activity zone. If so, a transfer parameter matrix within the zone is set to generate an activity pattern recognition model. If outside the zone, a transfer parameter matrix outside the zone is set to generate another activity pattern recognition model. Posture features, activity features, and sound features are input into the activity pattern recognition model, which identifies activity patterns such as fainting, startle, falls, and collisions. These activity patterns are considered abnormal. Once identified as abnormal, an electrocardiogram (ECG) event can be triggered, allowing medical personnel to analyze the animal's health status based on this ECG event.

[0049] In actual applications, medical staff may need to check whether the animal has exhibited abnormal behavior over the past period of time. In order to meet this viewing requirement, the mobile monitoring device can also display abnormality-related data in the form of an interface. One possible implementation is that the mobile monitoring device includes a communication module such as Bluetooth, SIM (Subscriber Identity Module) or other short-range or long-range wireless communication modules. The processing device 30 is communicatively connected to the communication module and is further configured to: in response to an abnormal event viewing instruction, determine abnormal parameter data corresponding to the abnormal event in the parameter data of the target physiological parameter; send the abnormal parameter data of the abnormal event and the activity pattern associated with the abnormal event to the display device through the communication module, so that the display device displays the abnormal parameter data and the activity pattern in association.

[0050] Specifically, when a user wishes to review an abnormal event that occurred on a target subject, they can send an abnormal event review instruction to the mobile monitoring device. For example, the user can send the abnormal event review instruction to the mobile monitoring device via a computer device such as a mobile terminal, desktop computer, or laptop computer that is in communication with the mobile monitoring device. After receiving the abnormal event review instruction, the mobile monitoring device can provide the user with two types of data related to the abnormal event: abnormal parameter data and the identified activity pattern. The display device can display these two types of data in a linked manner. The display device can be the display device of the aforementioned computer device.

[0051] It should be noted that the correlation between these two types of data is reflected in the time of occurrence. That is, if the target subject exhibits a certain activity pattern at a certain point in time, the parameter data of the target physiological parameter of the target subject may be abnormal at that time, and the abnormal parameter data at that time corresponds to an abnormal event. As mentioned earlier, after the mobile monitoring device identifies an abnormal activity pattern, it can generate an abnormal event for the target physiological parameter. Therefore, when searching for abnormal parameter data, the abnormal event can be found first, and the parameter data corresponding to the time of the abnormal event (which may be a period of time) can be determined as the abnormal parameter data.

[0052] Taking Figure 7 as an example, after the user sends a command to view an abnormal event of an animal to the mobile monitoring device through the terminal, it is assumed that the mobile monitoring device finds four activity patterns, namely, a fall that occurred at 2022-11-03 10:13:47, a fright that occurred at 2022-11-03 10:13:47, a fainting that occurred at 2022-11-03 10:14:48, and a collision that occurred at 2022-11-03 10:15:18. The corresponding ECG parameters are extracted at the time of each occurrence as abnormal ECG parameters, including ECG waveform data and heart rate (HR). The starting time point of the ECG waveform data is the occurrence time corresponding to the activity pattern. The terminal displays the name, occurrence time, and corresponding abnormal ECG parameters of each activity pattern. In this way, the user can intuitively view the activity patterns and abnormal ECG parameters corresponding to all ECG events, and the information relied on for the diagnosis and analysis of the animal's abnormal behavior is more comprehensive, and the analysis results are more accurate. It is understandable that this interface only displays the abnormal ECG parameters corresponding to the ECG event. If the user wants to view more comprehensive ECG parameters, they can select any ECG event in Figure 7. For example, if they select the ECG event in the first row, the display device can jump to the display interface shown in Figure 8, which can display the more complete ECG waveform data. Time information is displayed above the ECG waveform data, and the time of the ECG event, such as "10:15:18", is highlighted to indicate that an ECG event occurred at that time point. The user slides the ECG waveform data, and the display interface can display the ECG waveform data at the previous or next time point.

[0053] Furthermore, different activity modes can be divided into different danger levels to facilitate the subsequent review of activity modes according to the danger level. Specifically, the mobile monitoring device can also include a storage module, the processing device is in communication with the storage module, and is further configured to: determine the danger level of the activity mode, record the activity mode and the danger level of the activity mode. For example, the default danger level of fainting and collision can be higher, while the danger level of fright and fall can be lower. Of course, other forms of danger levels can also be set. In this way, the activity modes corresponding to different danger levels can be viewed later based on the danger level indicator.

[0054] The mobile monitoring device can also be equipped with an active triggering function, that is, the recording of abnormal events can be triggered manually remotely or closely. Specifically, the processing device 30 of the mobile monitoring device is further configured to: receive an event recording instruction, and generate an abnormal event of the target physiological parameter according to the time when the event recording instruction is received. The event recording instruction is generated by a physical trigger button of the mobile monitoring device. For example, if the user finds that the target object has an abnormal activity pattern, the physical button on the mobile monitoring device is actively triggered to record the abnormal event; or, the event recording instruction is received through the communication module of the mobile monitoring device. For example, if the user finds that the target object has an abnormal activity pattern, the event recording instruction can be remotely sent to the mobile monitoring device through a terminal or other device to trigger the recording of the abnormal event.

[0055] In addition, the present invention also provides a method for triggering an abnormal event, including:

[0056] obtaining motion data of the target animal, and determining activity characteristics of the target animal based on the motion data;

[0057] The activity characteristics of the target animal are input into a pre-trained activity pattern recognition model, and the activity pattern recognition model recognizes the activity pattern of the target animal based on the activity characteristics; in a specific implementation method, the activity pattern is any one of the activity patterns of collision, fainting, fright, fall, and normal.

[0058] If the activity pattern belongs to a preset abnormal pattern, an abnormal event of the target physiological parameter is generated according to the occurrence time of the activity pattern. In a specific implementation, the physiological parameter acquisition module includes an electrocardiogram module, the parameter data includes electrocardiogram waveform data, and the abnormal event is an electrocardiogram event.

[0059] In a specific implementation, identifying the activity pattern of the target animal according to the activity features by the activity pattern recognition model includes:

[0060] Obtaining the environmental sound data, and extracting sound features based on the environmental sound data;

[0061] The sound feature is input into the activity pattern recognition model, and the activity pattern recognition model recognizes the activity pattern of the target object according to the activity feature and the sound feature.

[0062] In a specific implementation, the motion data includes acceleration data and / or angular velocity data, and the activity characteristics include posture characteristics and / or activity characteristics; and determining the activity characteristics of the target object based on the motion data includes:

[0063] Determine the posture characteristics of the target object based on the acceleration data and / or angular velocity data of the target object; and determine the activity characteristics of the target object based on the acceleration data of the target object.

[0064] Among them, the posture characteristics include at least one of the following: heading angle, pitch angle, roll angle; the activity characteristics include at least one of the following: average maximum value of combined acceleration, standard deviation of combined acceleration, average value of combined acceleration, average maximum value of vertical acceleration, standard deviation of vertical acceleration, average value of vertical acceleration, average maximum value of horizontal acceleration, standard deviation of horizontal acceleration, average value of horizontal acceleration; the sound characteristics include at least one of the following: Mel-frequency cepstral coefficient, zero-crossing rate.

[0065] In a specific implementation, before the step of inputting the activity features of the target object into a pre-trained activity pattern recognition model is executed, the method further includes:

[0066] Determine target model parameters corresponding to the target object's activity area from model parameters pre-trained based on a machine learning model; and assign the target model parameters to the machine learning model to obtain the activity pattern recognition model. In one specific implementation, the activity pattern recognition model includes a neural network model trained using multiple sets of sample activity features, each of which has a corresponding activity pattern label.

[0067] In a specific implementation, before determining the target model parameters corresponding to the activity area of ​​the target object, the method further includes: obtaining position data of the target object, and determining the activity area of ​​the target object according to the position data.

[0068] In a specific implementation, the method for triggering an abnormal event further includes: storing the abnormal event and the activity pattern in association with each other according to the occurrence time of the activity pattern.

[0069] In a specific implementation, the method for triggering an abnormal event also includes: in response to an abnormal event viewing instruction, determining abnormal parameter data corresponding to the abnormal event in the parameter data; and associating and displaying the abnormal parameter data of the abnormal event and an activity pattern that has an associated relationship with the abnormal event.

[0070] In a specific implementation, the method for triggering an abnormal event further includes: determining a danger level of the activity pattern, and recording the activity pattern and the danger level of the activity pattern.

[0071] In a specific implementation, the method for triggering an abnormal event further includes: receiving an event recording instruction, and generating an abnormal event of the target physiological parameter according to the reception time of the event recording instruction.

[0072] It should be noted that for the relevant description of the method of triggering abnormal events, please refer to the corresponding content of the mobile monitoring device and will not be repeated here.

[0073] It should be noted that the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to in detail. With respect to the above description of the disclosed embodiments, the features described in the various embodiments in this specification can be replaced or combined with each other, so that professional and technical personnel in this field can implement or use this application.

[0074] It should also be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or device comprising the aforementioned elements.

[0075] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A mobile monitoring device, characterized in that: include: a physiological parameter acquisition module, configured to acquire parameter data of target physiological parameters of a target subject, the mobile monitoring device being worn by the target subject; A motion sensor, configured to collect motion data of the target object; A processing device is communicatively connected to the physiological parameter acquisition module and the motion sensor, and is configured to: obtain motion data of the target object and determine activity characteristics of the target object based on the motion data; input the activity characteristics of the target object into a pre-trained activity pattern recognition model, and have the activity pattern recognition model recognize the activity pattern of the target object based on the activity characteristics; if the activity pattern belongs to a preset abnormal pattern, generate an abnormal event of the target physiological parameter based on the occurrence time of the activity pattern.

2. The mobile monitoring device according to claim 1, characterized in that: Also includes a sound collection module; The sound collection module is used to collect environmental sound data of the environment where the target object is located; The processing device is also communicatively connected to the sound collection module and is configured to obtain the environmental sound data, extract sound features based on the environmental sound data; input the sound features into the activity pattern recognition model, and have the activity pattern recognition model recognize the activity pattern of the target object based on the activity features and the sound features.

3. The mobile monitoring device according to claim 1 or 2, characterized in that: The motion data includes acceleration data and / or angular velocity data, and the activity characteristics include posture characteristics and / or activity characteristics; when performing the step of determining the activity characteristics of the target object based on the motion data, the processing device is specifically configured as follows: determining a posture feature of the target object based on the acceleration data and / or angular velocity data of the target object; An activity characteristic of the target object is determined based on the acceleration data of the target object.

4. The mobile monitoring device according to claim 3, characterized in that: The posture feature includes at least one of the following: heading angle, pitch angle, and roll angle; The activity characteristics include at least one of the following: average maximum value of combined acceleration, standard deviation of combined acceleration, average value of combined acceleration, average maximum value of vertical acceleration, standard deviation of vertical acceleration, average value of vertical acceleration, average maximum value of horizontal acceleration, standard deviation of horizontal acceleration, and average value of horizontal acceleration; The sound feature includes at least one of the following: Mel-frequency cepstral coefficient and zero-crossing rate.

5. The mobile monitoring device according to claim 1, characterized in that: Before executing the step of inputting the activity features of the target object into a pre-trained activity pattern recognition model, the processing device is specifically configured to: Determining target model parameters corresponding to the activity area of ​​the target object from model parameters pre-trained based on a machine learning model; The target model parameters are configured in the machine learning model to obtain the activity pattern recognition model.

6. The mobile monitoring device according to claim 5, characterized in that: Also includes positioning module; The processing device is in communication with the positioning module, and is further configured to: obtain the position data of the target object collected by the positioning module, and determine the activity area of ​​the target object according to the position data.

7. The mobile monitoring device according to claim 1, characterized in that: The activity pattern recognition model includes a neural network model, which is trained by multiple groups of sample activity features, and the sample activity features have corresponding activity pattern labels.

8. The mobile monitoring device according to claim 1, characterized in that: Also included is a storage module; The processing device is in communication with the storage module and is further configured to: store the abnormal event and the activity pattern in association with each other according to the occurrence time of the activity pattern.

9. The mobile monitoring device according to claim 1, characterized in that: Also included is a communication module; The processing device is communicatively connected to the communication module and is further configured to: determine abnormal parameter data corresponding to the abnormal event in the parameter data in response to an abnormal event viewing instruction; and send the abnormal parameter data of the abnormal event and an activity pattern associated with the abnormal event to the display device through the communication module, so that the display device associates and displays the abnormal parameter data and the activity pattern.

10. The mobile monitoring device according to claim 1, characterized in that: Also included is a storage module; The processing device is in communication with the storage module and is further configured to: determine a danger level of the activity pattern, and record the activity pattern and the danger level of the activity pattern.

11. The mobile monitoring device according to claim 1, characterized in that: The processing device is further configured to: An event recording instruction is received, and an abnormal event of the target physiological parameter is generated according to a time when the event recording instruction is received.

12. The mobile monitoring device according to claim 11, characterized in that: The event recording instruction is generated by a physical trigger button of the mobile monitoring device; or The event recording instruction is received through the communication module of the mobile monitoring device.

13. The mobile monitoring device according to claim 1, characterized in that: The activity mode is any one of impact, fainting, fright, fall, and normal activity mode.

14. The mobile monitoring device according to claim 1, characterized in that: The physiological parameter acquisition module includes an electrocardiogram module, the parameter data includes electrocardiogram waveform data, and the abnormal event is an electrocardiogram event.

15. A method for triggering an abnormal event, characterized in that: include: obtaining motion data of the target animal, and determining activity characteristics of the target animal based on the motion data; Inputting the activity characteristics of the target animal into a pre-trained activity pattern recognition model, and allowing the activity pattern recognition model to recognize the activity pattern of the target animal based on the activity characteristics; If the activity pattern belongs to a preset abnormal pattern, an abnormal event of the target physiological parameter is generated according to the occurrence time of the activity pattern.

16. The method for triggering an abnormal event according to claim 15, characterized in that: Identifying the activity pattern of the target animal according to the activity characteristics by the activity pattern recognition model includes: Obtaining the environmental sound data, and extracting sound features based on the environmental sound data; The sound feature is input into the activity pattern recognition model, and the activity pattern recognition model recognizes the activity pattern of the target object according to the activity feature and the sound feature.

17. The method for triggering an abnormal event according to claim 15 or 16, characterized in that: The motion data includes acceleration data and / or angular velocity data, and the activity characteristics include posture characteristics and / or activity characteristics. Determining the activity characteristics of the target object based on the motion data includes: determining a posture feature of the target object based on the acceleration data and / or angular velocity data of the target object; An activity characteristic of the target object is determined based on the acceleration data of the target object.

18. The method for triggering an abnormal event according to claim 17, characterized in that: The posture feature includes at least one of the following: heading angle, pitch angle, and roll angle; The activity characteristics include at least one of the following: average maximum value of combined acceleration, standard deviation of combined acceleration, average value of combined acceleration, average maximum value of vertical acceleration, standard deviation of vertical acceleration, average value of vertical acceleration, average maximum value of horizontal acceleration, standard deviation of horizontal acceleration, and average value of horizontal acceleration; The sound feature includes at least one of the following: Mel-frequency cepstral coefficient and zero-crossing rate.

19. The method for triggering an abnormal event according to claim 15, characterized in that: Before executing the step of inputting the activity features of the target object into a pre-trained activity pattern recognition model, the method further includes: Determining target model parameters corresponding to the activity area of ​​the target object from model parameters pre-trained based on a machine learning model; The target model parameters are configured in the machine learning model to obtain the activity pattern recognition model.

20. The method for triggering an abnormal event according to claim 19, characterized in that: Before determining the target model parameters corresponding to the activity area of ​​the target object, the method further includes: The position data of the target object is obtained, and the activity area of ​​the target object is determined according to the position data.

21. The method for triggering an abnormal event according to claim 15, wherein: The activity pattern recognition model includes a neural network model, which is trained by multiple groups of sample activity features, and the sample activity features have corresponding activity pattern labels.

22. The method for triggering an abnormal event according to claim 15, characterized in that: Also includes: The abnormal event and the activity pattern are associated and stored according to the occurrence time of the activity pattern.

23. The method for triggering an abnormal event according to claim 15, characterized in that: Also includes: In response to an abnormal event viewing instruction, determining abnormal parameter data corresponding to the abnormal event in the parameter data; The abnormal parameter data of the abnormal event and the activity pattern associated with the abnormal event are displayed in association.

24. The method for triggering an abnormal event according to claim 15, characterized in that: Also includes: Determine the danger level of the activity pattern, and record the activity pattern and the danger level of the activity pattern.

25. The method for triggering an abnormal event according to claim 15, characterized in that: Also includes: An event recording instruction is received, and an abnormal event of the target physiological parameter is generated according to a time when the event recording instruction is received.

26. The method for triggering an abnormal event according to claim 15, characterized in that: The activity mode is any one of impact, fainting, fright, fall, and normal activity mode.

27. The method for triggering an abnormal event according to claim 15, characterized in that: The physiological parameter acquisition module includes an electrocardiogram module, the parameter data includes electrocardiogram waveform data, and the abnormal event is an electrocardiogram event.

Images