A false alarm prevention method and system based on user behavior self-learning of a loss prevention device

Abstract

The application provides a false alarm prevention method and system based on user behavior self-learning of a loss-prevention device. In the active alarm mode, dynamic data of the bound terminal is monitored and recorded. When an active alarm event of the loss-prevention device occurs, a first alarm-related event sequence within a first preset time range after the active alarm event of the loss-prevention device is identified. Whether the active alarm event is a false alarm event is determined according to the first alarm-related event sequence. When it is determined that the active alarm event is a false alarm event, a second alarm-related event sequence within a second preset time range before the active alarm event of the loss-prevention device is identified. The behavior mode of the user within the second preset time range is analyzed based on the second alarm-related event sequence. The behavior mode is recorded as a false alarm-prone behavior mode. The false alarm problem of the loss-prevention device can be effectively reduced.

Description

Technical Field

[0001] This invention relates to the field of anti-loss device technology, and in particular to a method and system for preventing false alarms based on user behavior self-learning of anti-loss devices. Background Technology

[0002] An anti-loss device is an electronic product placed inside items such as personal belongings, clothing, backpacks, and travel gear. It uses low-power communication technology to enable one-click retrieval, remote positioning, and lost item alerts. It effectively prevents items from being forgotten or lost, and allows for quick retrieval if items are lost. There are many types of anti-loss devices on the market. Based on the activation method of the alarm, anti-loss devices can be divided into active alarm anti-loss devices and passive alarm anti-loss devices. Active alarm anti-loss devices typically have a specific alarm distance range or geographical location range on a map. When the anti-loss device moves further away from the phone than this alarm distance range, or when its location exceeds this geographical range, an alarm is triggered. Anti-loss devices that are tied to a phone's alarm distance range are particularly prone to false alarms due to dynamic environmental changes. For example, communication signals may be affected by strong electromagnetic fields, or obstructed by obstacles such as metal or people, leading to communication interruptions or large distance measurement errors. Summary of the Invention

[0003] Based on the above-mentioned problems, this invention proposes a method and system for preventing false alarms based on user behavior self-learning of anti-loss devices, which can effectively reduce the false alarm problem of anti-loss devices.

[0004] In view of this, the first aspect of the present invention proposes a method for preventing false alarms based on user behavior self-learning of anti-loss devices, comprising:

[0005] The anti-loss device is configured to active alarm mode according to the user's operation instructions. In active alarm mode, an alarm is triggered when the distance between the anti-loss device and the bound terminal is greater than the pre-configured alarm distance.

[0006] Monitor and record dynamic data with the bound terminal, wherein the dynamic data of the bound terminal is data reflecting the real-time status of the bound terminal and / or the real-time association status between the bound terminal and the anti-loss device;

[0007] When the anti-loss device triggers an active alarm event, a first alarm-related event sequence is identified within a first preset time range after the anti-loss device triggers the active alarm event. The first alarm-related event sequence consists of dynamic data and / or related events of dynamic data changes of the anti-loss device within the first preset time range after the active alarm event.

[0008] Determine whether the active alarm event is a false alarm event based on the first alarm-related event sequence;

[0009] When it is determined that the active alarm event is a false alarm event, a second alarm-related event sequence within a second preset time range before the active alarm event of the anti-loss device is identified. The second alarm-related event sequence is the dynamic data and / or dynamic data change-related events of the anti-loss device within the second preset time range before the active alarm event.

[0010] Based on the second alarm-related event sequence analysis, the user's behavior pattern within the second preset time range is analyzed. The behavior pattern is a sequence of actions consisting of dynamic data or user behavior corresponding to changes in dynamic data within the second preset time range.

[0011] The aforementioned behavior pattern is recorded as a false alarm-prone behavior pattern.

[0012] Furthermore, after the step of recording the behavior pattern as a false alarm-prone behavior pattern, the method further includes:

[0013] Monitor the first distance d1 between the anti-loss device and the bound terminal;

[0014] Determine whether the first distance d1 between the anti-loss device and the bound terminal is greater than the pre-configured alarm distance d0;

[0015] When the first distance d1 between the anti-loss device and the bound terminal is greater than the alarm distance d0, the difference between the first distance d1 and the alarm distance d0 between the anti-loss device and the bound terminal is calculated:

[0016] Δd = d1 - d0;

[0017] Determine the user's current behavior pattern and the false alarm deviation threshold Δd0 between the anti-loss device and the current behavior pattern;

[0018] Determine whether the difference between the first distance d1 between the anti-loss device and the bound terminal and the alarm distance d0 is less than or equal to the false alarm deviation threshold Δd0;

[0019] When the difference between the first distance d1 between the anti-loss device and the bound terminal and the alarm distance d0 is less than or equal to the false alarm deviation threshold Δd0, it is determined whether the user's current behavior pattern is a false alarm-prone behavior pattern.

[0020] When the user's current behavior pattern is not a false alarm-prone behavior pattern, the anti-loss device is controlled to issue an alarm.

[0021] Furthermore, the method for preventing false alarms based on user behavior self-learning of the anti-loss device also includes:

[0022] When precise positioning is available, the anti-loss device and the bound terminal are precisely located, and the user's behavior patterns are monitored.

[0023] The second distance d2 between the anti-loss device and the bound terminal is obtained by measuring the distance through the wireless communication signal between the anti-loss device and the bound terminal;

[0024] Determine whether the second distance d2 between the anti-loss device and the bound terminal falls within the preset ranging range;

[0025] When the second distance d2 between the anti-loss device and the binding terminal falls within the preset ranging range, the ranging error between the anti-loss device and the binding terminal is calculated based on the accurate positioning result;

[0026] The ranging error is associated with and stored in relation to the user's behavior patterns.

[0027] Furthermore, after the step of associating and storing the ranging error with the user's behavior pattern, the method further includes:

[0028] Periodically iterate through the ranging error corresponding to each behavior pattern;

[0029] Remove outlier values ​​from the ranging error;

[0030] The average distance measurement error after removing outliers for each behavior pattern is determined as the false alarm bias threshold Δd0 for the corresponding behavior pattern.

[0031] Furthermore, after recording the behavior pattern as a false alarm-prone behavior pattern, the method further includes storing the action sequence and dynamic data corresponding to the false alarm-prone behavior pattern in a database. The specific steps for determining whether the user's current behavior pattern is a false alarm-prone behavior pattern include:

[0032] Obtain the action sequence corresponding to the user's current behavior pattern;

[0033] Match the database with action sequences that are prone to false alarms;

[0034] When a false alarm-prone behavior pattern with the same action sequence as the user's current behavior pattern exists in the database, the matching false alarm-prone behavior pattern will be identified as the target behavior pattern.

[0035] Obtain the first dynamic data corresponding to the user's current behavior pattern;

[0036] Read the second dynamic data of the target behavior pattern from the database;

[0037] Calculate the similarity between the first dynamic data and the second dynamic data;

[0038] When the similarity between the first dynamic data and the second dynamic data is greater than a preset similarity threshold, the user's current behavior pattern is determined to be a false alarm-prone behavior pattern.

[0039] Furthermore, the step of identifying the first alarm-related event sequence within a first preset time range after the anti-loss device triggers an active alarm event specifically includes:

[0040] Acquire the first dynamic data within the first preset time range;

[0041] A first event sequence within the first preset time range is generated based on the first dynamic data and / or the changes in the first dynamic data. The first event sequence consists of one or more of the following: user operation events on the bound terminal and / or user behavior causing position change events, motion change events, or posture change events of the bound terminal.

[0042] Extract alarm-related events from the first event sequence to generate the first alarm-related event sequence.

[0043] Furthermore, the step of determining whether the active alarm event is a false alarm event based on the first alarm-related event sequence specifically includes:

[0044] The first alarm-related event sequence is input into a pre-trained anti-loss device false alarm identification model to identify whether the active alarm event is a false alarm event. The anti-loss device false alarm identification model is a convolutional neural network model used to perform data classification.

[0045] Furthermore, the step of identifying the second alarm-related event sequence within a second preset time range prior to the occurrence of an active alarm event by the anti-loss device specifically includes:

[0046] Acquire the second dynamic data within the second preset time range;

[0047] A second event sequence is generated within the second preset time range based on the second dynamic data and / or changes in the second dynamic data. The second event sequence consists of one or more of the following: position change events, motion change events, and posture change events of the bound terminal caused by the user's behavior.

[0048] Extract alarm-related events from the second event sequence to generate the second alarm-related event sequence.

[0049] Furthermore, the step of analyzing the user's behavioral patterns within the second preset time range based on the second alarm-related event sequence specifically includes:

[0050] The second alarm-related event sequence is input into a pre-trained behavior pattern recognition model to identify the user's behavior pattern within a second preset time range. The false alarm recognition model of the anti-loss device is a convolutional neural network model used to perform data classification.

[0051] A second aspect of the present invention proposes a false alarm prevention system based on user behavior self-learning of an anti-loss device, comprising an anti-loss device and a binding terminal, wherein the false alarm prevention system is configured to implement the false alarm prevention method based on user behavior self-learning of an anti-loss device as described in any of the first aspects of the present invention.

[0052] This invention proposes a method and system for preventing false alarms based on user behavior self-learning of anti-loss devices. In the active alarm mode, dynamic data of the bound terminal is monitored and recorded. When the anti-loss device triggers an active alarm event, a first alarm-related event sequence within a first preset time range after the active alarm event is identified. Based on the first alarm-related event sequence, it is determined whether the active alarm event is a false alarm event. If the active alarm event is determined to be a false alarm event, a second alarm-related event sequence within a second preset time range before the active alarm event is identified. Based on the second alarm-related event sequence, the user's behavior pattern within the second preset time range is analyzed, and the behavior pattern is recorded as a false alarm-prone behavior pattern, which can effectively reduce the false alarm problem of the anti-loss device. Attached Figure Description

[0053] Figure 1 This is a flowchart of a method for preventing false alarms based on user behavior self-learning of an anti-loss device, provided by an embodiment of the present invention. Detailed Implementation

[0054] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0055] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0056] In the description of this invention, the term "multiple" refers to two or more. Unless otherwise explicitly defined, the terms "upper," "lower," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. The terms "connect," "install," "fix," etc., should be interpreted broadly. For example, "connect" can be a fixed connection, a detachable connection, or an integral connection; it can be a direct connection or an indirect connection through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances. Furthermore, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "multiple" means two or more.

[0057] In the description of this specification, the terms "one embodiment," "some implementations," "specific embodiment," etc., refer to a specific feature, structure, material, or characteristic described in connection with that embodiment or example, which is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0058] The technical solution of this invention is applicable to anti-loss devices with an active alarm mode. More specifically, the anti-loss device provided by this invention prevents false alarms caused by internal and external factors in active alarm mode through user behavior learning. The following describes, with reference to the accompanying drawings, a method and system for preventing false alarms based on user behavior self-learning of an anti-loss device according to some embodiments of the present invention.

[0059] like Figure 1 As shown, the first aspect of the present invention proposes a method for preventing false alarms based on user behavior self-learning of anti-loss devices, comprising:

[0060] The anti-loss device is configured to active alarm mode according to the user's operation instructions. In active alarm mode, an alarm is triggered when the distance between the anti-loss device and the bound terminal is greater than the pre-configured alarm distance.

[0061] Monitor and record dynamic data with the bound terminal, wherein the dynamic data of the bound terminal is data reflecting the real-time status of the bound terminal and / or the real-time association status between the bound terminal and the anti-loss device;

[0062] When the anti-loss device triggers an active alarm event, a first alarm-related event sequence is identified within a first preset time range after the anti-loss device triggers the active alarm event. The first alarm-related event sequence consists of dynamic data and / or related events of dynamic data changes of the anti-loss device within the first preset time range after the active alarm event.

[0063] Determine whether the active alarm event is a false alarm event based on the first alarm-related event sequence;

[0064] When it is determined that the active alarm event is a false alarm event, a second alarm-related event sequence within a second preset time range before the active alarm event of the anti-loss device is identified. The second alarm-related event sequence is the dynamic data and / or dynamic data change-related events of the anti-loss device within the second preset time range before the active alarm event.

[0065] Based on the second alarm-related event sequence analysis, the user's behavior pattern within the second preset time range is analyzed. The behavior pattern is a sequence of actions consisting of dynamic data or user behavior corresponding to changes in dynamic data within the second preset time range.

[0066] The aforementioned behavior pattern is recorded as a false alarm-prone behavior pattern.

[0067] Specifically, the binding terminal is a mobile terminal bound to the anti-loss device. The user manages and controls the anti-loss device through an application on the binding terminal, such as configuring the anti-loss device's working mode, configuring the alarm distance of the anti-loss device in active alarm mode, and controlling the anti-loss device to emit an alarm sound through the binding terminal when the anti-loss device is in passive alarm mode.

[0068] In some embodiments of the present invention, the dynamic data of the bound terminal includes, but is not limited to, real-time location data, real-time motion state data, real-time posture data, and usage state data of the bound terminal, reflecting the real-time state of the bound terminal. The real-time location data reflects the real-time location of the bound terminal, which may be GPS coordinates or location coordinates obtained through other positioning technologies. The real-time motion state data includes one or more of the following: motion speed, motion acceleration, motion direction, and motion distance, reflecting the motion state of the bound terminal. The real-time posture data is data detected by attitude sensors such as gyroscopes, reflecting the placement posture of the bound terminal. The usage state data includes data reflecting the current usage state of the bound terminal, such as screen-on status, foreground application name, and touch status.

[0069] In some embodiments of the present invention, the dynamic data of the binding terminal also includes data reflecting the real-time association status between the binding terminal and the anti-loss device, such as the communication connection status between the binding terminal and the anti-loss device, the strength of the wireless signal emitted by the anti-loss device detected by the binding terminal in real time, and the content and time of the control commands sent by the binding terminal to the anti-loss device. The communication connection status between the binding terminal and the anti-loss device refers to the state in which the binding terminal and the anti-loss device establish a communication connection through wireless communication signals such as Bluetooth or Wi-Fi. For example, in some cases, when the distance between the two is close enough that their communication modules can detect each other's communication signals, the communication connection status is "connected." Conversely, when the distance between the two is too far to establish a communication connection, the communication connection status is "disconnected."

[0070] The active alarm event is an event that triggers an alarm when the distance between the anti-loss device and the bound mobile terminal exceeds a pre-configured alarm distance.

[0071] The first preset time range is a time range starting from the time the anti-loss device triggers an active alarm event, with a pre-configured first duration as the time length. Similarly, the second preset time range is a time range ending at the time the anti-loss device triggers an active alarm event, with a pre-configured second duration as the time length. The first duration and the second duration are pre-configured empirical durations; they can be the same length or configured independently as different durations as needed. The first duration needs to be long enough to generate enough alarm-related events within the first preset time range to determine whether the active alarm event is a false alarm, but it cannot be too long to introduce events unrelated to the active alarm event within the first preset time range. Similarly, the second duration needs to be long enough to generate enough alarm-related events within the second preset time range to identify the user's behavioral patterns before the active alarm event, but it cannot be too long to introduce events unrelated to the active alarm event within the second preset time range.

[0072] Different user behavior patterns are constituted by different user actions. For example, user actions constituting these patterns include picking up a phone, turning the screen on or off, opening an application, checking the location of the anti-loss device, sending an alarm-off command to the anti-loss device, carrying the bound device close to or away from the anti-loss device, and putting or taking the bound device into a pocket or backpack. When the active alarm event is a false alarm, one or more user actions within the second preset time range may be key actions leading to the false alarm.

[0073] Furthermore, after the step of recording the behavior pattern as a false alarm-prone behavior pattern, the method further includes:

[0074] Monitor the first distance d1 between the anti-loss device and the bound terminal;

[0075] Determine whether the first distance d1 between the anti-loss device and the bound terminal is greater than the pre-configured alarm distance d0;

[0076] When the first distance d1 between the anti-loss device and the bound terminal is greater than the alarm distance d0, the difference between the first distance d1 and the alarm distance d0 between the anti-loss device and the bound terminal is calculated:

[0077] Δd = d1 - d0;

[0078] Determine the user's current behavior pattern and the false alarm deviation threshold Δd0 between the anti-loss device and the current behavior pattern;

[0079] Determine whether the difference between the first distance d1 between the anti-loss device and the bound terminal and the alarm distance d0 is less than or equal to the false alarm deviation threshold Δd0;

[0080] When the difference between the first distance d1 between the anti-loss device and the bound terminal and the alarm distance d0 is less than or equal to the false alarm deviation threshold Δd0, it is determined whether the user's current behavior pattern is a false alarm-prone behavior pattern.

[0081] When the user's current behavior pattern is not a false alarm-prone behavior pattern, the anti-loss device is controlled to issue an alarm.

[0082] Specifically, the binding terminal calculates the distance between the binding terminal and the anti-loss device using the Received Signal Strength Indication (RSSI) value of the wireless communication signal of the anti-loss device, or the propagation time of the wireless communication signal between the binding terminal and the anti-loss device.

[0083] Furthermore, after determining whether the first distance d1 between the anti-loss device and the bound terminal is greater than the pre-configured alarm distance d0, the method further includes:

[0084] When the first distance d1 between the anti-loss device and the bound terminal is less than or equal to the alarm distance d0, return to the step of monitoring the first distance d1 between the anti-loss device and the bound terminal.

[0085] Furthermore, after determining whether the difference between the first distance d1 between the anti-loss device and the bound terminal and the alarm distance d0 is less than or equal to the false alarm deviation threshold Δd0, the method further includes:

[0086] When the difference between the first distance d1 between the anti-loss device and the bound terminal and the alarm distance d0 is greater than the false alarm deviation threshold Δd0, the anti-loss device is controlled to issue an alarm.

[0087] Furthermore, after determining whether a user's behavior pattern is a false positive behavior pattern, the following steps are also included:

[0088] When the user's behavior pattern is a false alarm-prone behavior pattern, return to the step of monitoring the first distance d1 between the anti-loss device and the bound terminal.

[0089] Furthermore, the method for preventing false alarms based on user behavior self-learning of the anti-loss device also includes:

[0090] When precise positioning is available, the anti-loss device and the bound terminal are precisely located, and the user's behavior patterns are monitored.

[0091] The second distance d2 between the anti-loss device and the bound terminal is obtained by measuring the distance through the wireless communication signal between the anti-loss device and the bound terminal;

[0092] Determine whether the second distance d2 between the anti-loss device and the bound terminal falls within the preset ranging range;

[0093] When the second distance d2 between the anti-loss device and the binding terminal falls within the preset ranging range, the ranging error between the anti-loss device and the binding terminal is calculated based on the accurate positioning result;

[0094] The ranging error is associated with and stored in relation to the user's behavior patterns.

[0095] In some embodiments of the present invention, the conditions for accurate positioning can be that the anti-loss device and the bound terminal are simultaneously located in an indoor environment equipped with a positioning base station, or that the bound terminal falls within the coverage area of ​​the wireless communication signals of multiple bound anti-loss devices, and is located collaboratively through the wireless communication signals of multiple anti-loss devices.

[0096] Preferably, before the step of determining whether the second distance d2 between the anti-loss device and the binding terminal falls within a preset ranging range, the method further includes:

[0097] Configure the ranging radius r;

[0098] The ranging range is configured according to the ranging radius r, wherein the lower bound of the ranging range is d0-r and the upper bound of the ranging range is d0+r.

[0099] Furthermore, after determining whether the second distance d2 between the anti-loss device and the bound terminal falls within a preset ranging range, the method further includes:

[0100] When the second distance d2 between the anti-loss device and the bound terminal is outside the ranging range, return to the step of measuring the distance between the anti-loss device and the bound terminal by the wireless communication signal between the anti-loss device and the bound terminal to obtain the second distance d2 between the anti-loss device and the bound terminal.

[0101] Furthermore, after the step of associating and storing the ranging error with the user's behavior pattern, the method further includes:

[0102] Periodically iterate through the ranging error corresponding to each behavior pattern;

[0103] Remove outlier values ​​from the ranging error;

[0104] The average distance measurement error after removing outliers for each behavior pattern is determined as the false alarm bias threshold Δd0 for the corresponding behavior pattern.

[0105] Specifically, the abnormal values ​​in the ranging error are those that are significantly too large. Since ranging through wireless communication signals is easily affected by various environmental factors, there are measurement errors of varying magnitudes. Under some relatively extreme conditions, there may be some cases where the error is significantly too large. These cases are usually few and unreproducible. These abnormal values ​​can be removed from the database by configuring specific error thresholds.

[0106] In some other embodiments of the present invention, the maximum value of the ranging error after removing abnormal values ​​for each behavior pattern can be determined as the false alarm deviation threshold Δd0 for the corresponding behavior pattern.

[0107] Furthermore, after recording the behavior pattern as a false alarm-prone behavior pattern, the method further includes storing the action sequence and dynamic data corresponding to the false alarm-prone behavior pattern in a database. The specific steps for determining whether the user's current behavior pattern is a false alarm-prone behavior pattern include:

[0108] Obtain the action sequence corresponding to the user's current behavior pattern;

[0109] Match the database with action sequences that are prone to false alarms;

[0110] When a false alarm-prone behavior pattern with the same action sequence as the user's current behavior pattern exists in the database, the matching false alarm-prone behavior pattern will be identified as the target behavior pattern.

[0111] Obtain the first dynamic data corresponding to the user's current behavior pattern;

[0112] Read the second dynamic data of the target behavior pattern from the database;

[0113] Calculate the similarity between the first dynamic data and the second dynamic data;

[0114] When the similarity between the first dynamic data and the second dynamic data is greater than a preset similarity threshold, the user's current behavior pattern is determined to be a false alarm-prone behavior pattern.

[0115] Specifically, the database can be a local database built in the bound terminal, or a cloud database on a cloud server that establishes a communication connection with the bound terminal. In the step of storing the action sequence and dynamic data corresponding to the false alarm behavior pattern into the database, the dynamic data corresponding to the false alarm behavior pattern refers to the dynamic data recorded by the bound terminal within a second preset time range before determining the active alarm event of the false alarm behavior pattern. That is, the dynamic data corresponding to the false alarm behavior pattern is one or more data sequences recorded by the bound terminal within the second preset time range (depending on the specific implementation needs, different types of dynamic data can be recorded as a unified data sequence, or each type of dynamic data can be recorded as an independent data sequence).

[0116] In the step of matching false alarm behavior patterns in the database that have the same action sequence as the user's current behavior pattern, having the same action sequence means that the two behavior patterns are the same. However, the recording time of the corresponding dynamic data for the same behavior patterns at different times will be different, and the size of the dynamic data or the magnitude of the change in the dynamic data will inevitably have some differences.

[0117] Furthermore, the similarity between the first dynamic data and the second dynamic data can be represented using any one of the following: Euclidean distance, Manhattan distance, or cosine similarity between the two data sequences.

[0118] Furthermore, the step of identifying the first alarm-related event sequence within a first preset time range after the anti-loss device triggers an active alarm event specifically includes:

[0119] Acquire the first dynamic data within the first preset time range;

[0120] A first event sequence within the first preset time range is generated based on the first dynamic data and / or the changes in the first dynamic data. The first event sequence consists of one or more of the following: user operation events on the bound terminal and / or user behavior causing position change events, motion change events, or posture change events of the bound terminal.

[0121] Extract alarm-related events from the first event sequence to generate the first alarm-related event sequence.

[0122] In the technical solution of the above embodiments, the first dynamic data is a collection of recorded dynamic data of the bound terminal within the first preset time range. The dynamic data of the bound terminal or the changes in dynamic data are usually associated with some specific events, and the first event sequence is formed by arranging these specific events that occur within the first preset time range in chronological order. For example, the events in the first event sequence may include the event of the user picking up the phone corresponding to the posture change data of the bound terminal, the screen-on event corresponding to the bound terminal changing from a screen-off state to a screen-on state, the application opening event corresponding to the touch command input by the user received by the bound terminal through the touch screen, the alarm closing event corresponding to the dynamic data of the bound terminal sending an alarm closing command to the anti-loss device, the anti-loss device location viewing event where the bound terminal reads and displays the anti-loss device's location under the user's operation, and the event of the bound terminal moving closer to or further away from the anti-loss device corresponding to the change in the bound terminal's position, etc.

[0123] Of course, not all events in the first event sequence corresponding to the first dynamic data are necessarily related to the alarm event. It is necessary to analyze the correlation between the events in the first event sequence and the alarm event in order to extract the events associated with the alarm event from the first event sequence to generate the first alarm-related event sequence.

[0124] In the technical solution of this invention, the alarm-related events include events associated with specific applications, such as events related to operations on the anti-loss device's associated application or events related to control commands of the anti-loss device's associated application. The alarm-related events also include events of specified types, such as position change events, motion change events, and posture change events.

[0125] Furthermore, the step of determining whether the active alarm event is a false alarm event based on the first alarm-related event sequence specifically includes:

[0126] The first alarm-related event sequence is input into a pre-trained anti-loss device false alarm identification model to identify whether the active alarm event is a false alarm event. The anti-loss device false alarm identification model is a convolutional neural network model used to perform data classification.

[0127] In the technical solution of the above embodiments, by collecting a large amount of alarm-related event sequence data of the anti-loss device under false alarm or non-false alarm conditions, and labeling these data as false alarm or non-false alarm, a supervised training method can be used to train an anti-loss device false alarm recognition model with a convolutional neural network structure. Using the anti-loss device false alarm recognition model, the first alarm-related event sequence can be used as input data to identify whether the active alarm event is a false alarm event.

[0128] Furthermore, the step of identifying the second alarm-related event sequence within a second preset time range prior to the occurrence of an active alarm event by the anti-loss device specifically includes:

[0129] Acquire the second dynamic data within the second preset time range;

[0130] A second event sequence is generated within the second preset time range based on the second dynamic data and / or changes in the second dynamic data. The second event sequence consists of one or more of the following: position change events, motion change events, and posture change events of the bound terminal caused by the user's behavior.

[0131] Extract alarm-related events from the second event sequence to generate the second alarm-related event sequence.

[0132] In the technical solution of the above embodiments, the second dynamic data is a collection of recorded dynamic data of the bound terminal within the second preset time range. The second event sequence is composed of events related to dynamic data or changes in dynamic data occurring within the second preset time range, arranged chronologically. For example, the events in the second event sequence may include events such as the bound terminal's position change corresponding to its approaching or moving away from the anti-loss device, and events such as a significant decrease or increase in the signal strength of the anti-loss device corresponding to the bound terminal's position change or posture change.

[0133] Similarly, not all events in the second event sequence corresponding to the second dynamic data are necessarily related to the alarm event. It is necessary to analyze the correlation between the events in the second event sequence and the alarm event in order to extract the events associated with the alarm event from the second event sequence to generate the first alarm-related event sequence.

[0134] Furthermore, the step of analyzing the user's behavioral patterns within the second preset time range based on the second alarm-related event sequence specifically includes:

[0135] The second alarm-related event sequence is input into a pre-trained behavior pattern recognition model to identify the user's behavior pattern within a second preset time range. The false alarm recognition model of the anti-loss device is a convolutional neural network model used to perform data classification.

[0136] In the technical solution of the above embodiments, by collecting a large amount of alarm-related event sequence data of the anti-loss device under false alarm or non-false alarm conditions, and labeling these data as false alarm or non-false alarm, a supervised training method can be used to train an anti-loss device false alarm recognition model with a convolutional neural network structure. Using the anti-loss device false alarm recognition model, the first alarm-related event sequence can be used as input data to identify whether the active alarm event is a false alarm event.

[0137] A second aspect of the present invention proposes a false alarm prevention system based on user behavior self-learning of an anti-loss device, comprising an anti-loss device and a binding terminal, wherein the false alarm prevention system is configured to implement the false alarm prevention method based on user behavior self-learning of an anti-loss device as described in any of the first aspects of the present invention.

[0138] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0139] As described above, these embodiments of the present invention do not exhaustively cover all details, nor do they limit the invention to the specific embodiments described. Clearly, many modifications and variations can be made based on the above description. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to effectively utilize the invention and its modifications. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for preventing false alarms based on user behavior self-learning of anti-loss devices, characterized in that, include: The anti-loss device is configured to active alarm mode according to the user's operation instructions. In active alarm mode, an alarm is triggered when the distance between the anti-loss device and the bound terminal is greater than the pre-configured alarm distance. Monitor and record dynamic data with the bound terminal, wherein the dynamic data of the bound terminal is data reflecting the real-time status of the bound terminal and the real-time association status between the bound terminal and the anti-loss device; When the anti-loss device triggers an active alarm event, a first alarm-related event sequence within a first preset time range after the anti-loss device triggers an active alarm event is identified. The first alarm-related event sequence is a sequence of related events of dynamic data changes of the anti-loss device within a first preset time range after the active alarm event. Determine whether the active alarm event is a false alarm event based on the first alarm-related event sequence; When the active alarm event is determined to be a false alarm event, a second alarm-related event sequence within a second preset time range before the active alarm event of the anti-loss device is identified. The second alarm-related event sequence is a sequence of related events of dynamic data changes of the anti-loss device within a second preset time range before the active alarm event. Based on the second alarm-related event sequence analysis, the user's behavior pattern within the second preset time range is analyzed. The behavior pattern is a sequence of actions consisting of dynamic data or user behavior corresponding to changes in dynamic data within the second preset time range. Record the aforementioned behavior pattern as a false alarm-prone behavior pattern; After the step of recording the behavior pattern as a false alarm-prone behavior pattern, the method further includes: Monitor the first distance between the anti-loss device and the bound terminal. ; Determine the first distance between the anti-loss device and the bound terminal. Is it greater than the pre-configured alarm distance? ; When the first distance between the anti-loss device and the bound terminal Greater than the alarm distance At that time, calculate the first distance between the anti-loss device and the bound terminal. Distance from the alarm The difference between them: ； Determine the user's current behavior pattern and the false alarm deviation threshold between the anti-loss device and the current behavior pattern. ; Determine the first distance between the anti-loss device and the bound terminal. Distance from the alarm Is the difference between them less than or equal to the false alarm bias threshold? ; When the first distance between the anti-loss device and the bound terminal Distance from the alarm The difference between them is less than or equal to the false alarm bias threshold. At that time, determine whether the user's current behavior pattern is a false alarm-prone behavior pattern; When the user's current behavior pattern is not a false alarm-prone behavior pattern, the anti-loss device is controlled to issue an alarm.

2. The method for preventing false alarms based on user behavior self-learning of the anti-loss device according to claim 1, characterized in that, Also includes: When precise positioning is available, the anti-loss device and the bound terminal are precisely located, and the user's behavior patterns are monitored. The second distance between the anti-loss device and the bound terminal is obtained by measuring the distance through the wireless communication signal between the anti-loss device and the bound terminal. ; Determine the second distance between the anti-loss device and the bound terminal. Whether it falls within the preset ranging range; When the second distance between the anti-loss device and the binding terminal When the device falls within the preset ranging range, the ranging error between the anti-loss device and the binding terminal is calculated based on the accurate positioning result. The ranging error is associated with and stored in relation to the user's behavior patterns.

3. The method for preventing false alarms based on user behavior self-learning of the anti-loss device according to claim 2, characterized in that, After the step of associating and storing the ranging error with the user's behavior pattern, the method further includes: Periodically iterate through the ranging error corresponding to each behavior pattern; Remove outlier values ​​from the ranging error; The average ranging error after removing outliers for each behavior pattern is determined as the false alarm bias threshold for that behavior pattern. .

4. The method for preventing false alarms based on user behavior self-learning of the anti-loss device according to claim 1, characterized in that, After recording the behavior pattern as a false alarm-prone behavior pattern, the method further includes storing the action sequence and dynamic data corresponding to the false alarm-prone behavior pattern in a database. The specific steps for determining whether the user's current behavior pattern is a false alarm-prone behavior pattern include: Obtain the action sequence corresponding to the user's current behavior pattern; Match the database with action sequences that are prone to false alarms; When a false alarm-prone behavior pattern with the same action sequence as the user's current behavior pattern exists in the database, the matching false alarm-prone behavior pattern will be identified as the target behavior pattern. Obtain the first dynamic data corresponding to the user's current behavior pattern; Read the second dynamic data of the target behavior pattern from the database; Calculate the similarity between the first dynamic data and the second dynamic data; When the similarity between the first dynamic data and the second dynamic data is greater than a preset similarity threshold, the user's current behavior pattern is determined to be a false alarm-prone behavior pattern.

5. The method for preventing false alarms based on user behavior self-learning of the anti-loss device according to claim 1, characterized in that, The steps for identifying the first alarm-related event sequence within a first preset time range after the anti-loss device triggers an active alarm event specifically include: Acquire the first dynamic data within the first preset time range; Based on the changes in the first dynamic data, a first event sequence is generated within the first preset time range. The first event sequence consists of one or more of the following: user operation events on the bound terminal and / or user behavior causing position change events, motion change events, or posture change events of the bound terminal. Extract alarm-related events from the first event sequence to generate the first alarm-related event sequence.

6. The method for preventing false alarms based on user behavior self-learning of the anti-loss device according to claim 5, characterized in that, The steps for determining whether the active alarm event is a false alarm event based on the first alarm-related event sequence specifically include: The first alarm-related event sequence is input into a pre-trained anti-loss device false alarm identification model to identify whether the active alarm event is a false alarm event. The anti-loss device false alarm identification model is a convolutional neural network model used to perform data classification.

7. The method for preventing false alarms based on user behavior self-learning of the anti-loss device according to claim 5, characterized in that, The steps for identifying the sequence of second alarm-related events within a second preset time range prior to the occurrence of an active alarm event by the anti-loss device specifically include: Acquire the second dynamic data within the second preset time range; Based on the changes in the second dynamic data, a second event sequence is generated within the second preset time range. The second event sequence consists of one or more of the following: position change events, motion change events, and posture change events of the bound terminal caused by the user's behavior. Extract alarm-related events from the second event sequence to generate the second alarm-related event sequence.

8. The method for preventing false alarms based on user behavior self-learning of the anti-loss device according to claim 7, characterized in that, The steps for analyzing user behavior patterns within the second preset time range based on the second alarm-related event sequence specifically include: The second alarm-related event sequence is input into a pre-trained behavior pattern recognition model to identify the user's behavior patterns within a second preset time range. The behavior pattern recognition model is a convolutional neural network model used to perform data classification.

9. A false alarm prevention system based on user behavior self-learning of an anti-loss device, characterized in that, The system includes an anti-loss device and a binding terminal, and the false alarm prevention system is configured to implement the false alarm prevention method based on user behavior self-learning of the anti-loss device as described in any one of claims 1-8.

Images