Fall detection method based on MQTT-ROS communication bridge and multi-dimensional statistical features

Abstract

This application discloses a fall detection method based on MQTT-ROS communication bridging and multidimensional statistical features, comprising: acquiring data packets; decoding the data packets; if a human body is in an active state, sending continuous motion sensing states and active states to two sliding window queues of preset lengths respectively; determining whether the active state meets a first preset condition and / or a second preset condition, and simultaneously determining whether the motion sensing state meets a third preset condition and / or a fourth preset condition; if a fall event is determined, modifying the fall state field in the data packet to a valid state and adding a ROS system timestamp; based on the modified data packet, publishing alarm information to the ROS system through a ROS topic, triggering a single-trigger mechanism to shut down the ROS system node, thereby completing the abnormal event response process. This application effectively avoids false fall detections caused by historical data contamination and significantly improves detection accuracy in scenarios with frequent personnel entry and exit.

Description

Technical Field

[0001] This application relates to the field of human body detection technology, and in particular to a fall detection method based on MQTT-ROS communication bridging and multidimensional statistical features. Background Technology

[0002] Many existing fall detection technologies rely solely on simple single-frame data for judgment, failing to capture dynamic features over time and thus easily leading to false positives. Furthermore, some systems lack effective data stream gating and reset mechanisms. When the monitored target reappears after leaving, the algorithm may perform calculations based on "stale data" remaining in the buffer, resulting in inaccurate detection results. Additionally, traditional methods often lack filtering of sensor noise, leading to false alarms in highly active, non-fall-related states. Summary of the Invention

[0003] The purpose of this application is to provide a fall detection method based on MQTT-ROS communication bridging and multidimensional statistical features, in order to solve the technical problem of high false alarm rate in existing fall detection technologies. The various technical effects of the preferred technical solutions provided in this application are detailed below.

[0004] To achieve the above objectives, this application provides the following technical solutions: The first aspect of this application provides a fall detection method based on MQTT-ROS communication bridging and multidimensional statistical features, comprising: acquiring data packets in real time through an MQTT client; decoding the data packets to obtain a human presence sensing state, a motion sensing state, and an active state; if the human presence sensing state is active, sending consecutive motion sensing states and active states to two sliding window queues of preset length, wherein the sliding window queues are based on the ROS system; determining whether the active state meets a first preset condition and / or a second preset condition, and simultaneously determining whether the motion sensing state meets a third preset condition and / or a fourth preset condition; if both the active state and the motion sensing state meet their respective preset conditions, a fall event is determined to have occurred; if a fall event is determined, modifying the fall state field in the data packet to a valid state and adding a ROS system timestamp; publishing an alarm message to the ROS system through a ROS topic based on the modified data packet; triggering a single-trigger mechanism to shut down the ROS system node to complete the abnormal event response process.

[0005] In some embodiments, the first preset condition is that the average value of the active state is greater than a first preset value, and the second preset condition is that the difference between the maximum and minimum values ​​of the active state is greater than a second preset value.

[0006] In some embodiments, the third preset condition is that the average value of the motion sensing state is greater than a third preset value, and the fourth preset condition is that the variance of the motion sensing state is less than a fourth preset value.

[0007] In some embodiments, before publishing alarm information to the ROS system via ROS topic based on the modified data packet, the fall detection method based on MQTT-ROS communication bridging and multidimensional statistical features further includes: locking the critical section of the sliding window queue by thread lock.

[0008] In some embodiments, publishing alarm information to the ROS system via a ROS topic based on the modified data packet includes: calling the ROS Publisher interface to publish the modified data packet to a preset topic.

[0009] In some embodiments, the first preset value is 60 to 80, and the second preset value is 40 to 60.

[0010] In some embodiments, the third preset value is 90 to 110, and the fourth preset value is 20 to 40.

[0011] In some embodiments, when the human body is in a motion sensing state but not in the active state, the fall detection method based on MQTT-ROS communication bridging and multidimensional statistical features further includes: clearing the buffers of the motion sensing state and the active state.

[0012] A second aspect of this application provides a computer-readable storage medium storing a computer program that, when executed, implements the fall detection method based on MQTT-ROS communication bridging and multidimensional statistical features as described above.

[0013] A third aspect of this application provides a processing apparatus, comprising: one or more processors; a memory for storing one or more computer programs, wherein the one or more processors are configured to execute the one or more computer programs stored in the memory to cause the one or more processors to perform the fall detection method based on MQTT-ROS communication bridging and multidimensional statistical features as described above.

[0014] Implementing one of the above-mentioned technical solutions of this application has the following advantages or beneficial effects: In this application, it is first determined whether the human body is in an active state when sensing its presence. If it is in an active state, the continuous motion sensing status and active status are sent to the sliding window queue. Then, it is determined whether the motion sensing status and active status meet the corresponding preset conditions. If they are met, an alarm message is issued and the ROS system node is shut down.

[0015] In this case, by sending the data into a sliding window queue, continuous time slices before and after the fall can be recorded completely, thereby capturing the unique time series characteristics of the fall process, enabling the system to accurately distinguish between rapid falls and normal actions such as standing still or sitting down slowly.

[0016] Furthermore, the single-trigger mechanism immediately locks and stops routine monitoring once a fall event is triggered and an alarm is issued. This not only prevents redundant and repeated alarm messages from the same fall event, reducing the communication load on the upper-level monitoring system, but also ensures the uniqueness and urgency of the alarm signal, facilitating rapid intervention in subsequent rescue procedures. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings: Figure 1 This is a flowchart illustrating the fall detection method based on MQTT-ROS communication bridging and multidimensional statistical features according to an embodiment of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, various exemplary embodiments described below will be referenced to the accompanying drawings, which form part of the exemplary embodiments and depict various exemplary embodiments that may be adopted to implement this application. Unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. It should be understood that they are merely examples of processes, methods, and apparatuses consistent with some aspects of this application disclosed as detailed in the appended claims, and other embodiments may be used, or structural and functional modifications may be made to the embodiments listed herein without departing from the scope and spirit of this application.

[0019] In the description of this application, it should be understood that the terms "center," "longitudinal," "lateral," etc., indicate the orientation or positional relationship based on the accompanying drawings, and are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the referred element must have a specific orientation, or be constructed and operated in a specific orientation. 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. The term "multiple" means two or more. The terms "connected" and "linked" should be interpreted broadly, for example, they can be fixed connections, detachable connections, integral connections, mechanical connections, electrical connections, communication connections, direct connections, indirect connections through an intermediate medium, and can be the internal connection of two elements or the interaction relationship between two elements. The term "and / or" includes any and all combinations of one or more of the related listed items. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0020] To illustrate the technical solutions described in this application, specific embodiments are provided below, showing only the parts related to the embodiments of this application.

[0021] like Figure 1 As shown, this application provides a fall detection method based on MQTT-ROS communication bridging and multidimensional statistical features, including the following steps (steps S1 to S3): S1. Data packets are acquired in real time through the MQTT (Internet of Things Protocol) client. The data packets are decoded to obtain the human body presence sensing state, motion sensing state, and active state. If the human body presence sensing state is active, the motion sensing state and active state are sent to two sliding window queues of preset length, respectively. The sliding window queues are based on the ROS (Robot Operating System) system.

[0022] Specifically, the data packets can be JSON-formatted data streams published by millimeter-wave radar sensors and subscribed to through an MQTT client, i.e., sensor data. Meanwhile, by utilizing the ROS system mechanism in conjunction with the asynchronous loop thread of MQTT in the main control loop, parallel processing of external sensor data reception and internal algorithm scheduling can be achieved.

[0023] A sliding window queue can refer to a buffer queue, specifically a first-in, first-out (FIFO) buffer queue. This application utilizes the event loop mechanism of the ROS system to ensure that nodes can continuously respond to system commands during runtime and maintain the algorithm's lifecycle.

[0024] In some embodiments, the MQTT client can be used to decode and perform gating checks on the presence of a human body by bridging the ROS system, that is, to determine that the presence of a human body is in an active state before entering the sliding window algorithm logic based on the ROS framework.

[0025] In other embodiments, TCP / IP Socket direct connection, HTTP RESTful API (Representational State Transfer API) polling, or CoAP (Constrained Application Protocol) can also be used to transmit data packets. RESTful API is a web service interface based on the HTTP protocol and following the REST architecture constraints, while CoAP is an application layer protocol designed specifically for resource-constrained IoT devices.

[0026] The fall detection method based on MQTT-ROS communication bridging and multidimensional statistical features proposed in this application can be encapsulated as an independent edge computing node, which can be decoupled from other robot functions of the ROS system, such as navigation, voice, and emergency obstacle avoidance.

[0027] In some embodiments, the data packet may include a human state sensing field (HumanSenseState) and multidimensional motion statistics. The human state sensing field is the field corresponding to the human body's sensing state, and it mainly serves as a "hierarchical gating switch" for the system.

[0028] In some embodiments, when the human body is in a motion sensing state but not in an active state, the fall detection method based on MQTT-ROS communication bridging and multidimensional statistical features may further include: clearing the buffers of motion sensing state and active state.

[0029] Specifically, when the human body is in a motion sensing state but not in an active state, the value of the human body sensing state can be less than a preset threshold, such as less than 1. In this case, it can be determined that there is no one in the current area. By calling the clear method, the motion sensing state and active state buffer are forcibly cleared, that is, the buffer is reset. This ensures that each calculation is based on the latest real-time data, effectively avoiding false fall judgments caused by historical data contamination, and significantly improving the detection accuracy in scenarios where people frequently enter and exit.

[0030] Correspondingly, when the human presence sensor is active, and the value of the human presence sensor is greater than or equal to a preset threshold (e.g., greater than or equal to 1), it can be determined that someone is present in the current area. Real-time data packets can then be pushed into a sliding window queue of a preset length for calculation, for example, the preset length can be 30. In other words, it can accept consecutive data packets with the same value as the preset length, such as 30 packets, until a fall event is determined or the value of the human presence sensor is less than the preset threshold.

[0031] When the amount of data exceeds the preset length of the sliding window queue, for example, when the amount of data exceeds 30, the old data in the sliding window queue can be overflowed, thereby ensuring that the data is a continuous time slice of the most recent 30 time steps.

[0032] Multidimensional motion statistics refer to the core data extracted and pushed into a buffer from the raw data stream of an MQTT client subscription, including Motion Sense State and Active State. Motion Sense State reflects the target's motion amplitude, while Active State measures the target's energy level.

[0033] In the write and read operations of a sliding window queue, thread locks can be used to protect data consistency and ensure that statistical calculations are not interrupted by newly arriving data packets.

[0034] In some embodiments, a sliding window queue can be implemented using Python's collections.deque queue, a ring buffer, or a fixed-length array with pointer movement.

[0035] S2. Determine whether the active state meets the first preset condition and / or the second preset condition, and at the same time determine whether the motion sensing state meets the third preset condition and / or the fourth preset condition. If both the active state and the motion sensing state meet the corresponding preset conditions, then it is determined that a fall event has occurred.

[0036] Specifically, statistical analysis can be performed on the data within the sliding window sequence to determine whether the values ​​of the active state and motion sensing state meet preset conditions. Furthermore, the active state can meet either the first or the second preset condition, or both the first and second preset conditions simultaneously; correspondingly, the motion sensing state can meet either the third or the fourth preset condition, or both the third and fourth preset conditions simultaneously.

[0037] Preferably, when the active state satisfies the first and second preset conditions, and the motion sensing state satisfies the third and fourth preset conditions, it is considered that both the active state and the motion sensing state satisfy the corresponding preset conditions.

[0038] In some embodiments, the first preset condition can be that the average value of the active state is greater than a first preset value, and the second preset condition can be that the difference between the maximum and minimum values ​​of the active state is greater than a second preset value. Specifically, the first preset condition can be used to confirm that the target is at a high level of activity, and the first preset value can be 60 to 80, for example, 60, 70, 80, etc.; the second preset condition can be used to capture the drastic change in the signal at the moment of falling, and the second preset value can be 40 to 60, for example, 40, 50, 60, etc.

[0039] In some embodiments, the third preset condition can be that the average value of the motion sensing state is greater than a third preset value, and the fourth preset condition can be that the variance of the motion sensing state is less than a fourth preset value. Specifically, the third preset condition can be used to confirm that the overall motion amplitude is large, and the third preset value can be 90 to 110, for example, 90, 100, 110, etc.; the fourth preset condition can be used to eliminate high-frequency noise and ensure that the continuous high-intensity signal is not environmental noise, and the fourth preset value can be 20 to 40, for example, 20, 30, 40, etc. By limiting the variance of the motion sensing state, transient and unstable sensor noise interference can be effectively filtered out, and only those high-value and somewhat persistent, i.e., high-saturation real fall signals can be responded to, thereby significantly reducing the false alarm rate while ensuring high sensitivity.

[0040] In other embodiments, the fourth preset condition may be that the standard deviation of the motion sensing state is less than the fifth preset value, or the root mean square error of the motion sensing state is less than the sixth preset value.

[0041] It should be noted that the first, second, third, fourth, fifth, and sixth preset values ​​involved in the embodiments of this application, and their corresponding ActiveState and MotionSenseState values, are all dimensionless relative quantization indices output by the underlying signal processing of the millimeter-wave radar sensor. These indices are typically mapped to specific numerical ranges (such as 0-255 or 0-100) to characterize relative energy levels or motion amplitudes, rather than measurements with absolute physical units (such as velocity, acceleration, etc.).

[0042] Specifically, the average value of the active state can range from 0 to 100, the difference between the maximum and minimum values ​​of the active state can range from 0 to 100, the average value of the motion sensing state can range from 0 to 255, and the variance of the motion sensing state can range from 0 to 100.

[0043] In some embodiments, support vector machines (SVM), decision trees, or lightweight neural networks can be used to classify and train the data in the sliding window queue to determine whether a fall event has occurred.

[0044] Python's asynchronous I / O or multithreading technology can be used to enable the asynchronous loop thread of the MQTT client to be responsible for receiving JSON data streams in real time, while the ROS system mechanism is responsible for internal algorithm scheduling and topic publishing.

[0045] S3. If the event is determined to be a fall, the fall state field (FallState) in the data packet is modified to a valid state and a ROS system timestamp is added. Based on the modified data packet, an alarm message is published to the ROS system through the ROS topic, triggering the single-trigger mechanism to shut down the ROS system node, thus completing the abnormal event response process.

[0046] By modifying the FallState field to an active state, such as setting the FallState field to 1, the modified data packet contains both the original sensor values ​​that caused the fall determination and the alarm conclusion confirmed by the algorithm.

[0047] By forcibly shutting down the ROS system node, a closed-loop response from receiving data packets to safety circuit breaking can be completed. Specifically, the system shutdown function can be called to shut down the current ROS system node, preventing repeated alarms caused by residual limb movements after a fall, and locking the current abnormal state until an external reset is performed. Adding a ROS system timestamp ensures the traceability of alarm information.

[0048] In some embodiments, before publishing alarm information to the ROS system via ROS topic based on modified data packets, the fall detection method based on MQTT-ROS communication bridging and multidimensional statistical characteristics may further include locking the critical section of the sliding window queue using a thread lock. This ensures that no new data packets enter the buffer and interfere with the process of publishing alarm information and shutting down the ROS system node.

[0049] In some embodiments, publishing alarm information to the ROS system via a ROS topic based on the modified data packet may include: calling the ROS Publisher interface to publish the modified data packet to a preset topic, such as / mqtt_data (file system path), so that other modules of the ROS system, such as the alarm bell module or the emergency stop module, can immediately receive the alarm signal.

[0050] A one-shot trigger mechanism means that the trigger logic is activated immediately after the alarm message is published. This aims to ensure that a single fall event generates only one unique alarm message, avoiding redundant or irrelevant information. Finally, functions such as `rospy.signal_shutdown` can be called to terminate the current ROS system node, completely stopping all MQTT client subscriptions and data processing tasks, thus completing the exception response loop.

[0051] The application scenarios of this application embodiment are as follows: Elderly care service robots: As a mobile monitoring terminal, the robot patrols nursing homes or home environments, monitors the elderly’s condition non-contactly through millimeter-wave radar, and links with the robot’s emergency call or obstacle avoidance functions through the ROS system.

[0052] Complex personnel flow areas: This application includes a gating mechanism for sensing human presence and a buffer reset mechanism, which is particularly suitable for dynamic environments with frequent personnel entry and exit, such as hospitals and corridors, and can effectively avoid false alarms caused by historical residual data.

[0053] Edge computing and privacy protection requirements: The system is deployed directly at the edge of the robot for data processing, without uploading raw physiological / motor data to the cloud. Furthermore, millimeter-wave radar does not involve image privacy, making it suitable for privacy-sensitive areas such as bathrooms and bedrooms.

[0054] In a weak network IoT environment: By leveraging the high robustness of the MQTT client in unstable networks and combining it with the local real-time decision-making capabilities of the ROS system, we can ensure that fall detection and triggering of local alarms can still be completed when the network fluctuates.

[0055] In this application, it is first determined whether the human body is in an active state when sensing motion. If it is active, the continuous motion sensing status and active status are sent to a sliding window queue. Then, it is determined whether the motion sensing status and active status meet the corresponding preset conditions. If they do, an alarm message is issued and the ROS system node is shut down.

[0056] In this case, by sending the data into a sliding window queue, continuous time slices before and after the fall can be recorded completely, thereby capturing the unique time series characteristics of the fall process, enabling the system to accurately distinguish between rapid falls and normal actions such as standing still or sitting down slowly.

[0057] Furthermore, the single-trigger mechanism immediately locks and stops routine monitoring once a fall event is triggered and an alarm is issued. This not only prevents redundant and repeated alarm messages from the same fall event, reducing the communication load on the upper-level monitoring system, but also ensures the uniqueness and urgency of the alarm signal, facilitating rapid intervention in subsequent rescue procedures.

[0058] Those skilled in the art will understand that all or part of the features / steps of the above-described method embodiments can be implemented by methods, data processing systems, or computer programs. These features may be implemented without hardware, entirely in software, or in a combination of hardware and software. The aforementioned computer program may be stored in one or more computer-readable storage media. When the computer program is executed (e.g., by a processor), it performs the steps of the fall detection method embodiments based on MQTT-ROS communication bridging and multidimensional statistical features described above.

[0059] The aforementioned storage media capable of storing program code include: static hard disks, solid-state hard disks, random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), optical storage devices, magnetic storage devices, flash memory, magnetic disks or optical disks, and / or combinations of the above devices, that is, they can be implemented by any type of volatile or non-volatile storage devices or combinations thereof.

[0060] This application also provides a processing device embodiment, including one or more processors and a memory; wherein the memory is used to store one or more computer programs, and the one or more processors are used to execute the one or more computer programs stored in the memory, so that the processors execute the features / steps of the fall detection method embodiment based on MQTT-ROS communication bridging and multidimensional statistical features described above.

[0061] The above description is merely a preferred embodiment of this application. Those skilled in the art will understand that various changes or equivalent substitutions can be made to these features and embodiments without departing from the spirit and scope of this application. Furthermore, under the teachings of this application, these features and embodiments can be modified to adapt to specific situations and materials without departing from the spirit and scope of this application. Therefore, this application is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are within the protection scope of this application.

Claims

1. A fall detection method based on MQTT-ROS communication bridging and multi-dimensional statistical features, characterized in that, include: Data packets are acquired in real time through an MQTT client. The data packets are decoded to obtain the human body presence sensing state, motion sensing state, and active state. If the human body presence sensing state is active, the continuous motion sensing state and the active state are sent to two sliding window queues of preset length, respectively. The sliding window queues are based on the ROS system. Determine whether the active state meets the first preset condition and / or the second preset condition, and at the same time determine whether the motion sensing state meets the third preset condition and / or the fourth preset condition. If both the active state and the motion sensing state meet the corresponding preset conditions, it is determined that a fall event has occurred. If the event is determined to be a fall, the fall status field in the data packet is modified to a valid state and a ROS system timestamp is added. Based on the modified data packet, an alarm message is published to the ROS system via a ROS topic, triggering a one-time trigger mechanism to shut down the ROS system node, thus completing the abnormal event response process.

2. The fall detection method based on MQTT-ROS communication bridging and multi-dimensional statistical features according to claim 1, characterized in that, The first preset condition is that the average value of the active state is greater than a first preset value, and the second preset condition is that the difference between the maximum and minimum values ​​of the active state is greater than a second preset value.

3. The fall detection method based on MQTT-ROS communication bridging and multi-dimensional statistical features according to claim 1, characterized in that, The third preset condition is that the average value of the motion sensing state is greater than the third preset value, and the fourth preset condition is that the variance of the motion sensing state is less than the fourth preset value.

4. The fall detection method based on MQTT-ROS communication bridging and multi-dimensional statistical features according to claim 1, characterized in that, Before publishing alarm information to the ROS system via ROS topic based on the modified data packet, the fall detection method based on MQTT-ROS communication bridging and multidimensional statistical features further includes: locking the critical section of the sliding window queue by thread lock.

5. The fall detection method based on MQTT-ROS communication bridging and multidimensional statistical features according to claim 1, characterized in that, The step of publishing alarm information to the ROS system through a ROS topic based on the modified data packet includes: calling the ROS Publisher interface to publish the modified data packet to a preset topic.

6. The fall detection method based on MQTT-ROS communication bridging and multi-dimensional statistical features according to claim 2, characterized in that, The first preset value is 60 to 80, and the second preset value is 40 to 60.

7. The fall detection method based on MQTT-ROS communication bridging and multi-dimensional statistical features according to claim 3, characterized in that, The third preset value is 90 to 110, and the fourth preset value is 20 to 40.

8. The fall detection method based on MQTT-ROS communication bridging and multi-dimensional statistical features according to claim 1, characterized in that, When the human body is in a motion sensing state but not in the active state, the fall detection method based on MQTT-ROS communication bridging and multidimensional statistical features further includes: clearing the buffers of the motion sensing state and the active state.

9. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed, implements the fall detection method based on MQTT-ROS communication bridging and multidimensional statistical features as described in any one of claims 1-8.

10. A processing apparatus, characterized in that, include: One or more processors; A memory for storing one or more computer programs, and one or more processors for executing the one or more computer programs stored in the memory to cause the one or more processors to perform the fall detection method based on MQTT-ROS communication bridging and multidimensional statistical features as described in any one of claims 1-8.

Images