A wearable inertial sensor-based ankle pump exercise adherence monitoring method and system

Abstract

This invention discloses a method and system for monitoring ankle pump exercise compliance based on a wearable inertial sensor. By fixing an inertial measurement unit to a preset anatomical position below the patient's lateral malleolus, ankle joint motion data is collected. The data is filtered and angles are calculated to extract motion amplitude, motion cycle, flexion-extension symmetry, and active features. Valid ankle pump movements are identified according to preset judgment rules. Continuous valid movements are then aggregated into valid exercise groups, compared with preset medical advice targets, and compliance results are generated. Warning information is output when the exercise is not completed within the time limit or when compliance is insufficient. This invention improves the distinguishability of ankle pump motion signals from general whole-body motion signals by fixing the inertial measurement unit to a preset anatomical position on the ankle for localized acquisition of ankle pump nursing movements. This invention can simultaneously exclude walking, leg shaking, and passive activities, improving the nursing relevance of the movement recognition results.

Description

Technical Field

[0001] This invention relates to a method and system for monitoring ankle pump movement compliance based on wearable inertial sensors, belonging to the technical fields of medical wearable monitoring, patient rehabilitation compliance management, and hospital information system integration. Background Technology

[0002] Venous thromboembolism is a significant complication in hospitalized and postoperative patients, with deep vein thrombosis (DVT) frequently occurring in the lower extremities, and pulmonary embolism being a serious consequence. Ankle pump exercises, a common non-pharmacological preventative measure for perioperative and bedridden patients, promote venous return in the lower extremities through ankle dorsiflexion and plantarflexion, and have been widely adopted in clinical nursing practice. However, in actual ward and home-based continuing care settings, ankle pump exercises still primarily rely on verbal instruction, patient self-management, and nurse spot checks, leading to issues such as inconsistent frequency of administration, difficulty in assessing the extent of administration, and the inability to promptly detect instances of non-administration exceeding the recommended timeframe.

[0003] Existing ankle pump exercise compliance monitoring solutions can be broadly categorized into three types: The first type consists of mechanical or semi-mechanical ankle pump training devices, which typically assist patients in completing foot flexion and extension movements through supports, pedals, elastic components, or drive components. The focus is on assisted or passive training rather than continuous identification and quantification of patient's active ankle pump exercise performance. The second type comprises general ankle joint motion monitoring or rehabilitation monitoring devices, which can use angle sensors, pressure sensors, inertial sensors, or external terminals to record ankle joint activity. However, these devices are primarily geared towards general joint range of motion assessments, rehabilitation training records, or gait analysis, without establishing specific judgment rules for ankle pump exercise as a specific nursing action. The third type is a general rehabilitation compliance platform, which combines wearable sensors, mobile terminals, and healthcare management interfaces to count training sessions, send reminders, or synchronize patient data. However, it lacks the ability to differentiate between ankle pump exercise and walking, leg shaking, or passive activity.

[0004] For example, Chinese invention patent CN112516533B discloses a non-contact ankle pump movement achievement reminder method, system, and storage medium. It acquires target setting data and actual training data for ankle pump movements, compares the two to determine if the training has met the target, and then provides a reminder. However, this solution is explicitly limited to non-contact monitoring and does not disclose a technical solution for contact-based local signal acquisition by fixing an inertial measurement unit to a local area near the lateral malleolus in a ward setting. This presents the following problems: its "actual training data" is only a general description of the movement's effectiveness; it does not disclose effective ankle pump movement judgment rules based on the combination of four conditions: movement amplitude, movement cycle, flexion-extension symmetry, and initiative; nor does it disclose a specific mechanism for excluding passive activities, walking movements, and leg shaking movements through the initiative condition. Furthermore, its "reminder" is a feedback mechanism for achieving the target after training, using the detection of the movement and assessment of whether the target has been met as the trigger condition. It does not disclose a pre-configured execution schedule by nurses, and a pre-programmed mobile application on the patient's end actively outputs timed exercise reminders to the patient at the target execution time indicated by the execution schedule to drive the patient to begin exercise.

[0005] For example, Chinese invention patent CN116211290B discloses a method and system for monitoring and evaluating the posture of ankle pump movements. This method involves wearing a posture monitoring module on the patient's foot, constructing a carrier coordinate system with the module's center point, fusing angular velocity, acceleration, and magnetic force information through fuzzy adaptive Kalman filtering, obtaining state variables using the angular relationship between the carrier coordinate system and the ankle joint coordinate system, and performing cubic polynomial fitting and average error calculation on the state variables and standard state variables in a personal database to assess the effectiveness of the current ankle pump movement. However, this method wears the module on the sole of the foot rather than in the adjacent area below the lateral malleolus, does not disclose anatomical optimization for the signal-to-noise ratio of ankle pump flexion-extension signals, and its evaluation path-dependent state variable uses an overall template matching method. It does not consider flexion-extension symmetry and initiative as independent conditions for conjunction judgment, nor does it disclose the use of initiative conditions to distinguish active ankle pumps from passive activities, walking movements, or leg shaking movements. The patented solution does not disclose a compliance closed-loop mechanism that aggregates effective actions that have been performed a preset number of times into an effective exercise group and compares it with the nursing care order goals, nor does it involve nurses setting up execution plans and patients receiving timely reminders.

[0006] For example, Chinese invention patent application CN118614909A discloses an ankle pump movement monitoring and assessment system. This system uses a three-layer architecture of a data acquisition unit, a processing unit, and a motion unit to generate corresponding ankle pump movement plans for each of the patient's feet. The plans include a first and second expected goals. A first auxiliary component in the motion unit guides the patient's foot movement, and the system uses the local pressure distribution of the first auxiliary component to determine whether the patient has achieved the goal and their subjective initiative. However, this system relies on hardware auxiliary components to guide the movement and makes judgments based on pressure distribution rather than inertial signals. It does not disclose a method using a wearable inertial measurement unit for local data acquisition of the lateral ankle. This technical solution, based on the active / passive judgment method of "local pressure of auxiliary components" + "subjective initiative," is completely different from the technical path of active identification based on inertial signal characteristics such as peak angular velocity, amplitude of gravity component changes, and waveform regularity. Meanwhile, the technical solution does not disclose a closed-loop mechanism that aggregates effective movements into exercise groups and links them with multi-dimensional nursing orders that include execution intervals, minimum number of movements per group, duration of each group, and daily target number of groups to output timeout warnings. It also does not disclose a timed reminder mechanism that actively drives patients to start exercising at preset execution times.

[0007] Another general wearable compliance system for physical rehabilitation has been disclosed, with an overall architecture of "sensor + application + clinical portal". However, it is designed for general rehabilitation movements and cannot be directly adapted to nursing scenarios such as ankle pump exercises, which have special requirements for local amplitude, flexion and extension rhythm, initiative and doctor's order execution interval.

[0008] In summary, the existing solutions have at least the following structural shortcomings:

[0009] 1) It mainly targets general motion monitoring or non-contact monitoring, and cannot simultaneously achieve the acquisition of high signal-to-noise ratio inertial signals below the lateral malleolus and stably distinguish ankle pump movement from walking, leg shaking or other assisted activities; 2) There is a lack of an effective ankle pump action assessment mechanism that combines four conditions: range of motion, exercise cycle, flexion-extension symmetry, and initiative. It is impossible to determine whether the patient has truly achieved the required quality of movement and to exclude non-target movements. 3) Even if a single action can be recorded or the achievement of the target can be assessed, continuous effective actions are not aggregated into exercise groups and linked with the execution interval, number of times per group, duration of each group and daily target number of groups in the nursing orders in multiple dimensions, so as to form a compliance early warning closed loop that can be used for ward management. 4) Existing "reminder" or "early warning" mechanisms are generally post-training feedback after achieving the target or after exceeding the time limit. They do not realize a pre-driving mechanism in which nurses preset the execution schedule and the patient's mobile application actively drives the patient to start exercising at the preset target execution time.

[0010] Therefore, a dedicated monitoring technology solution is needed for ankle pump exercise care scenarios that can simultaneously address the above four shortcomings. Summary of the Invention

[0011] This invention aims to solve the problems of existing ankle pump exercise compliance monitoring programs being unable to objectively identify effective ankle pump movements in ward and home care scenarios, unable to quantify compliance, and unable to implement early warnings based on doctor's orders.

[0012] To address the aforementioned technical problems, the first aspect of the present invention discloses a method for monitoring ankle pump movement compliance based on a wearable inertial sensor, characterized by comprising the following steps: Step 1: Fix the wearable sensor containing the inertial measurement unit to the adjacent local area below the lateral malleolus of the patient, so that the inertial measurement unit is kept within a preset distance range relative to the ankle joint flexion and extension axis. Step 2: Collect ankle joint movement data during dorsiflexion and plantarflexion of the patient's ankle joint using wearable sensors; Step 3: After preprocessing the ankle joint motion data, extract the motion amplitude, motion cycle, flexion-extension symmetry, and active features, including: By using initiative-related characteristics, we can distinguish ankle pump movements that patients perform on their own from non-targeted movements such as foot movements assisted by others, getting out of bed and walking, and random leg shaking. Flexion-extension symmetry is determined based on the proportional relationship between the peak dorsiflexion and the peak plantarflexion within one motion cycle; Step 4: Determine whether the feature data simultaneously meets the preset rules for determining effective ankle pumping motion. If it does, the current ankle flexion and extension motion is determined to be an effective ankle pumping motion, and proceed to Step 5. If it does not meet the rules, the current ankle flexion and extension motion is determined to be an invalid ankle pumping motion and excluded, and return to Step 2. Step 5: Determine whether the number of consecutive effective ankle pump movements has reached the preset number. If not, return to step 2. If it has, record the consecutive effective ankle pump movements that have reached the preset number as a valid exercise group, and determine the end of the valid exercise group based on the duration of consecutive ineffective ankle pump movements reaching the preset duration. Step 6: Compare the effective exercise group with the preset medical advice goals to obtain compliance results; Step 7: For the current valid exercise set, determine: Whether the current time exceeds the execution interval and whether the target is met within the preset allowed time window; Or whether the compliance result is lower than a preset threshold; Or whether the difference in completion rates between the two sides exceeds a preset threshold; If yes, output a warning message; otherwise, return to step 2.

[0013] Preferably, in step 1, the adjacent local area is a local area located below the lateral malleolus and 0-5 cm from the center of the lateral malleolus, so as to obtain ankle joint motion data with a high signal-to-noise ratio during ankle dorsiflexion and plantarflexion by the inertial measurement unit.

[0014] Preferably, in step 3, the active related features include at least one of peak angular velocity, amplitude of change of gravity component, and waveform regularity.

[0015] Preferably, in step 3, the preset effective ankle pump action determination rule is: the current ankle joint flexion and extension movement is judged as an effective ankle pump action only when the feature data simultaneously meets the preset range of motion amplitude, preset range of motion cycle, preset range of flexion and extension symmetry, and preset active conditions; otherwise, it is judged as an invalid ankle pump action. Passive activities, walking movements, and leg shaking movements are identified as invalid ankle pump actions and excluded through the preset active conditions.

[0016] Preferably, in the preset effective ankle pump action determination rule: Regarding the aforementioned active-related characteristics, if at least one of the peak angular velocity, the amplitude of the change in the gravitational component, or the waveform regularity does not meet the preset active conditions, the ankle flexion and extension movement will not be recorded as an effective ankle pump action.

[0017] Regarding the flexion-extension symmetry, when the proportional relationship is within a preset range, it is determined that the preset flexion-extension symmetry range is met. The preset range can be individually calibrated based on the patient's baseline data.

[0018] Preferably, the parameters of the preset effective ankle pump action judgment rule are pre-calibrated based on the nursing plan, standard action samples, non-target action samples, and patient baseline data. During calibration, samples such as standard ankle pump actions, walking, leg shaking, and foot swinging assisted by others are collected to extract the amplitude of movement, movement cycle, peak dorsiflexion, peak plantarflexion, peak angular velocity, changes in gravity components, and waveform regularity. The parameter range or judgment boundary that can simultaneously distinguish between effective ankle pump actions and non-target actions is selected as the default parameter.

[0019] Preferably, in step 6, the preset medical order target includes at least one of the following: the duration of each group and a daily execution plan schedule determined by the number of target groups per day, the execution start time, and the execution interval; The compliance results include at least one of the following: number of completed sets, target number of sets, completion rate, action achievement rate, and most recent completion time.

[0020] The second aspect of the technical solution of the present invention discloses an ankle pump movement compliance monitoring system based on a wearable inertial sensor, used to implement the above-mentioned ankle pump movement compliance monitoring method, characterized in that it includes: Wearable ankle sensors are used to collect and upload ankle joint movement data; The patient-side mobile application is used to receive recognition results, output reminders and feedback to patients, including prompts after completing a set of effective ankle pump movements, reminders that the next target execution time is approaching, and guidance prompts when the movement is not performed correctly. The nursing management platform is used to receive patient adherence results, link medical order goals, summarize the status of multiple patients, and generate alerts and reports, including: The nursing management platform displays patient name, target group number, number of completed groups, most recent completion time, and current status in a bed view, and summarizes the compliance results of multiple patients by ward. The nursing management platform also provides nurses with an interface for configuring daily execution plans. Nurses can set at least one of the following for specific patients: the number of target groups per day, the minimum number of movements per group, the duration of each group, the start time of execution, and the execution interval. The nursing management platform generates a daily execution plan schedule based on the configuration. The execution plan schedule includes multiple target execution times for the day and the corresponding number of target movements or the target duration. The nursing management platform then distributes the execution plan schedule to the patient's mobile application. The patient's mobile application proactively sends timed exercise reminders to the patient within a preset lead time before each target execution time or at the exact time.

[0021] Preferably, the wearable sensor includes an energy storage unit, a power conversion unit, a processing unit, an inertial measurement unit, and a wireless communication module disposed within the housing, and a connecting part disposed on the housing. The power input terminal of the power conversion unit is connected to the power supply terminal of the processing unit, the inertial measurement unit, and the wireless communication module, respectively. The data output terminal of the inertial measurement unit is connected to the data input terminal of the processing unit. The data transmission terminal of the processing unit is connected to the wireless communication module. The connecting part is fixedly connected to an elastic strap. The wearable sensor is fixed to a nearby local area below the lateral malleolus protrusion by the elastic strap, so that the inertial measurement unit is kept within a preset distance range relative to the flexion and extension axis of the ankle joint.

[0022] Preferably, the inertial measurement unit is used to collect local angular velocity data and local acceleration data during ankle dorsiflexion and plantar flexion; the processing unit is used to execute the ankle pump exercise compliance monitoring method, identify effective ankle pump movements based on the collected data, aggregate effective exercise groups and compare them with preset medical order targets to obtain compliance results and warning information; the wireless communication module is used to send at least one of effective ankle pump movements, effective exercise groups, compliance results and warning information to the patient's mobile application or nursing management platform.

[0023] The core of this invention lies in establishing an effective ankle pump action determination rule: only when a single ankle flexion-extension movement collected by an inertial measurement unit fixed at a preset anatomical position below the lateral malleolus simultaneously meets preset movement amplitude conditions, preset movement cycle conditions, preset flexion-extension symmetry conditions, and active conditions is the movement recorded as an effective ankle pump action; then, effective ankle pump actions that continuously reach a preset number of times are aggregated into an effective movement group, and the effective movement group is compared with a preset medical order target, outputting a warning message when the interval exceeds the medical order or the completion rate is insufficient. This rule does not count general ankle joint movements, but rather is a local acquisition, movement identification, and nursing closed-loop mechanism specifically designed around ankle pump nursing actions.

[0024] Compared with existing technical solutions, the present invention has the following beneficial effects: 1) This invention improves the distinguishability of ankle pump motion signals from general whole-body motion signals by fixing an inertial measurement unit to a preset anatomical position on the ankle and collecting data locally for ankle pump care movements. 2) This invention uses the combination of movement amplitude, movement cycle, flexion-extension symmetry and active conditions as the effective ankle pump action judgment rule, which can simultaneously exclude walking, leg shaking and passive activity, and improve the nursing relevance of the action recognition results. 3) This invention further aggregates effective actions into effective movement groups and links them with medical order goals and execution intervals, which can directly output compliance results and nursing warnings, forming a closed loop of "medical order - execution - feedback"; 4) This invention supports nurses to pre-configure the daily target number of sets, minimum number of movements per set, duration of each set, start time of execution, and execution interval on the nursing management platform. Based on this, a daily execution plan schedule is generated, and the patient's mobile application actively outputs timed exercise reminders to the patient at the preset target execution time. This drives the patient to complete ankle pump exercises on time without relying on repeated verbal reminders from nurses or judgments of exceeding the time limit, thus bringing the nursing closed loop forward to the "timed reminder" stage. 5) The structured compliance data generated by this invention can be directly used for ward management, quality control evaluation and clinical research, and has both practical and promotional value. Attached Figure Description

[0025] Figure 1 This is a flowchart of an ankle pump motion compliance monitoring method based on a wearable inertial sensor disclosed in an embodiment of the present invention; Figure 2 This is a schematic diagram showing the wearing position of the wearable sensor in an embodiment of the present invention; Figure 3 This is a schematic diagram of an ankle pump motion compliance monitoring system based on a wearable inertial sensor disclosed in an embodiment of the present invention. Figure 4 This is a schematic diagram of a wearable inertial sensor. Detailed Implementation

[0026] The present invention will be further illustrated below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Furthermore, it should be understood that after reading the teachings of this invention, those skilled in the art can make various alterations or modifications to the invention, and these equivalent forms also fall within the scope defined by the appended claims.

[0027] like Figure 1 As shown, the first aspect of this invention discloses a method for monitoring ankle pump movement compliance based on a wearable inertial sensor, comprising the following steps: Step S100: Fix the wearable sensor containing the inertial measurement unit to the adjacent local area below the lateral malleolus of the patient, so that the inertial measurement unit is kept within a preset distance range relative to the flexion and extension axis of the ankle joint.

[0028] In one preferred embodiment of the present invention, as follows: Figure 2 As shown, the adjacent local area below the lateral malleolus protrusion is a local area located 0-5 cm below the lateral malleolus protrusion of the patient's ankle 20 and at a distance from the center of the lateral malleolus protrusion. The wearable sensor 10 is fixed to the adjacent local area 22 below the lateral malleolus protrusion by an elastic strap, so that the inertial measurement unit is kept within a preset distance range relative to the ankle joint flexion and extension axis 23, so as to obtain local angular velocity signals and local acceleration signals with high signal-to-noise ratio during ankle dorsiflexion and plantarflexion.

[0029] Step S110: Collect ankle joint motion data during the patient's ankle dorsiflexion and plantarflexion processes using wearable sensors. The ankle joint motion data includes local angular velocity data and local acceleration data.

[0030] Step S120: The ankle joint motion data is sequentially filtered, denoised, and angle calculated to obtain preprocessed data characterizing the flexion and extension state of the ankle joint.

[0031] Step S130: Extract four types of feature data from the preprocessed result data, including motion amplitude, motion cycle, flexion-extension symmetry, and active-related features.

[0032] In a preferred embodiment of this invention, the active-related features include at least one of peak angular velocity, amplitude of gravitational component variation, and waveform regularity. These active-related features distinguish ankle pump movements actively performed by the patient from non-targeted movements such as assisted foot movement, getting out of bed and walking, and random leg shaking. When at least one of the peak angular velocity, amplitude of gravitational component variation, and waveform regularity does not meet the preset active-related conditions, the ankle flexion and extension movements are not counted as effective ankle pump movements.

[0033] In a preferred embodiment of the present invention, the flexion-extension symmetry is determined based on the ratio between the peak dorsiflexion and the peak plantarflexion within one movement cycle. When the ratio falls within a preset range, it is determined to meet the preset flexion-extension symmetry range, wherein the preset range can be individually calibrated based on the patient's baseline data.

[0034] Step S140: Determine each ankle flexion and extension movement. Specifically, determine whether the feature data obtained in step S130 simultaneously meets the preset effective ankle pump action determination rules. If it meets the rules, proceed to step S150; otherwise, proceed to step S160.

[0035] In a preferred embodiment of the present invention, the preset effective ankle pump action determination rule is as follows: The current ankle flexion and extension movement is judged as an effective ankle pumping action only when the feature data simultaneously meets the preset range of motion amplitude, preset range of motion cycle, preset range of flexion and extension symmetry, and preset active conditions. Otherwise, it is judged as an invalid ankle pumping action. Passive activities, walking movements, and leg shaking movements are identified as invalid ankle pumping actions and excluded based on the preset active conditions.

[0036] In this embodiment, the effective ankle pump action determination rules include amplitude window, cycle window, flexion-extension symmetry window, active threshold, and passive activity exclusion threshold. These parameters are not fixed but are pre-calibrated based on the nursing plan, standard action samples, non-target action samples, and patient baseline data. During calibration, samples of standard ankle pump actions, walking, leg shaking, and assisted foot swinging can be collected. Features such as motion amplitude, motion cycle, peak dorsiflexion, peak plantarflexion, peak angular velocity, changes in gravity components, and waveform regularity are extracted. Parameter ranges or determination boundaries that can simultaneously distinguish between effective ankle pump actions and non-target actions are selected as default parameters.

[0037] Step S150: Determine the current ankle flexion and extension movement as an invalid ankle pump action and exclude it, then return to step S110.

[0038] Step S160: Determine the current ankle flexion and extension movement as an effective ankle pump action, and proceed to step S170.

[0039] Step S170: Determine whether the number of consecutive effective ankle pump movements has reached the preset number. If not, return to step S110; if so, proceed to step S180.

[0040] Step S180: Record the effective ankle pump movements that have reached a preset number of times as a valid exercise group, and determine the end of the valid exercise group based on the duration of continuous absence of effective ankle pump movements reaching a preset duration.

[0041] In one preferred embodiment of the present invention, the effective exercise group is not counted according to the number of arbitrary movements, but rather the grouping standard is based on the number of effective ankle pump movements that are continuously reached for a preset number of times, and the ending standard is the duration of no effective ankle pump movements for a preset duration. This allows the system to record the quality of the movements and also to correspond one-to-one with the "number of times per group", "duration of each group" and "execution interval" in the nursing orders.

[0042] Step S190: Compare the effective exercise group with the preset medical advice goals to obtain compliance results.

[0043] In one preferred embodiment of the present invention, the preset medical order target includes at least one of the following: execution interval, minimum number of actions per group, duration of each group, and number of target groups per day.

[0044] In a more preferred embodiment of this invention, the preset medical order target further includes at least one of the following: the duration of each group and a daily execution schedule determined by the number of target groups per day, the start time of execution, and the execution interval. The nursing management platform generates the daily execution schedule based on the parameters pre-configured by the nurse. The patient-side mobile application, based on the execution schedule, proactively outputs timed exercise reminders to the patient within a preset lead time before each target execution time, or upon reaching that time, without relying on timeout or target achievement judgments for effective ankle pump movements. The trigger time of the timed exercise reminder is earlier than the recognition time of any effective ankle pump movement within the current group of ankle pump movements. The timed exercise reminder is used to drive the patient to start the current group of ankle pump movements according to the execution schedule and to indicate the target duration or target number of movements for this group.

[0045] In a preferred embodiment of the present invention, the compliance result includes at least one of the following: number of completed groups, number of target groups, completion rate, action achievement rate, and most recent completion time.

[0046] Step S200: For the current valid exercise set, determine: Whether the current time exceeds the execution interval and whether the target is met within the preset allowed time window; Or whether the compliance result is lower than a preset threshold; Or whether the difference in completion rates between the two sides exceeds a preset threshold; If yes, proceed to step S210; otherwise, proceed to step S220.

[0047] Step S210: Output early warning information.

[0048] In a preferred embodiment of the present invention, the warning information includes at least one of the following: impending timeout warning, already timed out warning, low-quality movement warning, and bilateral asymmetric warning. The impending timeout warning and already timed out warning are triggered based on the time relationship between the "current moment" and the "current target execution moment in the execution plan timetable." The bilateral asymmetric warning is triggered based on the difference in completion rates between the left and right ankles. The low-quality movement warning is triggered based on the detected quality of effective ankle pump movements being lower than a preset quality threshold. Step S220: Return to step S110 and continue monitoring.

[0049] In addition to the above methods, this invention also provides another alternative method for monitoring ankle pump exercise compliance. This method first segments suspected ankle flexion and extension segments based on local data collected from the lateral malleolus. Then, it compares the suspected segments with pre-established standard ankle pump action templates and non-target action templates. Combined with the initiative verification results, it determines the effective ankle pump actions. Finally, it accumulates the effective ankle pump actions according to the doctor's order time window and generates compliance results and early warning information.

[0050] like Figure 3 As shown, a second aspect of this invention is to provide an ankle pump exercise compliance monitoring system 100, including a wearable ankle sensor 10, a patient-side mobile application 20, and a nursing management platform 30. The wearable ankle sensor 10 is used to collect and upload ankle joint movement data. The patient-side mobile application 20 is used to receive recognition results, output reminders and feedback to the patient. The nursing management platform 30 can be connected to a hospital information system 40 (e.g., HIS) to receive patient compliance results, link medical order goals, summarize the status of multiple patients, and generate early warnings and reports.

[0051] like Figure 2As shown, the wearable sensor 10 includes an energy storage unit 15, a power conversion unit 13, a processing unit 14, an inertial measurement unit 12, and a wireless communication module 11 disposed within a housing, and a connecting part 16 disposed on the housing. The housing has a flat elliptical shape and is made of a sterilizable material. The power input terminal of the power conversion unit 13 is connected to the energy storage unit 15, and the output terminal is connected to the power supply terminals of the processing unit 14, the inertial measurement unit 12, and the wireless communication module 11, respectively. The data output terminal of the inertial measurement unit 12 is connected to the data input terminal of the processing unit 14, and the data transmission terminal of the processing unit 14 is connected to the wireless communication module 11. The connecting part 16 is fixedly connected to an elastic strap 17, which fixes the wearable sensor 10 to a nearby local area 22 below the lateral malleolus protrusion, so that the inertial measurement unit 12 is kept within a preset distance range relative to the ankle joint flexion and extension axis 23. The elastic strap 17 is preferably an elastic fabric strap, which passes through the connecting part 16 on the back of the housing to form a closed loop structure around the ankle. The inertial measurement unit 12 may be a six-axis inertial sensor, used to collect local angular velocity and local acceleration data during ankle dorsiflexion and plantar flexion. The processing unit 14 may be a low-power microcontroller, used to execute the ankle pump exercise compliance monitoring method, identify effective ankle pump actions based on the collected data, aggregate effective exercise groups, and compare them with preset medical order targets to obtain compliance results and warning information. The wireless communication module 11 is used to send at least one of the effective ankle pump actions, effective exercise groups, compliance results, and warning information to a patient-end mobile application or nursing management platform. When worn, the wearable sensor 10 is located in the adjacent local area below the lateral malleolus protuberance, enabling the inertial measurement unit 12 to obtain more stable local angular velocity and acceleration signals during ankle joint movement.

[0052] In a preferred embodiment of the invention, the inertial measurement unit 12 collects triaxial acceleration data and triaxial angular velocity data at a sampling frequency of 50 Hz to 100 Hz, and sends the raw data to the processing unit 14. The processing unit 14 first performs low-pass filtering and zero-drift correction on the raw data, and then calculates the ankle joint flexion-extension related angle change curve according to a preset coordinate system. Subsequently, the processing unit 14 extracts the motion amplitude, motion cycle, peak dorsiflexion, peak plantarflexion, and active features according to a sliding time window.

[0053] The patient-side mobile application 20 receives sensor recognition results and provides real-time feedback 201 to the patient. Real-time feedback 201 includes prompts after completing a set of effective ankle pump exercises, reminders of the approaching target execution time, and guidance when the exercise is not performed correctly. Effective exercise sets can be configured with grouping and termination conditions based on nursing protocols or individualized patient orders. A set is defined as a series of consecutive effective ankle pump exercises reaching a preset number of repetitions. The current effective exercise set ends when no consecutive effective ankle pump exercises reach a preset duration. Medical order targets can be manually entered by nurses on the nursing management platform 30 or automatically retrieved via the hospital information system 40 interface. Medical order targets include at least one of the following: execution interval, minimum number of repetitions per set, and daily target number of sets. The system compares completed effective exercise sets with the medical order targets, generating completion rates, the most recent completion time, and the achievement rate. It also generates warnings for impending timeout and already expired timeouts based on preset rules.

[0054] The nursing management platform 30 displays the patient's name, number of target groups, number of completed groups, most recent completion time, and current status in a bed view 301, and can summarize the compliance results of multiple patients by ward. The nursing management platform 30 also provides nurses with an interface 302 for configuring daily execution plans. Nurses can set at least one of the following for specific patients: daily target number of groups, minimum number of actions per group, duration of each group, execution start time, and execution interval. Based on the above configuration, the nursing management platform 30 generates a daily execution plan schedule 303, which includes multiple target execution times for the day and the corresponding number of target actions or target duration. The nursing management platform distributes the execution plan schedule 303 to the patient-side mobile application 20. According to the execution plan schedule, the patient-side mobile application 20 proactively outputs timed exercise reminder information 202 to the patient within a preset lead time before each target execution time or upon reaching that time. The timed exercise reminder message 202 is presented via sound, vibration, screen pop-up, or a combination of at least two of the above, and instructs the patient to begin the current set of ankle pump exercises and specify the target number of movements or the target duration for this set. The timed exercise reminder message 202 is triggered at a preset time in the execution schedule, which differs from the existing technology that only outputs an alarm to the nurse's end after detecting timeout or insufficient compliance.

[0055] Before routine monitoring, the system disclosed in this embodiment of the invention can first collect data from the patient's resting segment and, under the guidance of a nurse, collect several standard ankle pump guided movements to individually correct the default parameters. During monitoring, when an ankle flexion-extension movement simultaneously falls within a preset amplitude window, a preset cycle window, and a preset flexion-extension symmetry window, and the active characteristics meet the requirements while the passive activity exclusion characteristics do not, it is recorded as a valid ankle pump movement; otherwise, it is recorded as an invalid movement.

[0056] Example 1: A post-surgical patient's medical order requires at least four sets of ankle pump exercises daily, each lasting two minutes. The nurse attaches a wearable ankle sensor to the area near the patient's right lateral malleolus and enters the patient's medical order into the nursing management platform. After the patient performs the ankle pump exercises, the system identifies valid ankle pump movements and forms valid exercise sets according to the judgment rules disclosed in this embodiment of the invention, synchronizing the results to the mobile application and the nursing management platform. If the patient has not formed a new valid exercise set within the preset reminder time window, the system outputs a timeout warning to the nurse.

[0057] Example 2: Before system deployment, standard ankle pump motion samples, walking samples, leg shaking samples, and family-assisted foot swinging samples were collected. Feature distributions such as motion amplitude, period, peak ratio, peak angular velocity, and gravitational component changes were extracted for each type of sample to determine the default judgment boundary used to distinguish effective ankle pump motions from non-target motions. For specific patients, after the initial wear, standard ankle pump guided motions during resting periods and under nurse guidance were further collected to individually correct the default judgment boundary. Therefore, the judgment parameters are derived from a pre-calibration process, rather than being arbitrarily specified.

[0058] Example 3: A patient took a short walk after getting out of bed wearing a sensor. The system collected significant changes in the gravity component and periodic characteristics different from ankle pump movements. Although foot activity signals were present, the walking activity was not recorded as a valid ankle pump movement because it did not simultaneously meet the preset amplitude window, preset period window, flexion-extension symmetry window, and active condition. This avoids misrepresenting ordinary activity as ankle pump training completion.

[0059] Example 4: A patient at high risk of bilateral lower extremity thrombosis wears sensors on both ankles. The system calculates the completion of effective movement groups on both sides separately. When the completion rate on the right side is consistently lower than that on the left side and the difference exceeds a preset threshold, a bilateral asymmetry warning is sent to the nurse. The nurse can then use this information to further intervene in cases of pain, discomfort, or incomplete movement on the patient's right side.

[0060] Example 5: A patient swings their foot with the assistance of a family member. Although the system collects periodic signals, the change in the gravity component matches the characteristics of passive activity, and the active correlation characteristics do not meet the requirements. Therefore, the corresponding movement is marked as a suspected passive activity and is not counted as a valid ankle pump movement. This example demonstrates that the present invention can distinguish between non-target movements actively performed by the patient and those assisted by others.

[0061] Example 6: A post-orthopedic patient's medical order requires them to perform 6 sets of ankle pump exercises daily, each set lasting 2 minutes, with a 2-hour interval between adjacent sets, and the first set scheduled for 8:00 AM. After the nurse enters the above parameters into the patient configuration interface of the nursing management platform, the platform generates an execution schedule for the day, including six target execution times: 8:00 AM, 10:00 AM, 12:00 PM, 2:00 PM, 4:00 PM, and 6:00 PM, based on the start time and execution interval. This schedule is then sent to the mobile application running on the patient's worn or held mobile device. Five minutes before each target execution time and upon arrival at that time, the mobile application proactively sends a timed exercise reminder to the patient via a pop-up window and vibration, displaying the message "Please start ankle pump exercises; this set is designed to last 2 minutes." After the patient begins exercising as reminded, the wearable ankle sensor identifies valid ankle pump movements according to the aforementioned judgment rules and aggregates them into valid exercise sets. The nursing management platform then updates the completion status of the current set accordingly. If a valid set of exercises that meets the target is not formed within the preset allowed time window after the current target execution time, the system will then output an overdue warning to the nurse according to the aforementioned warning rules. Thus, this invention automatically drives patients to perform ankle pump exercises on time according to a preset execution plan schedule without requiring nurses to manually remind them one by one, and links it with effective action recognition and compliance statistics to form a complete closed loop from nursing order entry, timed reminders, action collection to compliance feedback.

[0062] The embodiments of the present invention have been described above. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of this disclosure. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. The scope of the present invention is defined by the appended claims and their equivalents. Various substitutions and modifications can be made by those skilled in the art without departing from the scope of the present invention, and all such substitutions and modifications should fall within the scope of this disclosure.

Claims

1. A method for monitoring ankle pump movement compliance based on wearable inertial sensors, characterized in that, Includes the following steps: Step 1: Fix the wearable sensor containing the inertial measurement unit to the adjacent local area below the lateral malleolus of the patient, so that the inertial measurement unit is kept within a preset distance range relative to the ankle joint flexion and extension axis. Step 2: Collect ankle joint movement data during dorsiflexion and plantarflexion of the patient's ankle joint using wearable sensors; Step 3: After preprocessing the ankle joint motion data, extract the motion amplitude, motion cycle, flexion-extension symmetry, and active features, including: By using initiative-related characteristics, we can distinguish ankle pump movements that patients perform on their own from non-targeted movements such as foot movements assisted by others, getting out of bed and walking, and random leg shaking. Flexion-extension symmetry is determined based on the proportional relationship between the peak dorsiflexion and the peak plantarflexion within one motion cycle; Step 4: Determine whether the feature data simultaneously meets the preset rules for determining effective ankle pumping motion. If it does, the current ankle flexion and extension motion is determined to be an effective ankle pumping motion, and proceed to Step 5. If it does not meet the rules, the current ankle flexion and extension motion is determined to be an invalid ankle pumping motion and excluded, and return to Step 2. Step 5: Determine whether the number of consecutive effective ankle pump movements has reached the preset number. If not, return to step 2. If it has, record the consecutive effective ankle pump movements that have reached the preset number as a valid exercise group, and determine the end of the valid exercise group based on the duration of consecutive ineffective ankle pump movements reaching the preset duration. Step 6: Compare the effective exercise group with the preset medical advice goals to obtain compliance results; Step 7: For the current valid exercise set, determine: Whether the current time exceeds the execution interval and whether the target is met within the preset allowed time window; Or whether the compliance result is lower than a preset threshold; Or whether the difference in completion rates between the two sides exceeds a preset threshold; If yes, output a warning message; otherwise, return to step 2.

2. The ankle pump motion compliance monitoring method based on a wearable inertial sensor as described in claim 1, characterized in that, In step 1, the adjacent local area is a local area located below the lateral malleolus and 0-5 cm from the center of the lateral malleolus, so as to obtain ankle joint motion data with a high signal-to-noise ratio during ankle dorsiflexion and plantarflexion by the inertial measurement unit.

3. The ankle pump motion compliance monitoring method based on a wearable inertial sensor as described in claim 1, characterized in that, In step 3, the active correlation features include at least one of the following: peak angular velocity, amplitude of change of gravity component, and waveform regularity.

4. The ankle pump motion compliance monitoring method based on a wearable inertial sensor as described in claim 3, characterized in that, In step 3, the preset effective ankle pump action determination rule is: the current ankle joint flexion and extension movement is judged as an effective ankle pump action only when the feature data simultaneously meets the preset range of motion amplitude, preset range of motion cycle, preset range of flexion and extension symmetry and preset active conditions; otherwise, it is judged as an invalid ankle pump action. Passive activities, walking movements and leg shaking movements are identified as invalid ankle pump actions and excluded through the preset active conditions.

5. The ankle pump motion compliance monitoring method based on a wearable inertial sensor as described in claim 4, characterized in that, In the preset effective ankle pump action determination rules: Regarding the aforementioned active-related characteristics, if at least one of the peak angular velocity, the amplitude of the change in the gravitational component, or the waveform regularity does not meet the preset active conditions, the ankle flexion and extension movement will not be recorded as an effective ankle pump action. Regarding the flexion-extension symmetry, when the proportional relationship is within a preset range, it is determined that the preset flexion-extension symmetry range is met. The preset range can be individually calibrated based on the patient's baseline data.

6. The ankle pump motion compliance monitoring method based on a wearable inertial sensor as described in claim 4, characterized in that, The parameters of the preset effective ankle pump action judgment rule are pre-calibrated based on the nursing plan, standard action samples, non-target action samples, and patient baseline data. During calibration, samples such as standard ankle pump actions, walking, leg shaking, and foot swinging assisted by others are collected to extract the amplitude of movement, movement cycle, peak dorsiflexion, peak plantarflexion, peak angular velocity, changes in gravity components, and waveform regularity. The parameter range or judgment boundary that can simultaneously distinguish between effective ankle pump actions and non-target actions is selected as the default parameter.

7. The ankle pump motion compliance monitoring method based on a wearable inertial sensor as described in claim 1, characterized in that, In step 6, the preset medical order target includes at least one of the following: the duration of each group and the daily execution plan schedule determined by the number of target groups per day, the execution start time, and the execution interval; The compliance results include at least one of the following: number of completed sets, target number of sets, completion rate, action achievement rate, and most recent completion time.

8. An ankle pump movement compliance monitoring system based on a wearable inertial sensor, used to implement the ankle pump movement compliance monitoring method according to any one of claims 1 to 7, characterized in that, include: Wearable ankle sensors are used to collect and upload ankle joint movement data; The patient-side mobile application is used to receive recognition results, output reminders and feedback to patients, including prompts after completing a set of effective ankle pump movements, reminders that the next target execution time is approaching, and guidance prompts when the movement is not performed correctly. The nursing management platform is used to receive patient adherence results, link medical order goals, summarize the status of multiple patients, and generate alerts and reports, including: The nursing management platform displays patient name, target group number, number of completed groups, most recent completion time, and current status in a bed view, and summarizes the compliance results of multiple patients by ward. The nursing management platform also provides nurses with an interface for configuring daily execution plans. Nurses can set at least one of the following for specific patients: the number of target groups per day, the minimum number of movements per group, the duration of each group, the start time of execution, and the execution interval. The nursing management platform generates a daily execution plan schedule based on the configuration. The execution plan schedule includes multiple target execution times for the day and the corresponding number of target movements or the target duration. The nursing management platform then distributes the execution plan schedule to the patient's mobile application. The patient's mobile application proactively sends timed exercise reminders to the patient within a preset lead time before each target execution time or at the exact time.

9. The ankle pump motion compliance monitoring system based on a wearable inertial sensor as described in claim 8, characterized in that, The wearable sensor includes an energy storage unit, a power conversion unit, a processing unit, an inertial measurement unit, and a wireless communication module disposed within the housing, and a connecting part disposed on the housing. The power input terminal of the power conversion unit is connected to the energy storage unit, and the output terminal is connected to the power supply terminals of the processing unit, the inertial measurement unit, and the wireless communication module, respectively. The data output terminal of the inertial measurement unit is connected to the data input terminal of the processing unit, and the data transmission terminal of the processing unit is connected to the wireless communication module. The connecting part is fixedly connected to an elastic strap, and the wearable sensor is fixed to a nearby local area below the lateral malleolus protrusion by the elastic strap, so that the inertial measurement unit is kept within a preset distance range relative to the flexion and extension axis of the ankle joint.

10. The ankle pump motion compliance monitoring system based on a wearable inertial sensor as described in claim 8, characterized in that, The inertial measurement unit is used to collect local angular velocity data and local acceleration data during ankle dorsiflexion and plantar flexion; the processing unit is used to execute the ankle pump exercise compliance monitoring method, identify effective ankle pump movements based on the collected data, aggregate effective exercise groups and compare them with preset medical order targets to obtain compliance results and early warning information; the wireless communication module is used to send at least one of effective ankle pump movements, effective exercise groups, compliance results and early warning information to the patient's mobile application or nursing management platform.

Images