TW202202094A - Device for turning over detection - Google Patents

Abstract

A device for turning over detection is provided. The device including: an optical fiber sensor pad, adapting to allow people lie on it； an optical fiber sensor module, located inside the optical fiber sensor pad, turning the pressure into a light signal； a light sensor, coupling to the optical fiber sensor module, turning the light signal into a measured signal and delivering the measured signal to a signal processing module； a signal processing module, analyzing the measured signal and determining the event time for turning-over, back-slapping and lying base on the waveform feature of the measured signal.

Description

Turn over care detection system

The present invention relates to a turning over care detection system, in particular to a turning over care detection system capable of judging a care action.

When the skin is under concentrated pressure for a long time, hypoxia occurs in this part of the skin tissue due to insufficient blood perfusion, which will lead to necrosis of the skin tissue and the formation of bedsores. For the paralyzed or disabled care recipient, the caregiver must regularly turn the body of the care recipient or tap the pressure area to circulate blood flow to avoid pressure sores. Generally speaking, turning over care can be divided into turning over and turning over and slapping the back. For example, after the person being cared for eating or drinking, only turning over and not applying the slap on the back; and turning over can be divided into turning to the left or turning to the right. In order to avoid long-term pressure on the skin tissue in a specific location. A typical rollover care cycle is every two hours.

In modern families, there are few members. When there is a stroke in the family or a person who has difficulty in mobility due to other diseases, usually must entrust an external caregiver to take care of the person in the family when other family members go out, and whether the person in care is properly cared for That's what every client worries about. On the other hand, for the manager of the care center, when there are multiple beds of disabled care recipients, it is necessary to manually record the time of each turn over and back pat for each care recipient to ensure regular care. However, whether it is outsourcing or clinical management, it is inevitable that Baimi has some sparseness and is very labor-intensive.

The present invention proposes a rollover care detection system in order to solve the difficult clinical problem of care management of turning over and slapping the back.

In one embodiment, the turning over care detection system of the present invention includes an optical fiber sensing pad, an optical fiber sensing module, a photoelectric conversion module, and a signal processing module. When the care recipient is lying on the optical fiber sensing pad, the optical fiber sensing module located inside the optical fiber sensing pad can convert the pressure applied by the care recipient to the optical fiber sensing pad to affect the physical properties of the optical signal, such as intensity, frequency , phase and other physical properties. The photoelectric conversion module is coupled to the optical fiber sensing module, and can convert the optical signal into a measurement signal and then transmit it to the signal processing module. The signal processing module analyzes the measurement signal to determine the occurrence time point of the turning event, the back slapping event or the lying state according to the waveform characteristics of the measurement signal; wherein, the turning event corresponds to the first judgment signal, which means that the turning event has an amplitude greater than the first threshold. The back-beating event corresponds to the second judgment signal, which refers to the measurement signal whose periodic amplitude is less than the first threshold and greater than the second threshold; the lying state corresponds to the third judgment signal, which refers to the first judgment signal and the second judgment signal Measurement signals other than the time when the signal occurs.

By classifying the movement signals to obtain the occurrence time of each movement, the system can remind the caregiver to perform the turning movement regularly, and can also provide the records to the caregiver or the care center manager to ensure that all patients at home or in the hospital have a regular schedule. be well cared for.

Statistically significant in this manual refers to the statistical test result P>0.05; turning over refers to turning from a lying state to a state where the back is not completely attached to the bed surface; slapping the back refers to applying multiple times to the back in the turning state slap.

FIG. 1 is a schematic block diagram of a turning over care detection system according to an embodiment of the present invention. Referring to FIG. 1 , the turning over care detection system 1 includes an optical fiber sensing pad 11 , an optical fiber sensing module 12 , a photoelectric conversion module 13 and a signal processing module 14 .

In one embodiment, as shown in FIG. 2 , the user can lay one or more optical fiber sensing pads 11 on the mattress, and each optical fiber sensing pad 11 can maintain a certain distance to cover the bed of the care recipient. Apply pressure to larger areas such as the head, torso, or legs while lying down. The optical fiber sensing pad 11 may have one or more sets of optical fiber sensing modules 12 inside, and the optical fiber sensing modules 12 convert the pressure exerted by the caregiver on the optical fiber sensing pad 11 to affect the physical properties of the optical signal, such as amplitude, Physical properties such as intensity, frequency, phase, etc. The optical fiber sensing module 12 can use an intrinsic optical sensor (Intrinsic optic sensor), which utilizes the optical fiber itself to produce optical attenuation due to compression deformation or vibration; The light source is guided to the light modulating medium, and the light modulating medium changes its light transmission properties due to pressure. The photoelectric conversion module 13 is coupled to the optical fiber sensing module 12 by an optical fiber cable to receive the optical signal. The photoelectric conversion module 13 includes a light sensor to convert the optical signal into a measurement signal, and then transmits the measurement signal to the signal processing module 14 through a transmission line or a wireless transmission device. The signal processing module 14 can be an integrated circuit or software applied to a computer. The signal processing module 14 can first perform signal preprocessing on the measurement signal, including the steps of signal sampling, filtering, threshold calculation, and signal smoothing. Then, the classification process is performed, and the rollover event, the back slap event or the lying state is classified according to the waveform characteristics of the measurement signal.

For signal preprocessing, since the frequency of the turn-over-beat signal is below 5~10 Hz, the sampling frequency of the measurement signal is set to 10~20 Hz to avoid signal distortion under the condition of avoiding over-sampling. Signal filtering adopts 10 Hz low-pass filtering to avoid high-frequency electromagnetic wave noise interference. The threshold calculation can be customized according to the care recipients of different body types; the multiple of the standard deviation can also be used as the threshold; the dynamic threshold algorithm (Adaptive thresholding) can also be used to obtain the threshold. Signal smoothing adopts moving average or low-pass filtering to blur the signal, which is helpful to find the changing trend.

For the classification process, FIG. 3A , FIG. 3B and FIG. 3C are three situations classified according to the waveform characteristics of the measurement signal in an embodiment of the turning over care detection system 1 of the present invention. FIG. 3A is a waveform diagram of the turning event corresponding to the first determination signal 21 in an embodiment. When the caregiver turns over the person who is lying on the optical fiber sensing pad 11 of the present invention, the instantaneous pressure is applied to the optical fiber sensing pad 11 due to the turning action, which will cause the amplitude of the measurement signal to be greater than the first threshold 31 and the time scale A single pulse of about a typical rollover time. The aforementioned single pulse wave is a measurement signal with an amplitude greater than the first threshold 31 , and belongs to the first judgment signal 21 ; FIG. 3B is a waveform diagram corresponding to the second judgment signal 22 of the beat back event in one embodiment. When the caregiver slaps the back of the care recipient lying on the optical fiber sensing pad 11 of the present invention, periodic vibrations are generated by the back slap action to the optical fiber sensing pad 11, causing the measurement signal to have periodicity within a preset time. There is a high probability of occurrence of complex pulse waves with amplitudes less than the first threshold 31 and greater than the second threshold 32 . For the aforementioned complex pulse waves, the time scale of each pulse wave is about the slap time of a single slap on the back. The aforementioned complex pulse waves are measurement signals whose periodic amplitude is smaller than the first threshold value 31 and greater than the second threshold value 32, and belong to the second determination signal 22; FIG. 3C shows the waveform pattern corresponding to the third determination signal 23 in the lying state in one embodiment. picture. When the care receiver is lying on the fiber optic sensing pad 11 of the present invention, the care receiver cannot move by himself, so a constant pressure is applied to the fiber optic sensing pad 11 , so that the measurement signal is between the first judgment signal 21 and the second judgment signal 22 Others are mainly stable signals in a state less than the second threshold 32 . The aforementioned stable signal is the measurement signal other than the time when the first determination signal 21 and the second determination signal 22 are generated, and belongs to the third determination signal 23 .

In one embodiment, since the clinical caregiver will customize the back-beating cycle, the preset time is the back-beating cycle customized by the caregiver to determine whether the second judgment signal 22 occurs within the preset time. In one embodiment, since the usual turning care cycle is every two hours, the preset time is two hours to determine whether the second determination signal 22 occurs within the preset time.

In one embodiment, there is a first threshold 31, so that the amplitude of the peak of the first judgment signal 21 is statistically significantly larger than the amplitude of the measurement signal when the care recipient is in a lying state; there is a second threshold 32, so that the second The amplitude of the peak of the determination signal 22 is statistically significantly larger than the amplitude of the measurement signal when the care recipient is lying down, and is significantly lower than the amplitude of the peak of the first determination signal 21 .

In one embodiment, there is a first threshold 31, so that the peak amplitude of the first judgment signal 21 is statistically significantly larger than the moving average of the measurement signal; there is a second threshold 32, so that the peak amplitude of the second judgment signal 22 is Statistically, it is significantly larger than the moving average line of the measurement signal, and is significantly lower than the peak amplitude of the first judgment signal 21 .

In one embodiment, there is a first threshold 31, so that the occurrence of the first judgment signal 21 and the occurrence of turning over are statistically significantly positively correlated; there is a second threshold 32, so that the occurrence of the second judgment signal 22 There is a statistically significant positive correlation between the occurrence or not and the occurrence of back slap action.

In another embodiment of the present invention, in order to improve the recognition accuracy of the turning event, the back-slapping event or the lying state, the signal processing module 14 applies a machine learning algorithm to process. Because clinical turning over care can be divided into turning over and turning over and slapping the back, for example, after the person being cared for eating or drinking, only turning over and not applying the slap on the back. However, in order to perform back pat care, the caregiver must first turn the person under care from a lying state, and then perform back pat on the person under care. In short, the back-slapping action is not necessarily accompanied by the turning-over action, but the turning-over action must have occurred before the back-slapping action. The foregoing conclusion corresponds to the measurement signal, and it can be seen that the second determination signal 22 may not necessarily appear after the first determination signal 21 appears, but the first determination signal 21 must appear before the second determination signal 22 appears. Therefore, the appearance of the second determination signal 22 can be used as a training data feature for the machine learning algorithm to identify the first determination signal 21 .

FIG. 4 is a flow chart of another embodiment of the turning over care detection system of the present invention. Referring to FIG. 4 , after the signal processing module 14 receives the measurement signal, it starts signal preprocessing and classification processing (step S01 ). If the signal processing module 14 determines that the first determination signal 21 is generated during the classification process (step S02 ), the signal processing module 14 will intercept the waveform of the first determination signal 21 and temporarily store it in the memory, and start timing ( Step S03). After that, in the process of processing the measurement signal (step S04), if the signal processing module 14 does not determine that the second determination signal 22 is generated until the preset time expires (step S05), the timer is reset to zero (step S07) and the Empty the waveform of the first judgment signal 21 temporarily stored in the memory, and then continue to process the measurement signal (step S01); in the process of continuing to process the measurement signal (step S04), if it is determined that the second judgment signal occurs within the preset time 22 (step S05 ), the waveform of the first judgment signal 21 temporarily stored in the memory and the waveform of the second judgment signal 22 are input into the machine learning algorithm (step S06 ), and the waveform of the first judgment signal 21 is input to the machine learning algorithm (step S06 ) The flipping event is used as a label, and the waveform of the second judgment signal 22 is used as a label of the flipping event. After that, reset the timer (step S07 ) and clear the waveform of the first judgment signal 21 temporarily stored in the memory, and then continue to process the measurement signal (step S01 ). Through the processing method of the process shown in FIG. 4 , the machine learning algorithm can adapt to the differences of different individuals in the clinic, and can gradually acquire the corresponding relationship between the waveform of the first judgment signal 21 and the turning event and the second judgment signal 22 in real time. Correspondence between wave patterns and rollover events plus back-beat events. In the long run, the signal processing module 14 can simultaneously use the first threshold 31 and the second threshold 32 and the waveform as the classification basis to more accurately determine whether the first judgment signal 21 and the second judgment signal 22 occur.

For example, for a first-time admission patient, using traditional supervised learning algorithms to learn the correlation between the turning action and the measurement signal, it is necessary to turn the patient on the bed multiple times, record the turning time, and then reconcile with the patient. Measure signal comparisons to obtain training data. After the training is completed, the system can be applied to the first-time hospitalized patients; using the algorithm of the turning over care detection system 1 of the present invention in one embodiment, the system can be directly applied to the first-time hospitalized patients, first using the first threshold 31 as the classification basis Perform classification work, and at the same time obtain training data to input into machine learning algorithms to improve classification capabilities.

In another embodiment of the present invention, in order to distinguish whether the turning action is turning left or turning right, the optical fiber sensing pad 11 may have multiple sets of optical fiber sensing modules 12 . In clinical care, turning over can be divided into turning to the left or turning to the right to avoid long-term pressure on the skin tissue in a specific position. Since the choice of turning left or right will result in different pressure concentration points after turning over, the difference between turning left and turning right can be distinguished by means of multiple sets of optical fiber sensing modules 12 .

FIG. 5 is a schematic diagram of the use state of the turning over care detection system of the present invention in another embodiment. Referring to FIG. 5 , the turning over care detection system 1 includes an optical fiber sensing pad 11 , multiple sets of optical fiber sensing modules 12 , multiple sets of photoelectric conversion modules 13 and signal processing modules 14 . When the caregiver turns the care recipient over, the pressure of each optical fiber sensing module 12 will change. The signal processing module 14 can receive measurement signals from a plurality of groups of photoelectric conversion modules 13, and summarizes a plurality of groups of functional relationships that the pressure experienced by each optical fiber sensing module 12 increases or decreases with time, so as to understand the movement of the care recipient. direction.

For example, when the care recipient is lying in the middle position of the fiber optic sensing pad 11 shown in FIG. 5 , the fiber optic sensing module 12 in the middle is under the greatest pressure at this time. When the caregiver turns the care recipient to the right, the pressure on the optical fiber sensing module 12 in the middle will gradually decrease and the pressure on the optical fiber sensing module 12 on the right will gradually increase. The signal processing module 14 determines to turn right according to the decrease of the signal level measured by the optical fiber sensing module 12 in the middle and the increase of the signal level measured by the right optical fiber sensing module 12 .

For example, when the care recipient is lying on the right side of the fiber optic sensing pad 11 shown in FIG. 5 , the fiber optic sensing module 12 on the right side is under the greatest pressure. When the caregiver moves the care recipient to the right to get out of bed, the pressure on the fiber optic sensing module 12 on the right will gradually decrease while the pressure on the middle fiber optic sensing module 12 may remain unchanged or decrease gradually. The signal processing module 14 determines that the care recipient has left the optical fiber sensor according to the decrease of the signal level measured by the optical fiber sensing module 12 on the right and the maintenance or decrease of the signal level measured by the optical fiber sensing module 12 in the middle. Measuring pad 11.

In another embodiment of the present invention, the turning over care detection system 1 or the management system 4 includes a comparison module, which compares the level change of the measurement signal transmitted by each group of photoelectric conversion modules 13 with the level of the standard measurement signal Changes are compared for differences, and the result of whether the alignment is the same or not is output. In the general rollover care process, first, the hand of the lying patient must be placed on the chest; then, the patient's knees are erected to bend the legs; finally, the patient is turned sideways. The common clinical problem is that the caregivers lack perfect education and training or are unfamiliar with the turning technique, and use the wrong turning movements. Wrong turning over may cause the effect of preventing local compression to be unachievable, and may even cause the care recipient to be injured. Configure a plurality of turn over care detection systems 1 of the present invention to understand the pressure changes of various parts of the body when the care recipient is moved, and compare the pressure changes caused by a standard action to assist in performing education and training or reminding wrongdoers. Turn over care operations.

For example, as shown in FIG. 5 , a plurality of turning over care detection systems 1 are configured to correspond to the positions of the chest, legs and the like of the person being cared for. When the care recipient is lying down, the optical fiber sensing pads 11 located on the chest and legs are subject to stable pressure. At this time, the signal processing module 14 receives the stable measurement signal level sent from the photoelectric conversion module 13 of the chest and the legs; In the suspended state, the pressure of the fiber optic sensing pad 11 located on the leg drops sharply. At this time, the signal processing module 14 receives the level of the measurement signal transmitted by the photoelectric conversion module 13 from the chest and maintains the level, and the level of the measurement signal transmitted by the photoelectric conversion module 13 from the leg decreases. When the patient is turned to the right, the pressure on the right part of the fiber optic sensing pad 11 on the chest and legs rises sharply. At this time, the signal processing module 14 receives the measurement from the optical fiber sensing module 12 on the right side of the optical fiber sensing pad 11 of the chest and legs, and the level of the measurement signal transmitted by the photoelectric conversion module 13 increases. During the nursing process, the signal processing module 14 can establish the six sets of measurement signal time functions measured by all six sets of optical fiber sensing modules 12 on the chest and legs and transmitted by the photoelectric conversion module 13 to determine the time function of the nursing recipient. how to be moved. In one embodiment, the comparison module compares the sets of functions corresponding to the turning care operations of the trained caregivers with the sets of functions of the standard operations performed by the trained caregivers to confirm the Whether the procedures performed by the trained caregivers are correct. The user can also configure more rollover care detection systems 1 to cover multiple parts of the body to improve the accuracy of judgment.

In another embodiment of the present invention, in order to manage a plurality of turning over care and detection systems 1 for managing patients with multiple beds or managing single-bed patients, the turning over care and detection system 1 includes a communication module 15, which transmits data through wired or wireless transmission. Sent to management system 4. The management system 4 may be software applied to a computer. Traditionally, when there are multiple beds of care recipients with inconvenience, it is necessary to manually record the time of each turn over, turn over and back, turn left or right, and ensure that they are cared for regularly. The rollover care detection system 1 of the present invention not only provides automatic recording of the time points when the person under care performs different rollover care, but also can send relevant information externally to facilitate management.

FIG. 6 is a block schematic diagram of another embodiment of the turning over care detection system of the present invention. Referring to FIG. 6 , a plurality of sets of turning over care detection systems 1 each include an optical fiber sensing pad 11 , an optical fiber sensing module 12 , a photoelectric conversion module 13 , a signal processing module 14 and a communication module 15 . The communication module 15 can send the time point of the turning event, the back slapping event or the lying state to the management system 4 . The communication module 15 can send the measurement signal to the management system 4 . The management system 4 provides the administrator with a warning according to whether the last time the rollover event or the back pat event occurred in each rollover care detection system 1 has exceeded a preset time from the present.

1: Turn over care detection system 11: Optical fiber sensing pad 12: Optical fiber sensing module 13: Photoelectric conversion module 14: Signal processing module 15: Communication module 21: The first judgment signal 22: The second judgment signal 23: The third judgment signal 31: First Threshold 32: Second threshold 4: Management system S01~S07: Steps

[FIG. 1] is a block schematic diagram of an embodiment of the turning over care detection system of the present invention. [FIG. 2] is a schematic diagram of the use state of the optical fiber sensing pad of the present invention in an embodiment. FIG. 3A is a waveform diagram of a turning event corresponding to the first judgment signal in an embodiment. FIG. 3B is a waveform diagram of the second judgment signal corresponding to the back shot event in an embodiment. FIG. 3C is a waveform diagram of the third judgment signal corresponding to the lying state in an embodiment. [FIG. 4] is a flow chart of another embodiment of the turning over care detection system of the present invention. [Fig. 5] is a schematic diagram of the use state of the turning over care detection system of the present invention in another embodiment. Fig. 6 is a block schematic diagram of another embodiment of the turning over care detection system of the present invention.

1: Turn over care detection system

11: Optical fiber sensing pad

12: Optical fiber sensing module

13: Photoelectric conversion module

14: Signal processing module

Claims (9)

A rollover care detection system, comprising: an optical fiber sensing pad for the care recipient to lie on; an optical fiber sensing module, which is placed in the optical fiber sensing pad body, and converts a pressure experienced into an optical signal; a photoelectric conversion module, coupled to the optical fiber sensing module, to convert the optical signal into a measurement signal; and A signal processing module, coupled to the photoelectric conversion module, analyzes the measurement signal to determine a turning-over event, a back-beating event and a lying state according to the waveform characteristics of the measurement signal; wherein, the turning-over event corresponds to a first judgment signal, the first judgment signal is defined as a measurement signal with an amplitude greater than a first threshold; the beat back event corresponds to a second judgment signal, and the second judgment signal is defined as a periodic occurrence with an amplitude smaller than the first The threshold value is greater than a measurement signal of a second threshold value; the lying state corresponds to a third judgment signal, and the third judgment signal is defined as the measurement signal other than the time point when the first judgment signal and the second judgment signal occur. The rollover care detection system according to claim 1, wherein the first threshold is such that the occurrence of the first judgment signal is positively correlated with the occurrence of rollover. The turning over care detection system according to claim 1, wherein the second threshold is such that there is a statistically significant positive correlation between the occurrence of the second judgment signal and the occurrence of back pats. The rollover care detection system as claimed in claim 1, wherein the signal processing module applies a machine learning algorithm to perform classification of the rollover event, the back tapping event and the lying state. According to the turning over care detection system of claim 4, after the first determination signal is generated in the measurement signal, if the second determination signal occurs within a preset time, the waveform of the first determination signal is determined And/or the waveform of the second judgment signal is input to the machine learning algorithm as training data. The turning over care detection system according to claim 1, wherein the number of the optical fiber sensing module and the photoelectric conversion module is plural. According to the turning over care detection system as described in claim 6, the signal processing module establishes a plurality of functional relationships of the levels of the measurement signals transmitted by the photoelectric conversion modules, so as to determine whether the care recipient is moving. The turning over care detection system as claimed in claim 7 further includes a comparison module, which performs a difference comparison between the functional relationship and a functional relationship of a standard operation, and outputs a result of whether the comparison is the same. The rollover care detection system according to any one of claims 1 to 8, further comprising a communication module suitable for sending the measurement signal and/or the occurrence of the rollover event, the back-slapping event and the lying state Time to a management system.

Images