Old-age nursing turning-over device and method

Abstract

The turnover device comprises a bed body, the two ends of the bed body are rotationally connected with turnover plates which can turn over to achieve the turnover action of a patient, the two ends of the bed body are rotationally connected with guardrails, and the guardrails are connected with the bed body through tension springs; the guardrail is coaxially and fixedly connected with a large gear, the large gear is provided with a clamping groove, the bed body is slidably connected with a rack capable of sliding into the clamping groove, the end, away from the turning plate, of the clamping groove is provided with an inclined face, and the inclined face enables the guardrail to rotate outwards within a specified angle. According to the turnover bed, the arranged guardrails can play an auxiliary supporting role on the patient, the rotatable guardrails cannot extrude the patient, the safety of the turnover bed is effectively improved, and therefore the turnover action of the patient is facilitated.

Description

Technical Field

[0001] The present invention relates to the field of nursing technology, and in particular to a turning device and method for elderly nursing care. Background Art

[0002] In medical examinations and long-term home care, a turning bed is an auxiliary bed that helps patients who have difficulty turning over to turn over. If the patient relies solely on his or her own strength to turn over, it is easy to cause secondary injury to the patient, especially for some critically ill patients who can no longer turn over by their own strength. However, if manpower is used to help the patient turn over without the help of any equipment, it requires multiple people to work together and apply force together. Multiple people standing side by side will also prevent medical staff from conveniently operating on the patient's back, and the force applied by multiple people is likely to cause improper force application, causing secondary injury.

[0003] The patent with announcement number CN 211986012 U discloses a turning bed, including a bed frame, bed legs are arranged at the four corners of the bottom of the bed frame, and baffles are arranged on both sides of the bed frame; the turning bed also includes a bed board assembly for being placed on the bed frame, the bed board assembly includes a first plate body, a second plate body and a third plate body, the first plate body is close to the head of the turning bed, the second plate body and the third plate body are arranged side by side and close to the tail of the turning bed; a contraction assembly is arranged at both ends of the first plate body close to the head of the bed, the contraction assembly includes an elastic member and a connecting plate provided on the top of the elastic member, a slide rail is provided on the connecting plate, two magnetic members are slidably provided on the slide rail, and the two magnetic members attract each other.

[0004] However, there are still the following problems: when turning over, if the patient's lying position is off-center, or there are bedding or sundries piled on one side of the bed, the patient may be squeezed onto the railing when the bed rotates to turn the patient over, causing secondary injury to the patient; at the same time, since the patient's initial position is different each time, it is easy to over-turn or under-turn when turning over. Summary of the invention

[0005] In order to solve the above problems in the prior art, a turning device and method for elderly care are provided.

[0006] The technical solution adopted by the present invention to solve its technical problem is:

[0007] The present invention proposes a turning device for elderly care, comprising a bed body, characterized in that flaps that can be turned over to realize the patient turning over are rotatably connected at both ends of the bed body, and guardrails are rotatably connected at both ends of the bed body, and the guardrails are connected to the bed body through tension springs; the guardrail is coaxially fixedly connected to a large gear, and the large gear is provided with a slot, and the bed body is slidably connected to a rack that can slide into the slot, and an inclined surface is provided at the end of the slot away from the flap, and the inclined surface enables the guardrail to rotate outward within a specified angle.

[0008] Preferably, the bed body is rotatably connected with a spline shaft, the spline shaft is connected with the bed body through a torsion spring, a first gear and a second gear are fixedly sleeved on the spline shaft, the second gear meshes with the large gear, and the first gear meshes with the rack.

[0009] Preferably, the bed body is connected with a motor, the motor is connected with a main shaft, a spline sleeve is slidably connected to the main shaft, and the spline sleeve is connected with the main shaft through a spring.

[0010] Preferably, the elastic force of the spring can make the spline sleeve slide onto the spline shaft, the guardrail is fixedly connected with an insertion rod, and the insertion rod is located on one side of the spline sleeve; the elastic force of the tension spring is greater than that of the spring, and when the insertion rod contacts the spline sleeve, the spline sleeve can be disengaged from the spline shaft.

[0011] Preferably, both the first gear and the second gear are toothless gears. When the first gear disengages from the rack, the rack slides out of the card slot, and then the second gear will mesh with the large gear.

[0012] Preferably, a switch capable of controlling the operation of the motor is connected to the rack, and an alarm switch capable of controlling the motor to stop operating is connected to the guardrail; when the switch moves out of the card slot, the switch controls the motor to start operating, and when the motor drives the guardrail to rotate to a specified angle, the alarm switch controls the motor to stop operating and alarms simultaneously.

[0013] Preferably, the bed body is connected with a flap motor, the flap motor is fixedly connected with an output shaft, symmetrically arranged ejector rods are fixedly connected to the output shaft, and the ejector rods contact the lower surface of the flap and can slide along the lower surface of the flap.

[0014] An elderly care turning method is characterized in that the above-mentioned elderly care turning device is adopted, and it includes the following steps:

[0015] Data acquisition: The rotation of the flap realizes the turning action of the patient. Based on the data acquisition module, the positions of the patient before and after each turn on the bed body are collected, and the characteristic information of the position after turning is subjected to feature extraction by the feature extraction module. The feature extraction module includes a number of neural network models;

[0016] Data processing: Based on the feature extraction module, randomly combine a number of neural network models, and then reconstruct the position image of the extracted features through an autoencoder and compare it with the standard turning position;

[0017] When the reconstructed image is most similar to the standard turning position, it is determined that the combination of the current feature extraction network is the optimal combination, and the optimal features extracted by the feature extraction network of the current combination are forward propagated and the subsequent processing of the features is performed through the control module;

[0018] Model establishment: Record the control strategy of the elderly care turning device corresponding to the optimal combination, enhance the key feature data of the pre-turning position, optimize the representation of data features by dynamically adjusting weights, and after training the optimized features, obtain a feature library composed of image key descriptors;

[0019] Turning control: Collect the position where the patient is before turning, and match it with the image key descriptors in the feature library. Based on the K-means combined with the HITS algorithm, output the feature in the feature that is closest to the query image as the judgment result, and use the control strategy corresponding to this feature to control the elderly care turning device to realize the turning action of the patient.

[0020] Preferably, the method for comparing the reconstructed image with the standard turning position includes:

[0021] After flattening the multi-dimensional features of the original image and the reconstructed image into a one-dimensional vector, normalize the feature vector, then subtract the two one-dimensional vectors, then sum the absolute values, calculate the L1 norm of the two feature vectors, and use the L1 norm to calculate the distance between the two vectors to represent their similarity;

[0022] Then perform two fully connected operations on the obtained distance. Connect the second fully connected layer to a neuron, and take the Sigmoid function of its result. The role of the fully connected layer is to process the distance calculated by the L1 norm through two fully connected layers. The application of the Sigmoid function maps the output of the second fully connected layer to the [0,1] interval through the Sigmoid function to represent the similarity of the two input pictures;

[0023] When the similarity of the two pictures reaches the maximum, it means that the features extracted by the combination of the current neural network model are the optimal features, and forward propagate the optimal features extracted by the feature extraction network of the current combination, and perform subsequent processing of the features through the control module.

[0024] Preferably, the method for dynamically adjusting weights includes: Initializing the weight matrix: Assign an initial weight to each data feature, and the weight matrix is represented as W, and the formula is as follows:

[0025]

[0026] where, w i is the initial weight of the i-th feature, and n is the number of features;

[0027] Calculating the weighted eigenvalue: Use the weight matrix to weight the feature data X, and multiply each eigenvalue xi by its corresponding weight w i The formula is as follows:

[0028]

[0029] In the formula, Z is the weighted feature vector, representing the comprehensive index of the eigenvalues after weight processing, and x i is the original value of the i-th feature

[0030] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0031] 1. This application is provided with guardrails. When the patient turns over excessively and reaches the position of the guardrail, the guardrail can rotate around the bed body to support the patient. At the same time, after the switch of the rack disengages from the card slot, it can drive the motor to start. The motor can drive the guardrail to rotate until the guardrail rotates to a horizontal state. At this time, the motor can support the guardrail, so that both sides of the bed body are extended, realizing the support work of the guardrail for the patient. The role of the guardrail in assisting the support of the patient and the rotatable guardrail will not squeeze the patient, effectively improving the safety of the turning bed.

[0032] 2. This application is provided with a rack, and the rack can limit the flipping angle of the guardrail. When the guardrail rotates, at this time, the insertion rod releases the limit on the spline sleeve. At this time, the rotation of the motor can drive the spline shaft to rotate, and then drive the first gear to rotate. The first gear drives the rack to descend. If the rack has not been completely removed from the card slot, it means that the patient may just accidentally touch the guardrail at this time, or the patient has returned to the position of the bed body, then the guardrail will not be flipped. Only after the guardrail is squeezed for a specified time, the motor will drive the guardrail to flip at this time, so as to prevent the occurrence of false alarms and other situations caused by misoperations and other reasons, and reduce the waste of resources.

[0033] 3. This application combines deep learning with the turning technology, adopts the idea of turning position recognition, compares the recognized position after turning with the standard position to obtain the best turning control strategy, and then compares the initial position of the patient on the turning board with the position information in the database, and then can select the best turning strategy to control the turning board to flip at an appropriate angle to prevent over-turning or under-turning during turning. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The above and / or additional aspects and advantages of the present invention will become apparent and easy to understand from the following description of the embodiments in conjunction with the accompanying drawings, wherein:

[0035] Figure 1 is the overall three-dimensional view of the present invention;

[0036] Figure 2 is the overall front view of the present invention;

[0037] Figure 3 is the schematic structural view of the spline shaft part of the present invention;

[0038] Figure 4 is Figure 3 The sectional view of the structure of the first gear part in

[0039] Figure 5 is Figure 3 The sectional view of the structure of the second gear part in

[0040] Description of the reference numerals in the drawings:

[0041] 1. Bed body; 2. Flap; 3. Guardrail; 4. Rotating shaft; 5. Jack rod; 6. Flap motor; 7. Traveling wheel; 8. Large gear; 9. Tension spring; 10. Card slot; 11. Switch; 12. Rack; 13. First gear; 14. Second gear; 15. Insert rod; 16. Main shaft; 17. Spline shaft; 18. Spring; 19. Spline sleeve; 20. Volute spring; 21. Motor; 22. Slide groove. Specific embodiments

[0042] The embodiments of the present invention will be described in detail below. The examples of the embodiments are shown in the drawings, where the same or similar reference numerals denote the same or similar elements or elements with the same or similar functions from beginning to end. The embodiments described below by referring to the drawings are exemplary and are only used to explain the present invention, and should not be construed as a limitation to the present invention.

[0043] As Figures 1 - 5 shown, this embodiment proposes an elderly care turning device, including a bed body 1. At both ends of the bed body 1, there are rotatably connected flaps 2 that can be flipped to realize the turning action of the patient. There are two flaps 2 in total. In the middle part of the bed body 1, there is a fixed middle bed board. At both ends of the middle bed board, there are rotatably connected rotating shafts 4. The two flaps 2 are respectively fixedly connected to the two rotating shafts 4. The two flaps 2 can rotate around the rotating shafts 4 respectively, so as to realize the left turning and right turning of the patient.

[0044] The bed body 1 is fixedly connected with a flap motor 6. The flap motor 6 is arranged below the middle bed board. The flap motor 6 is fixedly connected with an output shaft. At both ends of the output shaft, there are fixedly connected symmetrically arranged jack rods 5. The jack rods 5 are in contact with the lower surface of the flap 2 and can slide along the lower surface of the flap 2. The flap motor 6 can drive the jack rods 5 to rotate. The jack rods 5 slide along the surface of the flap 2. At this time, the jack rods 5 can drive the flap 2 to rotate, so as to realize the turning action of the patient.

[0045] At both ends of the bed body 1, there are rotatably connected guardrails 3. The guardrails 3 are connected to the bed body 1 through tension springs 9. One end of the tension spring 9 is connected to the bed body 1, and the other end of the tension spring 9 is connected to the guardrail 3. Under the elastic force of the tension spring 9, the guardrails 3 can maintain a vertical state. At this time, the guardrails 3 are vertically arranged with the bed body 1.

[0046] The guardrail 3 is coaxially and fixedly connected to a large gear 8. When the guardrail 3 rotates, it can drive the large gear 8 to rotate simultaneously. When the large gear 8 rotates, it can drive the guardrail 3 to rotate. The large gear 8 is provided with a card slot 10. The bed body 1 is slidably connected with a rack 12 that can slide into the card slot 10. One end of the card slot 10 away from the flap 2 is provided with an inclined surface, and this inclined surface enables the guardrail 3 to rotate outward within a specified angle.

[0047] The bed body 1 is rotatably connected with a spline shaft 17. The spline shaft 17 is connected to the bed body 1 through a torsion spring 20. One end of the torsion spring 20 is connected to the spline shaft 17, and the other end of the torsion spring 20 is connected to the bed body 1. The torsion spring 20 can continuously provide a large restoring force in a narrow space, so as to realize the reset work of the spline shaft 17 after rotation, and thus facilitate the overall reset of the device.

[0048] The spline shaft 17 is fixedly sleeved with a first gear 13 and a second gear 14. The second gear 14 can mesh with the large gear 8, and the first gear 13 can mesh with the rack 12. When the spline shaft 17 rotates, it can drive the first gear 13 and the second gear 14 to rotate simultaneously. The second gear 14 can drive the large gear 8 to rotate, and then drive the guardrail 3 to rotate outward. The first gear 13 can drive the rack 12 to move downward, so as to drive the rack 12 to be able to disengage from the card slot 10 downward.

[0049] The bed body 1 is connected with a motor 21. The output shaft of the motor 21 is fixedly connected with a main shaft 16. The main shaft 16 is slidably connected with a spline sleeve 19. The main shaft 16 is also in a spline form, so the spline sleeve 19 can slide along the surface of the main shaft 16. The spline sleeve 19 is connected to the main shaft 16 through a spring 18. The spring 18 is sleeved on the main shaft 16. The right end of the spring 18 is connected to the main shaft 17, and the left end of the spring 18 is connected to the spline sleeve 19.

[0050] The elastic force of the spring 18 can make the spline sleeve 19 slide onto the spline shaft 17. In order to facilitate the spline sleeve 19 to slide onto the spline shaft 17, one end of the spline sleeve 19 and the spline shaft 17 close to each other has an inclined surface. The setting of this inclined surface facilitates the spline sleeve 19 to slide onto the spline shaft 17 and prevents jamming.

[0051] The guardrail 3 is fixedly connected with a plug rod 15. The plug rod 15 is located on one side of the spline sleeve 19. One end of the plug rod 15 close to the spline sleeve 19 is a conical head. The elastic force of the tension spring 9 is much greater than the elastic force of the spring 18. When the guardrail 3 is in a vertical state, at this time the plug rod 15 just inserts between the spline sleeve 19 and the spline shaft 17. When the plug rod 15 contacts the spline sleeve 19, it can make the spline sleeve 19 disengage from the spline shaft 17. At this time, the spline shaft 17 and the spline sleeve 19 can rotate independently, which is convenient for the reset of the spline shaft 17.

[0052] The first gear 13 and the second gear 14 are both gear wheels with missing teeth. When the motor 21 drives the spline shaft 17 to rotate, the spline shaft 17 will first drive the first gear 13 to engage with the rack 12. As the spline shaft 17 continues to rotate, when the first gear 13 and the rack 12 disengage, at this time the rack 12 slides out of the card slot 10, and then the second gear 14 will engage with the large gear 8. Only at this time will the second gear 14 drive the guardrail 3 to rotate through the large gear 8.

[0053] A switch 11 capable of controlling the operation of the motor 21 is connected to the rack 12, and an alarm switch capable of controlling the motor 21 to stop working is connected to the guardrail 3; when the switch 11 moves out of the card slot 10, the switch 11 controls the motor 21 to start working. When the motor 21 drives the guardrail 3 to rotate to a specified angle, the alarm switch controls the motor 21 to stop working and gives an alarm at the same time. The switch 11, the alarm switch and the motor 21 are connected to a controller.

[0054] The card slot 10 includes an inclined surface on the right side and a notch on the left side. When the guardrail 3 is in a vertical state, there is a certain gap between the inclined surface on the right side and the rack 12 at this time. This gap allows the guardrail 3 to rotate by a certain angle. This inclined surface can finally rotate to a vertical state, and at this time it is arranged parallel to the side surface of the rack 12. And at this time, the switch 11 can rotate to the notch position on the left side, and the card slot 10 no longer presses the switch 11, and the switch 11 can control the motor 21 to start working.

[0055] The alarm switch is arranged on the outside of the guardrail 3. The alarm switch is composed of a proximity switch and an alarm. The bed body 1 is connected with an induction piece that cooperates with the proximity switch. When the guardrail 3 rotates to a certain angle, at this time the proximity switch rotates to one side of the induction piece. After cooperating with the induction piece, it can transmit a corresponding signal to the controller. The controller controls the alarm to give an alarm and controls the motor 21 to stop working at the same time.

[0056] The proximity switch is a position switch that can be operated without mechanical direct contact with moving parts. When the guardrail 3 rotates until the induction surface of the proximity switch cooperates with the induction piece, the switch can be actuated without mechanical contact and without applying any pressure, so as to provide a control instruction to the controller and make the motor 21 stop working.

[0057] The switch 21 is a normally closed push switch or a normally closed push-button switch. Its characteristic is that when the button is pressed, the circuit is disconnected (open circuit), and after the button is released, the circuit automatically resumes closing (energized for work). The working principle of this kind of switch is based on the elastic restoring force of the spring. When the button is pressed, the spring is compressed, the contacts are separated, and the circuit is disconnected; after the button is released, the spring returns to its original state, the contacts are re-closed, and the circuit is connected. Thus, when the switch 21 is in the card slot 10, at this time the switch 21 is pressed, and when the switch 21 moves out of the card slot 10, the switch 21 resets, and the motor 21 is powered on.

[0058] An elderly care turning method, adopting one of the above, includes the following steps:

[0059] Data acquisition: The rotation of the turning board 2 realizes the turning action of the patient. Based on the data acquisition module, the positions of the patient before and after each turn on the bed body 1 are collected. The characteristic information of the position after turning is extracted through the feature extraction module, and the feature extraction module includes several neural network models;

[0060] Data processing: Based on the feature extraction module, randomly combine several neural network models, and then reconstruct the position image of the extracted features through an autoencoder and compare it with the standard turning position;

[0061] When the reconstructed image is most similar to the standard turning position, it is determined that the combination of the current feature extraction network is the optimal combination, and the optimal features extracted by the feature extraction network of the current combination are forward propagated, and the subsequent processing of the features is performed through the control module;

[0062] Model establishment: Record the control strategy of the elderly care turning device corresponding to the optimal combination, enhance the key feature data of the position before turning, optimize the representation of data features by dynamically adjusting weights, and after training the optimized features, obtain a feature library composed of image key descriptors;

[0063] Turning control: Collect the position of the patient before turning and match it with the image key descriptors in the feature library. Based on the K-means combined with the HITS algorithm, the feature closest to the query image in the features is used as the judgment result for output, and the control strategy corresponding to this feature is used to control the elderly care turning device to realize the turning action of the patient.

[0064] The comparison method between the reconstructed image and the standard turning position includes:

[0065] After flattening the multi-dimensional features (usually high-dimensional tensors) of the original image and the reconstructed image into one-dimensional vectors, by converting the multi-dimensional features into one-dimensional vectors, the distance between the two feature vectors can be calculated subsequently. Normalize the feature vectors to ensure that features in different dimensions are on the same scale, thereby improving the accuracy of similarity calculation;

[0066] Then subtract the two one-dimensional vectors, then sum the absolute values, calculate the L1 norm of the two feature vectors, use the L1 norm to calculate the distance between the two vectors, which is used to represent their similarity degree, and measure the distance between the two feature vectors through the L1 norm. The smaller the distance, the more similar the two images are;

[0067] Next, perform two fully connected operations on the obtained distance. Connect the second fully connected layer to a neuron and apply the Sigmoid function to its result. The role of the fully connected layer is to process the distance calculated by the L1 norm through two fully connected layers. The application of the Sigmoid function maps the output of the second fully connected layer to the interval [0, 1] through the Sigmoid function, which is used to represent the similarity degree between two input images;

[0068] The first fully connected layer: Input the distance of the L1 norm into a fully connected layer to learn its feature representation. The second fully connected layer: Input the output of the first fully connected layer into another fully connected layer and finally connect it to a neuron. The Sigmoid function: Apply the Sigmoid function to the output of the second fully connected layer and map the result to the interval [0, 1].

[0069] When the similarity degree between two images reaches the maximum, it means that the features extracted by the combination of the current neural network model are the optimal features. By maximizing the similarity, it is ensured that the features extracted by the model can best represent the similarity between the original image and the reconstructed image.

[0070] And forward propagate the optimal features extracted by the feature extraction network of the current combination, and perform subsequent processing of the features through the control module. The optimal features refer to the feature representation that can maximize the similarity under the current model structure and training strategy. These features are considered to be the features that can best reflect the similarity between the original image and the reconstructed image.

[0071] The method for dynamically adjusting weights includes: Initializing the weight matrix: Assign an initial weight to each data feature, and these weights will be used to adjust the importance of the feature in subsequent analysis. The weight matrix is denoted as W, and the formula is as follows:

[0072]

[0073] In the formula, w i is the initial weight of the i-th feature, and n is the number of features;

[0074] Calculating the weighted eigenvalue: Use the weight matrix to weight the feature data X, multiply each eigenvalue xi by its corresponding weight w i to calculate the weighted value of each feature, so as to obtain a comprehensive index Z. This index reflects the importance of the feature after weighted processing. The formula is as follows:

[0075]

[0076] In the formula, Z is the weighted feature vector, representing the comprehensive index of the eigenvalues after weight processing, x i$x_i$ is the original value of the $i$-th feature, and the weight matrix $W$ is a column vector, where each element represents the weight of the corresponding feature. These weights can be any positive numbers and are used to adjust the importance of the features in the analysis.

[0077] The turning-over method of the elderly care turning-over device includes the following steps:

[0078] S1: The turning-over plate motor 6 drives the ejector rod 5 to rotate, and the ejector rod 5 drives the turning-over plate 2 to rotate, thereby realizing the turning-over action of the patient. When the turning-over action of the patient is too large and the patient is overturned on the guardrail 3, the guardrail 3 rotates, causing the switch 11 to move out of the card slot 10. The switch 11 drives the motor 21 to work. The guardrail 3 drives the insertion rod 15 to rotate to one side of the spline sleeve 19, and the spring 18 causes the spline sleeve 19 to slide onto the spline shaft 17;

[0079] S2: The motor 21 drives the spline shaft 17 to rotate. The spline shaft 17 drives the first gear 13 and the second gear 14. The first gear 13 drives the rack 12 to descend. When the rack 12 completely disengages from the card slot 10, it indicates that the patient has pressed the guardrail 3 for the specified time at this time. At this time, the second gear 14 meshes with the large gear 8, and the second gear 14 drives the guardrail 3 to rotate through the large gear 8;

[0080] S3: The guardrail 3 is in an inclined state to support the patient. When the motor 21 drives the guardrail 3 to rotate to the specified angle, the alarm switch controls the motor 21 to stop working and alarms at the same time;

[0081] If the patient no longer presses the guardrail 3 when the rack 12 disengages from the card slot 10, the tension spring 9 pulls the guardrail 3 to reset at this time. The insertion rod 15 slides between the spline sleeve 19 and the spline shaft 17, and the spline sleeve 19 disengages from the spline shaft 17. At this time, the spiral spring 20 can drive the spline shaft 17 to reset. The spline shaft 17 drives the first gear 13 to reverse, and the first gear 13 drives the rack 12 to reset, causing the rack 12 to slide back into the card slot 10 again. After the switch 11 slides into the card slot 10, the switch 11 controls the motor 21 to stop working

[0082] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the claims and their equivalents.

Claims

1. A turning device for elderly care, comprising a bed body (1), characterized in that: The two ends of the bed body (1) are rotatably connected to a flap (2) capable of being turned over to realize a patient turning over action; the two ends of the bed body (1) are rotatably connected to a guardrail (3), and the guardrail (3) is connected to the bed body (1) via a tension spring (9); the guardrail (3) is coaxially fixedly connected to a large gear (8), and the large gear (8) is provided with a slot (10); the bed body (1) is slidably connected to a rack (12) capable of sliding into the slot (10); the slot (10) is provided with an inclined surface at one end away from the flap (2), and the inclined surface enables the guardrail (3) to rotate outward within a specified angle.

2. The turning device for elderly care according to claim 1, characterized in that: The bed body (1) is rotatably connected to a spline shaft (17), the spline shaft (17) being connected to the bed body (1) via a coil spring (20), the spline shaft (17) being fixedly sleeved with a first gear (13) and a second gear (14), the second gear (14) being meshed with the large gear (8), and the first gear (13) being meshed with the rack (12).

3. The turning device for elderly care according to claim 2, characterized in that: The bed body (1) is connected to a motor (21), the motor (21) is connected to a main shaft (16), the main shaft (16) is slidably connected to a spline sleeve (19), and the spline sleeve (19) is connected to the main shaft (16) via a spring (18).

4. The turning device for elderly care according to claim 3, characterized in that: The elastic force of the spring (18) can make the spline sleeve (19) slide onto the spline shaft (17); the guardrail (3) is fixedly connected with an insertion rod (15), and the insertion rod (15) is located on one side of the spline sleeve (19); the elastic force of the tension spring (9) is greater than the elastic force of the spring (18); when the insertion rod (15) and the spline sleeve (19) are in contact, the spline sleeve (19) can be separated from the spline shaft (17).

5. The turning device for elderly care according to claim 2, characterized in that: The first gear (13) and the second gear (14) are both toothless gears. When the first gear (13) and the rack (12) are disconnected from the meshing, the rack (12) slides out of the slot (10), and the second gear (14) will mesh with the large gear (8).

6. The turning device for elderly care according to claim 3, characterized in that: The rack (12) is connected to a switch (11) capable of controlling the motor (21) to work, and the guardrail (3) is connected to an alarm switch capable of controlling the motor (21) to stop working; when the switch (11) is moved out of the slot (10), the switch (11) controls the motor (21) to start working, and when the motor (21) drives the guardrail (3) to rotate to a specified angle, the alarm switch controls the motor (21) to stop working and simultaneously issues an alarm.

7. The turning device for elderly care according to claim 1, characterized in that: The bed body (1) is connected to a flap motor (6), the flap motor (6) is fixedly connected to an output shaft, the output shaft is fixedly connected to a symmetrically arranged push rod (5), the push rod (5) is in contact with the lower surface of the flap (2) and can slide along the lower surface of the flap (2).

8. A method for turning over the elderly in nursing care, characterized in that: The turning device for elderly care according to any one of claims 1 to 7 comprises the following steps: Data collection: The rotation of the flap (2) realizes the patient's turning over action. Based on the data collection module, the patient's position before and after turning over on the bed (1) is collected each time, and the feature information of the position after turning over is extracted by the feature extraction module. The feature extraction module includes several neural network models; Data processing: Based on the feature extraction module, several neural network models are randomly combined, and then the extracted features are reconstructed into position images through the automatic encoder and compared with the standard turning position; When the reconstructed image is most similar to the standard turning position, the combination of the current feature extraction network is determined to be the optimal combination, and the optimal features extracted by the feature extraction network of the current combination are forward propagated, and the subsequent feature processing is performed through the control module; Model establishment: record the control strategy of the elderly care turning device corresponding to the optimal combination, enhance the key feature data of the position before turning, optimize the representation of data features by dynamically adjusting the weights, and obtain the feature library composed of key image descriptors after training the optimized features; Turning control: The patient's position before turning over is collected and matched with the key image descriptors in the feature library. Based on the K-means combined with the HITS algorithm, the feature closest to the query image is output as the judgment result. The control strategy corresponding to the feature is used to control the elderly care turning device to realize the patient's turning action.

9. The turning device for elderly care according to claim 8, characterized in that: Methods for comparing the reconstructed image to the standard turning position include: After flattening the multi-dimensional features of the original image and the reconstructed image into a one-dimensional vector, the feature vector is normalized, and then the two one-dimensional vectors are subtracted, and then the absolute values ​​are summed to obtain the L1 norm of the two feature vectors. The L1 norm is used to calculate the distance between the two vectors to represent the similarity between the two. Then, the obtained distance is fully connected twice, the second fully connected layer is connected to a neuron, and the result is subjected to the Sigmoid function. The function of the fully connected layer is to process the distance calculated by the L1 norm through two fully connected layers. The Sigmoid function is applied to map the output of the second fully connected layer to the [0,1] interval through the Sigmoid function to represent the similarity between the two input images. When the similarity between two images reaches the maximum, it means that the features extracted by the current combination of neural network models are the optimal features, and the optimal features extracted by the feature extraction network of the current combination are forward propagated, and the subsequent feature processing is performed through the control module.

10. The turning device for elderly care according to claim 8, characterized in that: The dynamic weight adjustment method includes: Initializing the weight matrix: assigning an initial weight to each data feature. The weight matrix is ​​represented by W, and the formula is as follows: In the formula, w i is the initial weight of the i-th feature, and n is the number of features; Calculate weighted eigenvalues: Use the weight matrix to weight the feature data X, and each eigenvalue xi is associated with its corresponding weight w i Multiplying, the formula is as follows: In the formula, Z is the weighted eigenvector, which represents the comprehensive index of the eigenvalue after weight processing, x i is the original value of the i-th feature.

Images