A control system and method for an intelligent electric standing bed

Abstract

The present invention relates to the technical field of electric standing beds, and specifically to a control system and method for an intelligent electric standing bed. The control system includes a manual controller, a human-computer interaction controller and an intelligent controller. The manual controller is installed on the bed surface and is electrically connected to the drive motor. The human-computer interaction controller is used to interact with the user, collect the user's intention to stand up, generate control instructions, and then control the start and stop state of the drive motor. The intelligent controller uses an actor-critic-based prediction model to predict the optimal angle and optimal frequency between the bed surface and the bed frame, and converts the optimal angle and optimal frequency into control data of the motor, generates a control strategy, and then controls the start and stop state of the drive motor. The present invention can select different controllers according to the customer's health status. The user can autonomously control the intelligent electric standing bed according to his or her own comfort or the advice of medical staff, or can perform adaptive control through the intelligent controller, which has stronger adaptability.

Description

Technical Field

[0001] The present invention relates to the technical field of electric standing beds, and in particular to a control system of an intelligent electric standing bed and a control method of the intelligent electric standing bed. Background Art

[0002] An intelligent electric standing bed is a bed that combines electric lifting function with an intelligent control system, designed to help users (especially the elderly or those with limited mobility) stand up, lie down and adjust their posture more conveniently. It is usually equipped with components such as electric motors, sensors, control panels or mobile phone applications to achieve height and angle adjustment.

[0003] The main functions of the smart electric standing bed include: Users can easily adjust the height of the bed through the remote control or mobile phone application. Support multiple posture adjustments, such as sitting up, lying down, tilting, etc., to provide better comfort. Intelligent monitoring: Some high-end models are equipped with sensors that can monitor the user's health data (such as heart rate, sleep quality, etc.). Safety features: including anti-pinch design, emergency stop button, etc. to ensure user safety.

[0004] The control system of intelligent electric standing beds has been widely used in recent years, especially in the fields of medical care and elderly care. However, there are still some defects and shortcomings in the existing technology, mainly including the following aspects: 1. Complexity of the control system: The control system interface of many intelligent electric standing beds is complex and the operation is not intuitive enough, especially for elderly users or people with physical disabilities, it may be difficult to understand and use. In addition, the existing system may only support a single control method (such as a remote control) and lacks the integration of multiple control methods such as mobile phone applications and voice control. 2. Insufficient sensors and feedback mechanisms: The existing systems often lack the ability to monitor the user's status in real time and cannot adjust the height and angle of the bed in time to meet the user's needs. 3. Insufficient adaptability and personalization: The existing systems often cannot be personalized according to the specific needs of the user, resulting in a poor user experience, and for users of different body shapes and health conditions, the existing systems are not adaptable enough and cannot provide the best support. Summary of the invention

[0005] Based on this, it is necessary to propose a control system and method for an intelligent electric standing bed in order to address the problems that the existing intelligent electric standing beds have complex and single control systems and cannot be adaptively adjusted according to user needs.

[0006] The present invention is achieved through the following technical solutions: a control system of an intelligent electric standing bed, comprising a manual controller, a human-computer interaction controller and an intelligent controller.

[0007] The manual controller is installed on the bed surface and is electrically connected to the drive motor. The human-computer interaction controller is used to interact with the user, collect the user's intention to stand up, generate control instructions, and then control the start and stop state of the drive motor. The intelligent controller uses the actor-critic prediction model to predict the optimal angle and optimal frequency between the bed surface and the bed frame, and converts the optimal angle and optimal frequency into control data of the motor, generates a control strategy, and then controls the start and stop state of the drive motor.

[0008] The above control system can select different controllers according to the customer's health status, including manual controller, human-computer interaction controller, and intelligent controller. Users can achieve autonomous control through manual controller and human-computer interaction controller, and control the intelligent electric standing bed according to their own comfort or the advice of medical staff, which has stronger adaptability.

[0009] Furthermore, the human-computer interaction controller includes a voice interaction controller and a brain wave interaction controller. The voice interaction controller is used to identify the user's voice information and generate control instructions. The brain wave interaction controller is used to identify the user's brain wave information and generate control instructions.

[0010] Furthermore, the brainwave interaction controller includes a brainwave acquisition module, a brainwave identification module and a control instruction generation module. The brainwave acquisition module is used to collect the user's brainwave signals. The brainwave identification module is used to identify the characteristics of the brainwaves and output the corresponding intention category. The control instruction generation module is used to generate corresponding control instructions according to the intention category.

[0011] Furthermore, the recognition process of the EEG recognition module is as follows: Preprocess the EEG signal to remove interference. Divide the EEG signal into multiple continuous and non-overlapping time windows according to the preset period. Extract the time-frequency features of each time window. Build and train the CNN recognition model, input the time-frequency features into the CNN recognition model, and output the corresponding intent category.

[0012] Furthermore, the process of constructing the actor-critic based prediction model by the intelligent controller is as follows: Define the environment parameters. Establish the DNN optimized actor network. Establish the DNN optimized critic network. Adjust the network weights using the gradient descent method. Set the loss function. Update the parameters using the gradient descent method.

[0013] Furthermore, the intelligent controller trains the actor-critic prediction model as follows: At each time step, select an action based on the current state. Perform the action and observe the new state and reward. Use the critic network to evaluate the value of the new state. Update the actor network's policy based on the true reward and the critic network's evaluation value. Increase exploration and prevent overfitting through experience replay.

[0014] The present invention also provides a control method for an intelligent electric standing bed, comprising the following steps:

[0015] Choose autonomous control or intelligent control according to the user's rehabilitation status. Autonomous control includes manual control, voice control, and brainwave control. Among them, the method of brainwave control is as follows: Collect the user's brainwave signal. Preprocess the brainwave signal. Divide the brainwave signal into multiple continuous and non-overlapping time windows according to the preset period. Extract the time-frequency features of each time window. Build and train a CNN recognition model, input the time-frequency features into the CNN recognition model, and output the corresponding intent category. Generate control instructions according to the intent category.

[0016] The specific method of intelligent control is as follows: Construct a prediction model based on actor-critic. Use DNN network to optimize the prediction model. Train with the prediction model and update the model parameters. Input the user's rehabilitation data into the prediction model and output the optimal angle and optimal frequency. Calculate the start-stop frequency, output power, and driving direction of the drive motor based on the optimal angle and optimal frequency, and generate a control strategy.

[0017] Compared with the prior art, the present invention has the following beneficial effects:

[0018] The present invention can select different controllers according to the health status of the customer, including manual controllers, human-computer interaction controllers, and intelligent controllers. Users can achieve autonomous control through manual controllers and human-computer interaction controllers, and control the intelligent electric standing bed according to their own comfort or the advice of medical staff, which has stronger adaptability. In addition, users can also choose intelligent controllers for adaptive control. The intelligent controller uses an optimized actor-critic prediction model, which can not only adaptively adjust the angle and frequency of the electric standing bed according to the rehabilitation status of each user, but also can make real-time corrections to the angle and adjustment frequency of the electric standing bed according to the health data fed back by the user, thereby achieving precise control of the electric standing bed, effectively improving the user's rehabilitation efficiency, and improving the safety and comfort of the patient's rehabilitation training. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a schematic diagram of the structure of the intelligent electric standing bed in Example 1 of the present invention;

[0020] Figure 2 A schematic diagram of the framework structure of a control system of an intelligent electric standing bed according to an embodiment of the present invention;

[0021] Figure 3 This is a flow chart of a control method for an intelligent electric standing bed according to Embodiment 2 of the present invention. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] It should be noted that when a component is referred to as being "mounted on" another component, it may be directly on the other component or there may be a central component. When a component is considered to be "set on" another component, it may be directly set on the other component or there may be a central component at the same time. When a component is considered to be "fixed to" another component, it may be directly fixed on the other component or there may be a central component at the same time.

[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention belongs. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "or / and" used herein includes any and all combinations of one or more of the related listed items.

[0025] Example 1: Please refer to Figure 1 , this embodiment provides a control system for an intelligent electric standing bed. Specifically, the electric standing bed includes a bed frame, a bed surface, a driving motor, a push rod, a foot pedal, a strap, an armrest, an angle sensor, a caster and a control system. The bed frame is rotatably connected to one side of the bed surface, and the bed surface can be flipped relative to the bed frame. One end of the push rod is rotatably connected to the middle part of the bed surface, and the other end is fixedly connected to the output end of the driving motor. The driving motor is rotatably mounted on the bed frame to drive the push rod to move. The driving motor is a linear driving motor. During operation, according to the input control signal, it forms an electric telescopic component with the push rod to adjust the angle between the bed frame and the bed surface. The angle sensor is installed on the bed surface, and the driving motor can cooperate with the angle sensor to adjust the bed surface according to the input angle, so that the angle between the bed surface and the bed frame reaches the preset input angle. The foot pedal is installed on one side of the bed surface to provide support for the user to stand when the bed surface rises. The armrest table is installed in the middle of the bed surface, and multiple straps are set at different positions on the bed surface. After the user puts on the straps, it can effectively prevent falling. At the same time, the armrest table can provide support for the user after the bed surface is raised, preventing the user from leaning forward too much and causing physical discomfort. The casters are installed at the four ends of the bed frame, which can drive the bed frame to move, thereby making the electric standing bed move as a whole.

[0026] See also Figure 2The control system is used to control the start and stop state of the driving motor of the electric standing bed to adjust the angle between the bed frame and the bed surface of the electric standing bed. The control system includes: a manual controller, a human-computer interaction controller, and an intelligent controller.

[0027] The manual controller is installed on the bed surface and is electrically connected to the drive motor. The manual controller includes an up gear, a down gear and a pause gear. In this embodiment, the manual controller adopts a button-type controller, including an up key and a down key. The user or guardian starts the drive motor by pressing the up key or the down key to drive the bed surface up or down. When not in operation, the button-type controller is in the pause gear, and the bed frame and the bed surface are relatively fixed.

[0028] The manual controller is suitable for users with normal arm function. Users can manually adjust the angle between the bed surface and the bed frame according to their own feelings or the doctor's advice to assist rehabilitation exercises and improve rehabilitation efficiency. Especially for some elderly people, it is difficult to rely on their own strength to sit up, but they can control it by pressing a switch, such as using the manual controller that comes with the electric standing bed or other remote controls.

[0029] The human-computer interaction controller is used to interact with the user, collect the user's intention to stand up, generate control instructions, and then control the start and stop state of the drive motor. In this embodiment, the human-computer interaction controller includes a voice interaction controller and a brain wave interaction controller. Among them, the voice interaction controller can use AI inquiry method or voice recognition module to collect the user's voice control instructions. For example, when using AI inquiry, the user actively wakes up AI and issues control instructions to AI. AI interacts with the user to confirm the control information, and then generates control instructions for the drive motor, so that the drive motor starts to adjust the angle between the bed surface and the bed frame. The voice interaction controller is suitable for users with normal voice function. Users can adjust the angle between the bed surface and the bed frame by voice calling according to their own needs or according to the advice of medical staff, thereby assisting rehabilitation exercises and improving rehabilitation efficiency. In some embodiments, the voice interaction controller can also be connected to the medical end, so that users can call medical staff in time and conduct simple voice communication, which is convenient for rapid handling of emergency events, such as calling medical staff in time for processing when the control system of the electric standing bed fails.

[0030] The EEG interactive controller includes an EEG acquisition module, an EEG identification module and a control instruction generation module. The EEG acquisition module uses an EEG cap. After the user wears the EEG cap, the user's EEG signals are collected through multiple electrodes in the EEG cap.

[0031] The brainwave identification module is used to output the user's intention category according to the user's brainwave signal, including rising, falling, stopping, and other four categories.

[0032] Specifically, the identification process of the brain wave identification module is as follows:

[0033] First, the EEG signal is preprocessed to remove interference. EEG signals are weak and need to be processed by specific amplifiers and filters, and the filtered EEG is standardized to extract useful signals.

[0034] Secondly, the EEG signal is divided into multiple continuous and non-overlapping time windows according to a preset period.

[0035] The processed EEG signal is analyzed to extract features that can represent the user's intention. In this embodiment, the frequency spectrum features of each window are extracted by fast Fourier transform, and the time-frequency features of each window are extracted by wavelet transform.

[0036] Subsequently, a CNN recognition model is constructed and trained using EEG big data, and the user's EEG signal label is then identified using the trained recognition model.

[0037] The CNN model includes input layer, convolution layer, activation function, pooling layer and fully connected layer. The time-frequency features of the EEG are input into the input layer, the convolution layer is used to extract local features, the activation function uses the ReLU activation function, the pooling layer reduces the dimension by downsampling, and finally the fully connected layer maps the features to the output intent categories, such as rise, fall, stop, and others.

[0038] The control instruction generation module can generate corresponding control instructions according to the output category to control the drive motor to adjust the angle between the bed surface and the bed frame. For example, after a short period of training, the user actively thinks about the downward movement control instruction, and the brain wave interactive controller recognizes the user's intention and controls the drive motor to start and adjust the bed surface downward. When the user perceives that the angle of the bed surface is appropriate, he actively thinks about the pause instruction, and the brain wave interactive controller recognizes the user's intention and controls the drive motor to pause, and continues to recognize the user's intention. If the user's intention is to stop or other, the control of the electric standing bed is completed.

[0039] The brainwave interactive controller is suitable for users with hand dysfunction and speech dysfunction. If the user has good thinking ability and can cooperate with medical staff in rehabilitation training, he or she can achieve autonomous control of the electric standing bed by collecting brainwave signals.

[0040] In practical applications, a user-friendly interface can be developed first, allowing users to control the electric bed by thinking, including thinking specific ideas or performing specific tasks, and then converting the brain wave recognition results into actual motor control signals to achieve precise control of the electric bed. In order to ensure the safety and reliability of the system, instant feedback can be provided during the control process to ensure the safety and comfort of the user.

[0041] The intelligent controller is used to build an actor-critic based prediction model, which is trained with big data. It then outputs the optimal angle and frequency between the bed surface and the bed frame according to the user's rehabilitation data, and converts the optimal angle and frequency into control data of the motor to achieve automatic control of the motor.

[0042] Among them, the construction and prediction process of the prediction model are as follows:

[0043] First, define the environment parameters, including:

[0044] State: user's current body posture u, angle w between the bed surface and the bed frame.

[0045] Action: Control the speed v and frequency f of the electric bed.

[0046] Reward: Keep the user moving from one state to another in a smooth and comfortable manner.

[0047] Secondly, establish an actor network optimized by DNN (deep neural network). Strategy learning of actor network: The actor network is responsible for selecting the optimal action according to the current state. Its output is the speed v and frequency f of the electric bed. The strategy can be expressed as:

[0048] π(a∣s)=softmax{A(s,a)}

[0049] Among them, a is the action, s is the state, and A is the scoring function based on CNN.

[0050] Building a critic network:

[0051] Value function learning: The critic network evaluates the value of the state-action pair, that is, predicts the expected value of the environment reward. DNN (deep neural network) is used to optimize the critic network, and the formula is expressed as:

[0052]

[0053] Where V(s) is the predicted value function, and Q(s, a) is the state-action value function, which represents the expected return after executing action a in state s.

[0054] The network weights are adjusted using the gradient descent method to optimize the strategy and value function. The gradient strategy can be expressed as:

[0055]

[0056]

[0057] Where J(θ) is the expected return of the strategy, θ is the parameter of the actor model, and Gt (a) is the reward gradient of action a, R t is the actual return, V(S t ) is the predicted value of the critic network.

[0058] The loss function of the critic network is set as:

[0059]

[0060] The parameters of the critic model are updated by the gradient descent method, and the gradient expression is:

[0061]

[0062] Where ω is the parameter of the critic model.

[0063] The training process of the actor-critic prediction model is as follows: at each time step, select an action based on the current state (output by the actor network); perform the action and observe the new state and reward (provided by the environment); use the critic network to evaluate the value of the new state; update the actor network's strategy based on the real reward and the evaluation value of the critic network; increase exploration and prevent overfitting through experience replay. In this embodiment, L1 or L2 regularization can be used to prevent overfitting and improve the stability of learning.

[0064] By outputting the optimal angle and frequency based on the actor-critic prediction model, adaptive control of the electric standing bed can be achieved, and the control strategy can be continuously optimized through machine learning to improve safety and comfort.

[0065] The control data of the motor, including the start and stop frequency, output power, and driving direction, is calculated based on the optimal angle and optimal frequency, and then the corresponding control instructions are generated to automatically adjust to the optimal angle within the expected time period, thereby improving the user's rehabilitation efficiency.

[0066] In summary, the control system of the intelligent electric standing bed of this embodiment can select different controllers according to the health status of the customer, including manual controller, human-computer interaction controller, and intelligent controller. The user can realize autonomous control through the manual controller and the human-computer interaction controller, and control the intelligent electric standing bed according to his or her own comfort or the advice of medical staff, which has stronger adaptability. In addition, the user can also choose the intelligent controller for adaptive control. The intelligent controller adopts the optimized actor-critic prediction model, which can not only adaptively adjust the angle and frequency of the electric standing bed according to the rehabilitation status of each user, but also can make real-time corrections to the angle and adjustment frequency of the electric standing bed according to the health data fed back by the user, thereby realizing precise control of the electric standing bed, effectively improving the rehabilitation efficiency of the user, and improving the safety and comfort of the patient's rehabilitation training.

[0067] Example 2: Please refer to Figure 3 This embodiment provides a control method for an intelligent electric standing bed, which can be applied to the control system of the intelligent electric standing bed in Embodiment 1, and includes the following steps:

[0068] Choose autonomous control or intelligent control according to the user's recovery status.

[0069] Autonomous control includes manual control, voice control, and brain wave control. Among them, the method of manual control is as follows: the user uses a manual controller installed on the bed surface or a remote controller to control the electric standing bed, and the control content includes rising, falling, and stopping.

[0070] The voice control method is as follows: the user actively wakes up the voice controller and issues a voice control command to the voice controller, and the voice controller extracts keywords and generates corresponding control instructions.

[0071] The specific method of brainwave control is as follows: Collect the user's brainwave signals. Preprocess the brainwave signals. Divide the brainwave signals into multiple continuous and non-overlapping time windows according to a preset period. Extract the time-frequency features of each time window. Build and train a CNN recognition model, input the time-frequency features into the CNN recognition model, and output the corresponding intent category. Generate control instructions based on the intent category.

[0072] The specific method of intelligent control is as follows: Construct a prediction model based on actor-critic. Use DNN network to optimize the prediction model. Train with the prediction model and update the model parameters. Input the user's rehabilitation data into the prediction model and output the optimal angle and optimal frequency. Calculate the start-stop frequency, output power, and driving direction of the drive motor based on the optimal angle and optimal frequency, and generate a control strategy.

[0073] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0074] The above-mentioned embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A control system for an intelligent electric bed, used to control the start and stop state of the driving motor of the electric bed to adjust the angle between the bed frame and the bed surface of the electric bed; characterized in that: The control system comprises: A manual controller, mounted on the bed surface and electrically connected to the drive motor; A human-computer interaction controller, used to interact with the user, collect the user's intention to stand up, generate control instructions, and then control the start and stop state of the drive motor; An intelligent controller, which uses an actor-critic prediction model to predict an optimal angle and an optimal frequency between a bed surface and a bed frame, and converts the optimal angle and the optimal frequency into control data of a motor, generates a control strategy, and further controls the start and stop state of the drive motor; The process of constructing the actor-critic prediction model by the intelligent controller is as follows: Define environment parameters; including: State: user's current body posture u, angle w between the bed surface and the bed frame; Action: Control the speed v and frequency f of the electric standing bed; Reward: Keep the user moving from one state to another in a smooth and comfortable manner. A DNN-optimized actor network is established; the output of the actor network is the speed v and frequency f of the electric standing bed; the strategy π is expressed as: π(a∣s)=softmax{A(s,a)} Among them, a is the action, s is the state, and A is the scoring function based on CNN; Establish a DNN optimized critic network; the formula is expressed as: Where V(s) is the predicted value function, Q(s, a) is the state-action value function, which represents the expected return after executing action a in state s; The network weights are adjusted using the gradient descent method; the gradient strategy is expressed as: In the formula, is the gradient operator, J(θ) is the expected return of the strategy, θ is the parameter of the actor model, G t (a) is the reward gradient of action a, R t is the actual return, V(S t ) is the predicted value of the critic network; Set the loss function to: In the formula, L is the gradient and T is the number of gradients; The gradient descent method is used to update the parameters; the gradient expression is: Where ω is the parameter of the critic model; The process of training the actor-critic based prediction model by the intelligent controller is as follows: At each time step, an action is selected based on the current state; Perform actions and observe new states and rewards; Use a critic network to assess the value of new states; Update the actor network’s policy based on the true reward and the critic network’s evaluation value; Increase exploration and prevent overfitting through experience replay.

2. The control system of the intelligent electric standing bed according to claim 1, characterized in that: The human-computer interaction controller includes a voice interaction controller and a brain wave interaction controller; the voice interaction controller is used to identify the user's voice information and generate control instructions; the brain wave interaction controller is used to identify the user's brain wave information and generate control instructions.

3. The control system of the intelligent electric standing bed according to claim 2 is characterized in that: The brainwave interactive controller includes a brainwave acquisition module, a brainwave identification module and a control instruction generation module; the brainwave acquisition module is used to collect the user's brainwave signals; the brainwave identification module is used to identify the characteristics of the brainwaves and output the corresponding intention category; the control instruction generation module is used to generate corresponding control instructions according to the intention category.

4. The control system of the intelligent electric standing bed according to claim 3 is characterized in that: The recognition process of the brain wave identification module is as follows: Pre-process brain wave signals to remove interference; Divide the brain wave signal into multiple continuous and non-overlapping time windows according to a preset period; Extract the time-frequency features of each time window; Construct and train a CNN recognition model, input the time-frequency features into the CNN recognition model, and output the corresponding intent category.

5. A control method for an intelligent electric standing bed, which is applied to a control system of an intelligent electric standing bed as claimed in any one of claims 1 to 4, characterized in that: The steps include: Choose autonomous control or intelligent control according to the user's recovery status; Autonomous control includes manual control, voice control, and brain wave control; The method of brainwave control is as follows: collecting the brainwave signal of the user; preprocessing the brainwave signal; dividing the brainwave signal into a plurality of continuous and non-overlapping time windows according to a preset period; extracting the time-frequency features of each time window; constructing and training a CNN recognition model, inputting the time-frequency features into the CNN recognition model, and outputting the corresponding intention category; generating a control instruction according to the intention category; The specific method of intelligent control is as follows: construct a prediction model based on actor-critic; use a DNN network to optimize the prediction model; train with the prediction model and update model parameters; input the user's rehabilitation data into the prediction model, and output the optimal angle and optimal frequency; calculate the start and stop frequency, output power, and driving direction of the drive motor according to the optimal angle and the optimal frequency, and generate a control strategy.

Images