A bed-mounted integrated intelligent rehabilitation training system and a rehabilitation training method thereof

Abstract

This invention provides a bed-mounted integrated intelligent rehabilitation training system and its rehabilitation training method, belonging to the field of rehabilitation training technology. It includes an integrated bed-controlled support platform with an electrical interface, a lower limb servo stress loading mechanism mounted on the platform, upper limb multi-dimensional feedback training mechanisms on both sides of the platform, a multimodal state sensing network, and a central intelligent processing terminal. The lower limb mechanism can apply axial mechanical stress to the lower limbs and drive movement training, while the upper limb mechanism provides upper limb joint mobility and resistance training. The sensing network collects training load, joint angle, and movement trajectory parameters. The central intelligent processing terminal communicates and manages each mechanism and executes training prescriptions, performs safety monitoring, and provides feedback control. The beneficial effects are that bedridden patients can conduct coordinated upper and lower limb rehabilitation training without leaving the bed, breaking the limitations of fragmented training; multimodal real-time sensing combined with terminal closed-loop control improves the accuracy of training load and safety response capability, ensuring safe and efficient early active rehabilitation for bedridden patients.

Description

Technical Field

[0001] This invention relates to the field of rehabilitation training technology, and in particular to a bed-mounted integrated intelligent rehabilitation training system and its rehabilitation training method. Background Technology

[0002] With the accelerating aging of the population, the number of patients requiring prolonged bed rest due to post-fracture surgery, spinal cord injury, severe cardiovascular and cerebrovascular diseases, and other chronic wasting diseases continues to expand. Prolonged immobilization can rapidly induce disuse syndrome, leading to multi-system functional decline, including muscle atrophy, joint contractures, bone mineral loss, and reduced cardiopulmonary compensatory capacity, significantly increasing the probability of complications and the difficulty of rehabilitation. Clinical research and practice have fully demonstrated that initiating systematic active rehabilitation intervention as early as possible after the patient's vital signs stabilize is the core strategy for preventing or delaying the progression of disuse syndrome and rebuilding limb function.

[0003] However, training devices currently widely deployed in clinical and rehabilitation settings are generally difficult to match the actual needs of bedridden patients, which is particularly evident in the following key aspects.

[0004] First, the functions are limited, and upper and lower limb training is fragmented. Existing upper limb rehabilitation equipment is usually a stand-alone upright or desktop structure, focusing on the range of motion and muscle strength training of the shoulder, elbow, and wrist joints; lower limb rehabilitation equipment is mostly a stand-alone continuous passive joint movement machine or a bedside lower limb trainer. The two cannot be deployed collaboratively on the same platform, and doctors cannot implement a holistic training program that links the upper and lower limbs simultaneously within the same time window. For bedridden patients, this means that each functional rehabilitation intervention is fragmented, and the motor control, coordination training, and overall physical recovery of the upper and lower limbs cannot be organically connected.

[0005] Secondly, the training load control precision is insufficient, and sensing methods and safety protection mechanisms are lacking. Most existing lower limb axial loading devices use simple weights, springs, or pneumatic dampers to provide resistance or thrust, making it difficult to accurately quantify and dynamically adjust the loading force. Furthermore, these devices often only have limited mechanical sensors or rely solely on estimated current for force prediction, lacking multimodal real-time sensing of the patient's joint range of motion, plantar pressure distribution, and actual force. More critically, in the event of overload, spasticity, or abnormal force loading due to operational errors during active training, existing devices lack hardware safety mechanisms to quickly cut off power transmission at the physical level, exposing patients, especially those with fragile bone structures in the early postoperative period, to a high risk of overload injury.

[0006] Third, most equipment is not integrated with hospital beds, making clinical deployment and usage procedures cumbersome. Conventional rehabilitation training equipment is often placed as a large, independent device in the rehabilitation center. Bedridden patients needing to use it must be moved from their wards to the equipment location by transferring beds. This process not only consumes a significant amount of physical and time resources for medical staff, but the movement itself may also strain the patient's immobilized fracture site or postoperative wound, increasing pain and the risk of complications. Furthermore, due to limitations in manpower and transfer procedures, the frequency and duration of rehabilitation training for patients are often difficult to guarantee, preventing true early bedside intervention and continuous, repeated training, thus underutilizing the golden window of rehabilitation.

[0007] As can be seen from the above analysis, there is currently a lack of intelligent rehabilitation systems on the market that can be deeply integrated with hospital beds, have integrated upper and lower limb training capabilities, have precise and controllable loading force, and have embedded hardware-level safety protection mechanisms. This results in early active rehabilitation for bedridden patients facing bottlenecks in terms of equipment, safety, and technical means. Summary of the Invention

[0008] To address the problems existing in the prior art, the present invention provides a bed-mounted integrated intelligent rehabilitation training system, comprising: An integrated bed control support platform, wherein the integrated bed control support platform is equipped with an electrical interface; The lower limb servo stress loading mechanism, installed on the integrated bed control support platform, includes a servo drive component and a wearable limb fixation bracket connected thereto, used to apply axial mechanical stress to the patient's lower limbs and drive the lower limbs to perform exercise training; The upper limb multi-dimensional feedback training mechanism is installed on both sides of the integrated bed control support platform, including a multi-joint training arm and an end tool connected to the end, for providing upper limb joint movement and resistance training; A multimodal state perception network is distributed in the lower limb servo stress loading mechanism and the upper limb multidimensional feedback training mechanism to collect training load, joint motion angle and motion trajectory parameters in real time. The central intelligent processing terminal is connected to the lower limb servo stress loading mechanism, the upper limb multidimensional feedback training mechanism and the multimodal state perception network through the electrical interface. It is used to control the motion mechanism to execute the preset training prescription and to perform safety monitoring and feedback control based on sensor data.

[0009] Preferably, the wearable limb fixation brace includes a thigh fixation part, a calf fixation part, and a weight-bearing boot connected in sequence. Both the thigh fixation part and the calf fixation part are length-adjustable structures to adapt to limbs of different body types. The weighted boot has a foot pressure sensor embedded in its sole, and the forefoot area and heel area of ​​the weighted boot are respectively provided with fixed anchor points for connecting the tension band.

[0010] Preferably, the wearable limb fixation bracket further includes a knee joint transmission part connected between the thigh fixation part and the lower leg fixation part. The knee joint transmission part includes multiple sets of gear transmission structures to simulate the movement trajectory of the knee joint, and has a knee joint angle sensor inside for real-time monitoring of the knee joint flexion and extension angle.

[0011] Preferably, the servo drive assembly includes a servo motor and a robotic arm, wherein the servo motor is connected to the load-bearing boot via the robotic arm to apply a thrust along the lower leg axis to the sole of the foot; The lower limb servo stress loading mechanism has a tension sensor and an electromagnetic clutch connected in series in the power transmission path. The tension sensor is used to monitor the actual applied force value in real time. The electromagnetic clutch is controlled by the central intelligent processing terminal and disconnects to physically interrupt the power transmission when the measured force value exceeds a preset safety threshold.

[0012] Preferably, the lower limb servo stress loading mechanism further includes a pulley system disposed at the ankle position opposite the weight-bearing boot; One end of the tension band is fixed to a fixed anchor point in the forefoot area, and the other end passes over the pulley system and is fixed to a fixed anchor point in the heel area, forming a closed-loop traction to transmit the reaction force generated by the patient's upper limb pulling, thereby achieving axial simulated load on the sole of the foot.

[0013] Preferably, the multi-joint training arm has a foldable and retractable structure, and its end is equipped with a force sensor and a quick-change interface; The end effector includes a handle, a grip ball, or a simulated handle, which is detachably mounted to the end of the multi-joint training arm via the quick-change interface, and the grip ball or grip device integrates a pressure sensor for monitoring grip strength.

[0014] Preferably, the multimodal state-aware network includes: A tension sensor is installed in the power transmission path of the lower limb servo stress loading mechanism to monitor axial load; A plantar pressure sensor, embedded in the sole of the weight-bearing boot in the wearable limb fixation bracket, is used to monitor pressure distribution; A knee joint angle sensor is installed in the knee joint transmission part of the wearable limb fixation bracket to monitor the joint angle of the lower limb. A six-dimensional force sensor is installed at the end of the multi-joint training arm to monitor the magnitude and direction of force applied by the upper limb; Joint angle sensors are installed at each joint of the multi-joint training arm to monitor the range of motion of the upper limb. A pressure sensor, located inside the end tool, is used to monitor grip strength; The central intelligent processing terminal displays the data transmitted by the multimodal state perception network in real time in the form of numerical values, waveforms, or virtual human animation.

[0015] Preferably, the central intelligent processing terminal has a built-in coordinate transformation algorithm module, which is used to generate a personalized three-dimensional motion trajectory for the patient after recording the motion parameters of the robotic arm and the patient's healthy lower limb, and call the personalized three-dimensional motion trajectory to drive the servo drive component to reproduce the motion.

[0016] This invention also provides a rehabilitation training method, which utilizes the aforementioned bed-mounted integrated intelligent rehabilitation training system and includes the following steps: S1, retrieve patient records and issue training prescriptions, wherein the training prescriptions include training mode, target load and training duration; S2, for patients using it for the first time, guide them to complete standard movements using their unaffected lower limbs, record motion parameters through the multimodal state perception network, and generate personalized three-dimensional motion trajectories by the central intelligent processing terminal through coordinate transformation algorithms; S3, according to the training prescription and personalized three-dimensional motion trajectory, control the upper limb multidimensional feedback training mechanism and / or the lower limb servo stress loading mechanism to perform at least one of the preset training actions; S4. During the training process, the multimodal state perception network is used to monitor force data in real time. When the force data exceeds a preset safety threshold, the central intelligent processing terminal controls the electromagnetic clutch in the lower limb servo stress loading mechanism to disengage and physically interrupt the power transmission.

[0017] Preferably, the preset training actions include: Upper limb independent training: Through the multi-joint training arm and end effector, the user completes the specified movements under the guidance of animation and voice. The central intelligent processing terminal adjusts the resistance according to the training prescription to carry out joint range of motion training, progressive resistance training, task-oriented training or grip strength training. Lower limb active and passive training: The personalized three-dimensional motion trajectory is invoked to drive the affected lower limb to reproduce the motion. Upper and lower limb coordinated simulated weight-bearing training: In response to the pulling of the upper limb on the end tool, the tension band generates an axial reaction force along the leg direction on the sole of the weight-bearing boot; Lower limb simulated movement training: The servo drive component is controlled to drive the lower limbs to alternately extend-flex or circular movements to perform simulated walking or cycling training.

[0018] The above technical solution has the following advantages or beneficial effects: By integrating the lower limb servo stress loading mechanism and the upper limb multidimensional feedback training mechanism onto a unified bed-controlled support platform, bedridden patients can perform lower limb axial stress loading training and upper limb joint range of motion and resistance training separately or simultaneously on a single platform without leaving the bed. This fundamentally breaks the limitation of the separation between upper and lower limb rehabilitation training, making the training process more holistic and continuous. The multimodal state perception network can collect core parameters such as training load, joint range of motion, and movement trajectory in real time and in multiple dimensions. Based on this, the central intelligent processing terminal can continuously perform safety monitoring and closed-loop feedback control while controlling the motion mechanism to execute the training prescription. This significantly improves the accuracy of training load application and the safety response capability in abnormal states, effectively ensuring the safety and effectiveness of early active rehabilitation for bedridden patients. Attached Figure Description

[0019] Figure 1 A schematic diagram of the structure of a bed-mounted integrated intelligent rehabilitation training system is shown in a preferred embodiment of the present invention. Figure 2 This is a flowchart illustrating a preferred embodiment of the present invention, illustrating a rehabilitation training method. Detailed Implementation

[0020] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. The present invention is not limited to this embodiment; other embodiments that conform to the spirit of the present invention may also fall within the scope of the present invention.

[0021] In a preferred embodiment of the present invention, based on the above-mentioned problems existing in the prior art, a bed-mounted integrated intelligent rehabilitation training system is provided, such as... Figure 1 As shown, it includes an integrated bed control support platform 1, a lower limb servo stress loading mechanism 2, an upper limb multi-dimensional feedback training mechanism 3, a multi-modal state perception network, and a central intelligent processing terminal 4.

[0022] The integrated bed-controlled support platform 1 serves as the physical foundation and load-bearing framework of the entire system. Using a structurally reinforced hospital bed as its base, the platform can withstand various additional loads generated during training. Power and signal lines are pre-embedded within the platform, and standardized mechanical and electrical interfaces are located at the foot and sides of the bed. Various institutions using the platform for rehabilitation training can quickly install, disassemble, and adjust the position through these interfaces without the need for additional tools, ensuring flexibility in modular combination and convenience for clinical deployment.

[0023] The lower limb servo stress loading mechanism 2 is mounted on the integrated bed control support platform and is used to apply axial mechanical stress and drive movement to the lower limbs of bedridden patients. This mechanism includes a servo drive assembly 22 and a wearable limb fixation bracket 21 connected to the servo drive assembly. During use, the wearable limb fixation bracket is fixed to the patient's lower limb and can accurately transmit the controllable mechanical thrust generated by the servo drive assembly to the lower limb, thereby simulating axial stress along the tibia or driving the lower limb to complete a set trajectory for movement training.

[0024] The upper limb multi-dimensional feedback training mechanism 3 is installed on both sides of the integrated bed-controlled support platform to meet the patient's needs for upper limb joint range of motion and muscle strength training. This mechanism includes symmetrically arranged multi-joint training arms 31 and end effectors 32 connected to the ends of each training arm 31. The multi-joint training arms can perform multi-degree-of-freedom movements and resistance loading in space, guiding the patient to complete range of motion training, progressive resistance training, and task-oriented training, including those for the shoulder, elbow, and wrist joints.

[0025] A multimodal state perception network is distributedly integrated into the lower limb servo stress loading mechanism 2 and the upper limb multidimensional feedback training mechanism 3. Its function is to collect various key parameters during the training process in real time, including the magnitude of the training load, the real-time joint range of motion, and the three-dimensional motion trajectory parameters of the end effector and the limb. The collected sensor data is aggregated in real time to the central intelligent processing terminal, providing a data foundation for training control and safety assessment.

[0026] The central intelligent processing terminal 4 is located in the control cabinet on one side of the platform. Through the electrical interface of the integrated bed control support platform 1, it communicates with the lower limb servo stress loading mechanism 2, the upper limb multi-dimensional feedback training mechanism 3, and all sensors constituting the multimodal state perception network. This terminal stores executable training prescriptions and can control the aforementioned movement mechanisms to perform actions according to preset training modes, target loads, and durations based on medical instructions. Simultaneously, the terminal continuously reads sensor data and performs real-time safety monitoring and feedback control. For example, it adjusts the output or triggers protection mechanisms when abnormal loads are detected, ensuring that the entire training process remains quantifiable, controllable, and safe.

[0027] This embodiment integrates the drive unit, actuator, sensor network, and control unit onto a single bed-mounted platform, making the system a truly integrated bed-mounted intelligent training terminal. Bedridden patients can perform lower limb axial stress loading training and upper limb multi-mode resistance training separately or simultaneously, according to a prescription, while in a supine position. The training load and key joint movement parameters are quantified and sensed in real time, and dynamically adjusted and protected against anomalies by a central intelligent processing terminal. This systematically improves the accessibility, safety, and traceability of training data for early active rehabilitation of long-term bedridden patients.

[0028] In another preferred embodiment of the present invention, based on the above-mentioned bed-mounted integrated intelligent rehabilitation training system, the wearable limb fixation bracket further includes a thigh fixation part, a calf fixation part, and a weight-bearing boot connected in sequence.

[0029] Both the thigh fixation and calf fixation are length-adjustable structures. For example, the length can be changed by means of telescopic sleeves or adjusting grooves in conjunction with locking devices to adapt to the lower limb size of different patients. When in use, both are reliably fixed to the affected limb by flexible straps, which not only ensures the stability of force transmission, but also prevents excessive local pressure.

[0030] The weighted boots have embedded plantar pressure sensors in the soles that can detect changes in force in different areas of the foot in real time. The forefoot and heel areas of the weighted boots are equipped with anchor points for connecting the tension bands. These anchor points are designed as high-strength connectors that can withstand large tensile forces from external pulls without deformation.

[0031] This structure enables the fixed bracket to not only transmit the axial thrust and torque provided by the servo drive components, but also allows the introduction of auxiliary traction forces from other directions via tension bands, providing a structural interface for subsequent upper and lower limb coordinated loading modes.

[0032] The adjustable wearable limb fixation brace of this embodiment can conform to the affected limbs of different body types, ensuring the physiological rationality of the force transmission path and the comfort of the patient. At the same time, by integrating a plantar pressure sensor into the sole of the boot and setting multiple anchor points in the boot body, this passive load-bearing component is upgraded into an intelligent terminal structure with dual functions of sensing and force transmission interface, providing a hardware foundation for more complex closed-loop training strategies.

[0033] In another preferred embodiment of the present invention, the wearable limb fixation brace also includes a knee joint transmission unit connected between the thigh fixation unit and the lower leg fixation unit. This knee joint transmission unit contains multiple sets of gear transmission structures. These gears mesh with each other and, with a specific transmission ratio, match the instantaneous change in the center of motion of the human knee joint during flexion and extension movements. This accurately simulates the physiological movement trajectory of the knee joint, making the lower leg movement driven by the drive mechanism more closely resemble natural joint movement and avoiding unreasonable stress concentration.

[0034] The knee joint transmission unit also integrates an angle sensor, specifically a non-contact angle encoder, to monitor the flexion and extension angles of the knee joint in real time and upload the data to a multimodal state perception network in real time.

[0035] Therefore, the system can not only control the movement trajectory of the limb ends, but also dynamically acquire and monitor the real-time angle of the knee joint, providing key data for compliant control and over-limit protection.

[0036] This embodiment simulates the physiological movement of the knee joint through gear transmission, ensuring the biomechanical consistency of joint movement during passive exercise or active assisted training, reducing the risk of joint injury during training. At the same time, the built-in angle sensor supplements the multimodal state perception network of this system with key joint position feedback.

[0037] In another preferred embodiment of the invention, the servo drive assembly includes a servo motor and a robotic arm. The servo motor is connected to a weighted boot in a wearable limb fixation bracket via the robotic arm to apply a thrust along the lower leg axis to the sole of the foot, thereby achieving precise control of the axial load on the lower limb.

[0038] In the power transmission path from the servo motor to the load-bearing shoe, a tension sensor and an electromagnetic clutch are connected in series. The tension sensor is used to monitor the actual force value transmitted along the path in real time and transmit the measured data to the central intelligent processing terminal. The electromagnetic clutch is controlled by the central intelligent processing terminal and is normally in a closed state to transmit power. When the central intelligent processing terminal determines that the measured force value exceeds the preset safety threshold, it immediately cuts off the power supply to the electromagnetic clutch, physically interrupting the power transmission and achieving millisecond-level overload protection.

[0039] When the measured force value falls back to a safe range, the electromagnetic clutch can automatically re-engage and resume training. This dual safety mechanism, consisting of real-time sensor monitoring and direct physical disengagement by the electromagnetic clutch, can protect patients from instantaneous or cumulative overload injuries at the lowest level, without relying on software responses.

[0040] This embodiment achieves high-precision axial stress loading capability through specific design of servo drive components and power transmission paths, and enables the system to reliably protect itself through hardware physical interruption at the moment of overload. This effectively solves the core problem of achieving precise control and absolute safety protection of axial mechanical stress in bedridden rehabilitation.

[0041] In another preferred embodiment of the invention, the lower limb servo stress loading mechanism 2 further includes a pulley system disposed on a structural member at the ankle position opposite the weighted boot. Simultaneously, the system provides a tension band, one end of which is fixed to a fixed anchor point in the forefoot area of ​​the weighted boot, and the other end, after passing over the pulley system, is fixed to a fixed anchor point in the heel area, thus forming a closed-loop traction path around the sole of the foot. During training, a portion of the tension band can extend to a force-applying area operable by the patient's upper limb via a connector provided by the system. When the patient pulls on the force-applying area with their upper limb, the resulting tension is transmitted along the tension band, redirected by the pulley system, and generates a downward reaction force along the calf axis on the weighted boot, thereby applying an axial load to the sole of the foot simulating an upright position. In this mode, the patient can autonomously control the magnitude of the upper limb tension to simultaneously adjust the axial stress on the lower limb, truly achieving coordinated force application and loading of the upper and lower limbs. The entire process remains under real-time monitoring and protection by the tension sensor and the electromagnetic clutch.

[0042] This embodiment cleverly transforms the active pulling force of the patient's upper limb into a simulated axial load of the lower limb by utilizing a pulley system and a closed-loop traction belt. This achieves a physiological weight-bearing simulation mode without an external power source, which not only enriches the training mode but also promotes the functional linkage between the upper and lower limbs, conforming to the muscle synergistic recruitment mode under natural human weight-bearing conditions.

[0043] In another preferred embodiment of the invention, the multi-joint training arm is designed as a foldable and retractable structure, which can be folded to the side of the bed when not in use, avoiding space occupation and facilitating clinical nursing operations. The end effector of the multi-joint training arm is equipped with a force sensor and a quick-change interface. The force sensor is specifically a six-dimensional force sensor, capable of accurately sensing the magnitude and direction of the force applied by the patient in three-dimensional space. End effectors include different types such as handles, grip balls, or simulated handles, each suitable for different training tasks. They are all detachably mounted to the end effector of the multi-joint training arm via quick-change interfaces, allowing medical staff to quickly change them according to the training prescription. The grip ball or grip device integrates a pressure sensor for real-time monitoring of the patient's grip strength, and the values ​​are incorporated into a multimodal state sensing network for unified processing and display.

[0044] With its foldable structure and quick-change interface, the upper limb multi-dimensional feedback training mechanism provided in this embodiment achieves rapid switching of end-effectors and comprehensive perception of mechanical parameters without affecting the normal use of the hospital bed, providing a highly flexible and quantifiable hardware platform for upper limb rehabilitation training for bedridden patients.

[0045] In another preferred embodiment of the present invention, the multimodal state-aware network includes the following specific sensors: A tension sensor is installed in the power transmission path of the lower limb servo stress loading mechanism to monitor axial load. A plantar pressure sensor embedded in the sole of a weight-bearing boot in a wearable limb fixation brace is used to monitor plantar pressure distribution. A knee joint angle sensor, installed in the knee joint transmission unit, is used to monitor the joint angle of the lower limb. A six-dimensional force sensor is installed at the end of the multi-joint training arm to monitor the magnitude and direction of the force applied by the upper limb. Joint angle sensors are installed at each joint of the multi-joint training arm to monitor the range of motion of the upper limb. And a pressure sensor located inside the end tool to monitor grip strength.

[0046] All of these sensors are connected to a central intelligent processing terminal. The terminal has a built-in visualization software module that can display the data collected by the sensors in the form of numerical values, real-time waveforms, or virtual human body animations on a touch screen, so that medical staff and patients can intuitively grasp the training status, such as the current joint angles, axial load value change curves, and plantar pressure distribution heat maps.

[0047] This embodiment specifies the sensor configuration scheme and data presentation method of the multimodal state perception network, and constructs a complete perception matrix covering key parameters of upper and lower limb biomechanics and kinematics, making the rehabilitation training process fully visualized and quantifiable, and providing objective data support for clinical assessment and program adjustment.

[0048] In another preferred embodiment of the present invention, the central intelligent processing terminal has a built-in coordinate transformation algorithm module. Specifically, when a patient using the system for the first time performs a standard movement (e.g., ankle pumping or knee flexion-extension) using their unaffected lower limb, the system synchronously records data such as the robotic arm motion parameters in the servo drive component and the knee joint angle measured by the knee joint transmission unit through a multimodal state perception network.

[0049] The terminal runs a coordinate transformation algorithm, using a kinematic model to correlate and transform the joint space parameters of the robotic arm with the actual anatomical motion parameters of the affected limb, ultimately fitting and generating a personalized 3D motion trajectory that is entirely unique to the patient. This personalized trajectory is stored in the central intelligent processing terminal. In subsequent training, the terminal can call upon this personalized 3D motion trajectory to instruct the servo drive components to accurately reproduce the movement patterns and range of motion of the unaffected side of the patient's affected lower limb, thereby implementing highly individualized training.

[0050] This embodiment utilizes coordinate transformation and trajectory learning technology to achieve a complete closed loop from the acquisition of movement data of the healthy leg to the execution of training for the affected leg. This enables passive active training of the lower limbs to be truly tailored to each individual, avoiding problems such as mismatch in joint range of motion and inadequate muscle stimulation that may result from using fixed universal trajectories. This significantly improves the accuracy of training and patient compliance.

[0051] The present invention also provides a rehabilitation training method using the above-mentioned bed-mounted integrated intelligent rehabilitation training system.

[0052] In a preferred embodiment of the present invention, such as Figure 2 As shown, the method includes the following steps: Step S1: Retrieve patient records and issue training prescriptions. Medical staff retrieve the corresponding patient's electronic rehabilitation records through the human-machine interface of the central intelligent processing terminal, and issue a prescription for this training based on the assessment results. The prescription data includes parameters such as the selected training mode, the target workload to be achieved, and the training duration.

[0053] Step S2: For patients using the system for the first time, the system guides them to complete a set of standard movements using their unaffected lower limb. This process is guided through screen animations and voice prompts. During the movements, the multimodal state perception network records the motion parameters of the robotic arm and knee joint angle of the servo drive component in real time. The central intelligent processing terminal then processes these parameters using its built-in coordinate transformation algorithm module to generate a personalized three-dimensional motion trajectory specific to the patient and stores it for later use.

[0054] In step S3, the central intelligent processing terminal controls the upper limb multidimensional feedback training mechanism and / or the lower limb servo stress loading mechanism to perform at least one preset training action specified in the prescription, based on the issued training prescription and the aforementioned personalized three-dimensional motion trajectory.

[0055] Step S4: During the training process, the multimodal state perception network continuously monitors various force values ​​in real time.

[0056] The central intelligent processing terminal compares the measured force value with a preset safety threshold. When the force value exceeds the threshold, it immediately sends a control command to the electromagnetic clutch to disengage it, thus physically interrupting power transmission and achieving absolute overload protection for the affected limb. At the same time, the system can issue warnings through sound and screen prompts.

[0057] In step S2, the specific process for generating personalized three-dimensional motion trajectories for patients using this system for the first time is as follows: First, the system completes preparatory work, including mechanism calibration and initial pose recording. The patient lies supine on the integrated bed-controlled support platform. Medical staff correctly attach the wearable limb fixation brace to the patient's unaffected lower limb, ensuring that the thigh fixation, calf fixation, and weight-bearing boot are securely fixed and rigidly connected to the end effector of the robotic arm via the servo-driven assembly and the interface on the weight-bearing boot. The central intelligent processing terminal controls the servo motor to enter zero-force control mode or low-impedance following mode, so that the robotic arm only compensates for its own weight and applies almost no additional load to the patient's lower limb. Subsequently, the system calibrates various sensors, including zeroing the tension sensor and plantar pressure sensor, recording the reference angle of the current extension position by the knee joint angle sensor, and recording the coordinate values ​​of each joint of the robotic arm at the starting position by the servo motor encoder, thereby determining the initial end effector pose in the robotic arm's base coordinate system.

[0058] Secondly, the system guides the patient to perform a set of preset standard movements using the unaffected lower limb through on-screen animations and voice prompts. These standard movements are designed according to the needs of rehabilitation medicine and typically include, but are not limited to: slow ankle dorsiflexion and plantar flexion (ankle pump), gradual knee flexion from full extension to the maximum tolerable angle followed by slow extension, and straight leg raises. The animations demonstrate the movements, and the voice prompts provide guidance on the required speed and range of motion, ensuring that the movements cover the main joint range of motion and typical movement patterns of the unaffected lower limb in daily activities. During the movements, the patient can use the voluntary muscle strength of the unaffected lower limb to complete all the movements, while the robotic arm follows the movement in zero-force mode without hindering the movement.

[0059] During motion execution, the multimodal state perception network synchronously collects multi-dimensional data at a fixed frequency. The collected data includes: angle and angular velocity values ​​of each joint of the robotic arm output by the servo motor encoder, used to calculate the three-dimensional spatial pose of the robotic arm's end effector; real-time knee flexion and extension angles output by the knee joint angle sensor built into the knee joint transmission unit; plantar pressure distribution sensed by the plantar pressure sensor on the sole of the weight-bearing boot; and axial force values ​​monitored by the tension sensor connected in series along the power transmission path. All sampled data is timestamped and recorded in real-time by the central intelligent processing terminal.

[0060] Next, the central intelligent processing terminal uses its built-in coordinate transformation algorithm module to process the collected data and generate personalized 3D motion trajectories. The algorithm steps are as follows: (1) Forward kinematics calculation. Using the known link parameters and time series of joint angles of the robotic arm, the three-dimensional position coordinates (x, y, z) and attitude angles (such as Euler angles or quaternions) of the robotic arm end effector in the base coordinate system are calculated as a sequence of time through the robot's forward kinematics model. This end effector position sequence represents the motion path of the center point of the load-bearing boot in the workspace of the robotic arm.

[0061] (2) Knee joint angle fusion. The knee joint angle time series collected by the knee joint angle sensor is aligned and fitted with the end-effector pose sequence of the robotic arm. Since there is a biomechanical constraint between the knee joint angle change and the end-effector pose change, a mapping relationship between the end-effector pose and the knee joint angle can be constructed by performing correlation analysis between the two.

[0062] (3) Personalized anatomical trajectory conversion and fitting. A simplified kinematic model of the human lower limb is pre-stored in the terminal. This model uses the hip joint as an approximate ball-and-socket joint and the knee joint as a single-degree-of-freedom hinge joint as a simplified structure, and has adjustable thigh length and calf length parameters (which can be set according to the patient's height or direct measurement values). Using the end pose sequence obtained in step (1) and the knee joint angle sequence obtained in step (2), the time series of the corresponding hip joint flexion-extension angle and abduction-inversion angle are calculated by numerical inverse kinematics solution or optimization fitting method, thereby obtaining a set of complete lower limb joint spatial motion data containing the hip joint angle, knee joint angle and ankle joint corresponding end motion trajectory. These data are defined as the patient's personalized three-dimensional motion trajectory, which is essentially a quantitative expression of the coordinated motion pattern of the joints of the healthy lower limb at the kinematic level.

[0063] (4) Trajectory smoothing and storage. The obtained joint angle sequences and end-effector trajectories are subjected to Butterworth low-pass filtering or spline smoothing to remove jitter caused by slight muscle tremors or sensor noise, ensuring smooth and continuous trajectory. The smoothed trajectory data is packaged into a unique motion trajectory file for the patient, stored in the memory of the central intelligent processing terminal, and associated with the patient ID.

[0064] Finally, the system performs a verification reproduction. The central intelligent processing terminal calls the newly generated personalized 3D motion trajectory and controls the servo drive component to passively reproduce the movement at extremely low speeds, moving the affected lower limb (or the unaffected limb, depending on clinical needs). Simultaneously, the real-time joint angles and motion trajectory are displayed on the screen, and the patient or medical staff confirms the comfort and appropriateness of the movement. If discomfort is observed, the trajectory amplitude or speed ratio can be fine-tuned under prescription permissions; the adjusted trajectory is used as the final prescription.

[0065] This process is entirely based on the existing physical structure and sensor array of this system. By collecting active motion data of the healthy side through zero-force following, fusing the pose and joint angle data of the robotic arm end effector, and using kinematic transformation and fitting, a highly personalized three-dimensional motion trajectory is generated. This enables subsequent training of the affected lower limb to reproduce the patient's own physiological movement pattern, achieving precise rehabilitation treatment tailored to each individual.

[0066] This embodiment, starting from the clinical workflow, details the entire closed-loop process from prescription issuance, personalized trajectory generation, training execution to dynamic safety protection, enabling the system of this invention to be applied to clinical practice in a standardized manner. The healthy side learning mechanism during initial use ensures individualized adaptation of training movements, while the force value monitoring throughout the process and the millisecond-level physical disconnection protection strategy minimize the risk of secondary injury caused by sudden spasms or mechanical failures.

[0067] In another preferred embodiment of the present invention, the preset training actions performed in step S3 specifically include the following multiple modalities, which can be performed individually or in combination according to the prescription.

[0068] For the independent upper limb training modality, the multi-dimensional feedback training mechanism is controlled by a central intelligent processing terminal. Specifically, the system guides the patient to complete designated movements on the multi-joint training arm and end-effectors such as the handle through screen animations and voice prompts.

[0069] The training content includes basic joint range of motion training such as shoulder flexion / extension, abduction / adduction, elbow flexion / extension, and wrist flexion / extension and rotation. Simultaneously, the system can adjust the resistance provided by the training arm in real time according to the prescription to implement progressive resistance training. The resistance mode can be constant resistance, progressively increasing resistance, or adaptive resistance that changes according to the patient's exertion, with the resistance range continuously adjustable.

[0070] Furthermore, this modality supports task-oriented training, which involves changing the end effector to simulate everyday actions such as drinking (holding a cup), pushing a door (forward pushing motion), and grasping objects (opening and closing grasping). Its training action library is designed with reference to upper limb function assessment standards such as FMA-UE and ARAT. Finally, when a grip ball or grip strengthener is installed, grip strength training can be performed independently, with the grip strengthener's built-in pressure sensor monitoring and providing feedback on grip strength in real time.

[0071] For the active and passive training modalities of the lower limbs, the system first calls up the patient's stored personalized three-dimensional motion trajectory, and then controls the servo drive component to move the affected lower limb, which is wearing a fixation brace, to accurately reproduce the trajectory. During the training process, it can be switched as needed to passive training (the robotic arm fully drives the movement), active assisted training (the patient actively exerts force, and the robotic arm assists to supplement), or active resistance training (the robotic arm applies reverse resistance to the patient's movement), so as to achieve personalized rehabilitation that combines active and passive methods.

[0072] The upper and lower limb coordinated simulated weight-bearing training modality is a unique training method that converts upper limb tension into lower limb axial load. During training, the patient uses their upper limbs to pull on the force application part (such as the grip ring) connected to the resistance band. This tension is transmitted through the main unit to the pulley, and then through the resistance band connected to the front and rear fixed anchor points of the weight-bearing boot and passing around the ankle pulley, forming a reaction force on the sole of the foot along the leg direction, thus simulating the axial load when standing. In this mode, the magnitude of the upper limb tension directly determines the magnitude of the axial stress borne by the affected limb. The patient can autonomously adjust the degree of force according to their own feelings, achieving coordinated training of upper and lower limb functions. Throughout the process, the system's tension sensor monitors the load value in real time. Once the preset weight-bearing safety threshold is exceeded, the electromagnetic clutch immediately disconnects the power transmission, protecting the affected limb from overload damage in a hardware manner.

[0073] The lower limb simulated movement training modality further includes two sub-modes. The first is simulated walking training, where the system controls servo drive components to alternately extend and flex both lower limbs, with the left leg extending while the right leg flexes, alternating repeatedly to simulate the rhythm of lower limb movement during walking. During training, parameters such as cadence and stride length can be set by medical staff in the prescription, with stride length automatically adapted to the patient's height by the system. The second is simulated cycling training, where the patient lies supine with both lower limbs raised and secured by weighted boots. The system controls servo drive components to move both lower limbs in circular motion in space, simulating the action of cycling. Training parameters such as cadence can be set between 10 and 40 revolutions per minute, resistance can be continuously adjusted within the range of 0 to 20 kg, and training duration can also be determined according to the prescription. This mode effectively trains the coordinated movement of the hip, knee, and ankle joints, improves joint mobility, and strengthens lower limb muscles. During simulated walking or cycling training, axial loads can also be applied simultaneously as needed to better simulate changes in foot load under real standing or movement conditions.

[0074] This embodiment elaborates on the implementation mechanisms, specific movement designs, and adjustable parameters of different training modalities, fully demonstrating the diversity and freely combinable nature of the training content of this invention. Through the free selection and combination of modalities such as independent upper limb training, active and passive lower limb training, coordinated simulated weight-bearing, and simulated gait / cycling, this system can comprehensively cover the full rehabilitation cycle needs of bedridden patients, from early passive activity to later active resistance, greatly expanding the system's clinical applicability.

[0075] The above are merely preferred embodiments of the present invention and are not intended to limit the implementation methods and protection scope of the present invention. Those skilled in the art should recognize that any equivalent substitutions and obvious changes made using the content of this specification and illustrations should be included within the protection scope of the present invention.

Claims

1. A bed-mounted integrated intelligent rehabilitation training system, characterized in that, include: An integrated bed control support platform, wherein the integrated bed control support platform is equipped with an electrical interface; The lower limb servo stress loading mechanism, installed on the integrated bed control support platform, includes a servo drive component and a wearable limb fixation bracket connected thereto, used to apply axial mechanical stress to the patient's lower limbs and drive the lower limbs to perform exercise training; The upper limb multi-dimensional feedback training mechanism is installed on both sides of the integrated bed control support platform, including a multi-joint training arm and an end tool connected to the end, for providing upper limb joint movement and resistance training; A multimodal state perception network is distributed in the lower limb servo stress loading mechanism and the upper limb multidimensional feedback training mechanism to collect training load, joint motion angle and motion trajectory parameters in real time. The central intelligent processing terminal is connected to the lower limb servo stress loading mechanism, the upper limb multidimensional feedback training mechanism and the multimodal state perception network through the electrical interface. It is used to control the motion mechanism to execute the preset training prescription and to perform safety monitoring and feedback control based on sensor data.

2. The bed-mounted integrated intelligent rehabilitation training system according to claim 1, characterized in that, The wearable limb fixation brace includes a thigh fixation part, a calf fixation part, and a weight-bearing boot connected in sequence. Both the thigh fixation part and the calf fixation part are length-adjustable structures to adapt to limbs of different body types. The weighted boot has a foot pressure sensor embedded in its sole, and the forefoot area and heel area of ​​the weighted boot are respectively provided with fixed anchor points for connecting the tension band.

3. The bed-mounted integrated intelligent rehabilitation training system according to claim 2, characterized in that, The wearable limb fixation brace also includes a knee joint transmission part connected between the thigh fixation part and the lower leg fixation part. The knee joint transmission part includes multiple sets of gear transmission structures to simulate the movement trajectory of the knee joint, and has a knee joint angle sensor inside for real-time monitoring of the knee joint flexion and extension angle.

4. The bed-mounted integrated intelligent rehabilitation training system according to claim 2, characterized in that, The servo drive assembly includes a servo motor and a robotic arm. The servo motor is connected to the load-bearing boot via the robotic arm to apply a thrust along the lower leg axis to the sole of the foot. The lower limb servo stress loading mechanism has a tension sensor and an electromagnetic clutch connected in series in the power transmission path. The tension sensor is used to monitor the actual applied force value in real time. The electromagnetic clutch is controlled by the central intelligent processing terminal and disconnects to physically interrupt the power transmission when the measured force value exceeds a preset safety threshold.

5. The bed-mounted integrated intelligent rehabilitation training system according to claim 4, characterized in that, The lower limb servo stress loading mechanism also includes a pulley system located at the ankle position opposite the weight-bearing boot; One end of the tension band is fixed to a fixed anchor point in the forefoot area, and the other end passes over the pulley system and is fixed to a fixed anchor point in the heel area, forming a closed-loop traction to transmit the reaction force generated by the patient's upper limb pulling, thereby achieving axial simulated load on the sole of the foot.

6. The bed-mounted integrated intelligent rehabilitation training system according to claim 1, characterized in that, The multi-joint training arm has a foldable and retractable structure, and its end is equipped with a force sensor and a quick-change interface. The end effector includes a handle, a grip ball, or a simulated handle, which is detachably mounted to the end of the multi-joint training arm via the quick-change interface, and the grip ball or grip device integrates a pressure sensor for monitoring grip strength.

7. The bed-mounted integrated intelligent rehabilitation training system according to claim 1, characterized in that, The multimodal state-aware network includes: A tension sensor is installed in the power transmission path of the lower limb servo stress loading mechanism to monitor axial load; A plantar pressure sensor, embedded in the sole of the weight-bearing boot in the wearable limb fixation bracket, is used to monitor pressure distribution; A knee joint angle sensor is installed in the knee joint transmission part of the wearable limb fixation bracket to monitor the joint angle of the lower limb. A six-dimensional force sensor is installed at the end of the multi-joint training arm to monitor the magnitude and direction of force applied by the upper limb; Joint angle sensors are installed at each joint of the multi-joint training arm to monitor the range of motion of the upper limb. A pressure sensor, located inside the end tool, is used to monitor grip strength; The central intelligent processing terminal displays the data transmitted by the multimodal state perception network in real time in the form of numerical values, waveforms, or virtual human animation.

8. The bed-mounted integrated intelligent rehabilitation training system according to claim 4, characterized in that, The central intelligent processing terminal has a built-in coordinate transformation algorithm module, which generates a personalized three-dimensional motion trajectory for the patient after recording the motion parameters of the robotic arm and the patient's healthy lower limb. The personalized three-dimensional motion trajectory is then used to drive the servo drive component to reproduce the motion of the affected lower limb.

9. A rehabilitation training method, characterized in that, The application of the bed-mounted integrated intelligent rehabilitation training system as described in any one of claims 1-8 includes the following steps: S1, retrieve patient records and issue training prescriptions, wherein the training prescriptions include training mode, target load and training duration; S2, for patients using it for the first time, guide them to complete standard movements using their unaffected lower limbs, record motion parameters through the multimodal state perception network, and generate personalized three-dimensional motion trajectories by the central intelligent processing terminal through coordinate transformation algorithms; S3, according to the training prescription and personalized three-dimensional motion trajectory, control the upper limb multidimensional feedback training mechanism and / or the lower limb servo stress loading mechanism to perform at least one of the preset training actions; S4. During the training process, the multimodal state perception network is used to monitor force data in real time. When the force data exceeds a preset safety threshold, the central intelligent processing terminal controls the electromagnetic clutch in the lower limb servo stress loading mechanism to disengage and physically interrupt the power transmission.

10. The rehabilitation training method according to claim 9, characterized in that, The preset training actions include: Upper limb independent training: Through the multi-joint training arm and end effector, the user completes the specified movements under the guidance of animation and voice. The central intelligent processing terminal adjusts the resistance according to the training prescription to carry out joint range of motion training, progressive resistance training, task-oriented training or grip strength training. Lower limb active and passive training: The personalized three-dimensional motion trajectory is invoked to drive the affected lower limb to reproduce the motion. Upper and lower limb coordinated simulated weight-bearing training: In response to the pulling of the upper limb on the end tool, the tension band generates an axial reaction force along the leg direction on the sole of the weight-bearing boot; Lower limb simulated movement training: The servo drive component is controlled to drive the lower limbs to alternately extend-flex or circular movements to perform simulated walking or cycling training.

Images