A lower limb assistance training method based on a treadmill

Abstract

The application relates to a kind of lower limb assisted training methods based on walking machine assistance.The assisted training method realized according to gait equation planning fixed training track, and training track is discretized to the drive unit of robot according to step length, and passive walking machine matched with lower limb rehabilitation robot is used as the dynamic platform of training to provide the training speed of the training track of robot.The assisted training method collects the real-time position and interaction torque of drive unit during operation as the judgment basis of assisted scheme selection.In addition, the interaction force threshold is set for training anti-shake.The assisted training assistance scheme includes pure passive assistance scheme and compliance assistance scheme.

Description

Technical Field

[0001] This invention relates to the field of rehabilitation robot control, and more particularly to a lower limb assisted training method based on a treadmill. Background Technology

[0002] A significant factor contributing to lower limb dysfunction is damage to the central nervous system. Correspondingly, the fundamental goal of rehabilitation training is to stimulate the reorganization and compensation of the central nervous system, promote the recovery of motor perception function, and ultimately rebuild lower limb motor function. Therefore, after lower limb function is impaired, in addition to medication, rehabilitation training should be conducted to improve rehabilitation efficiency. Applying advanced robotics technology to lower limb clinical rehabilitation can leverage its advantages of high automation, precision, and suitability for repetitive, high-intensity physical labor. This can effectively reduce the workload of medical staff, improve the effectiveness of lower limb rehabilitation training, compensate for the shortage of rehabilitation resources, and help accelerate the rehabilitation process. Based on lower limb rehabilitation robots, several rehabilitation training methods, including passive training and resistance training, have been developed. However, for patients with some initial mobility, passive training cannot meet their needs for improved rehabilitation outcomes. Furthermore, active training such as resistance training requires a higher level of mobility and cannot provide sufficient training for patients. Therefore, developing a lower limb assisted training method that can adapt to patients with some initial mobility and provides corresponding assistance programs based on their training performance is crucial. Summary of the Invention

[0003] The purpose of this invention is to provide a lower limb assisted training method based on a treadmill to solve the above-mentioned problems.

[0004] This invention provides a lower limb assisted training method based on a treadmill, characterized by using a passive treadmill to provide a training platform and a suitable movement speed for the robot. The method comprises two assistance schemes: passive assisted training and compliant assisted training. During movement, the robot collects force-position feedback data from the drive unit and switches between assistance schemes based on the trajectory equation.

[0005] First, a passive treadmill is combined with a lower limb rehabilitation robot. The treadmill provides a dynamic platform for assisted lower limb training. The treadmill is driven by friction applied to its moving surface. The treadmill's movement speed can be adapted to the specific assistive device.

[0006] Furthermore, kinematic functions for the hip and knee joints are established, representing the robot's motion trajectory. During the motion process, the robot discretizes the set motion trajectory to the motion position of each drive unit based on the step length. Based on the kinematic functions, the target position of the drive unit is updated in real time at fixed step intervals; that is, the target position is sent to the driver of each servo motor at fixed step intervals to achieve interpolation.

[0007] Furthermore, the assisted training method provides both passive and compliant assistance schemes for lower limb rehabilitation robots. The choice of assistance scheme depends on the feedback parameters collected during robot operation. During operation, the robot acquires the real-time interaction torque and position of the drive unit and sets an interaction torque threshold. When the interaction torque is opposite to the movement direction of the drive unit or the interaction torque is below the threshold, the robot completes the preset movement trajectory by fixing the step length ΔS, thus achieving passive assistance. Conversely, when the interaction torque is consistent with the joint movement direction, the robot's movement speed is increased by increasing the step length 2ΔS, thus achieving compliant assistance.

[0008] Furthermore, based on the control strategy described above, when the robot reaches the 0.04T and 0.94T positions of the motion cycle during the assisted training motion trajectory, due to the large fluctuations in the interactive torque, a passive assist scheme is selected during this motion period to ensure operational stability.

[0009] Compared with existing technologies, the advantages of this invention are: the treadmill can provide an adaptive movement speed based on the movement trajectory provided by the robot, thus improving coordination. The developed assisted training provides a variety of assisted movement schemes, improving existing training models for lower limb rehabilitation. Attached Figure Description

[0010] Figure 1 This is a flowchart of the assisted training control process;

[0011] Figure 2 This diagram illustrates how interpolation is achieved by sending the target position of the driving unit through step intervals.

[0012] Figure 3 This is the logic diagram for controlling the speed of assisted training. Detailed Implementation

[0013] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0014] The lower limb assistive training system based on a treadmill used in this invention includes a passive treadmill and a lower limb rehabilitation robot. The treadmill provides the robot with a platform for movement, while the robot provides the assistive mechanism.

[0015] In the assisted training, the motion trajectory planning of the robot uses motion data from OpenSim software (developed by Stanford University). One motion cycle is 1 second. To obtain the motion tracking curve for training, the data is processed and put into MATLAB. Multi-order trigonometric functions are used to fit the discrete data points in the normal motion curve, and the coefficients that minimize the residual error are found. Finally, the function formula (1) is determined. Considering that the actual motion speed is relatively low, the motion cycle of the lower limb rehabilitation robot is set to 4 seconds, and the amplitude coefficient a, frequency coefficient ω, and initial phase in the standard trajectory are adjusted. The offset k is adjusted to obtain a suitable new motion trajectory. The motion trajectory of the lower limb rehabilitation robot planned by the kinematic function is shown in (1).

[0016]

[0017] Where a represents the vibration coefficient and ω represents the frequency coefficient. represents the initial phase, and k represents the offset distance.

[0018] Since the robot's pose during operation is periodic and symmetrical, the robot's trajectory throughout the entire cycle can be deduced based on the acquired new gait. The kinematic equations for the hip joint are shown in equation (2), and the kinematic equations for the knee joint are shown in equation (3).

[0019]

[0020]

[0021] Where, θ hip The kinematic function representing the hip joint, θ knee This represents the motion function of the knee joint, where t represents the stride length.

[0022] The specific control process for assisted training is shown in the appendix. Figure 1 As shown, the training mode is set to assisted training mode in the human-computer interaction interface. After the mode starts, the torque sensor reads the joint torque information in real time. The lower computer in the control system filters and processes the torque signal and transmits the data to the PC through the serial port. The upper computer executes the robot's control decision, analyzes the current motion trend of the patient relative to the robot, and calculates the target motion angle and speed of each joint based on the interaction torque. It then hands the data over to the servo motor controller so that it can dynamically adjust the angle output of the servo motor according to the desired angle, and guide the robot to execute a specific predetermined trajectory at the desired speed.

[0023] In assisted training mode, set the torque threshold F. start The torque region is divided into [0, F] using the threshold as the dividing point. start ] and [Fstart The feedback parameters acquired by the robot are used as control signals for training. In this method, the control signal can be interpreted as an interaction torque F. Different assist schemes in the assist mode are triggered based on the relationship between the interaction torque and the direction of the robotic leg's movement. When F is less than a preset value or its direction is opposite to the direction of the robotic leg's movement, the lower limb robot will execute a passive motion trajectory; when F is greater than a preset value and the direction of the force is consistent with the direction of the robotic leg's movement, the control system commands the robot to adjust the joint motors to accelerate rotation in the direction of the force.

[0024] According to the control principle of assisted training, the initial velocity V0 in assisted training methods can be understood as the initial step length. To avoid disturbances, the threshold value F of the hip joint is set... start The threshold F for the knee joint is set to 2.5 (Nm). start The value is set to 0.7 Nm. Communication between the PC and the lower-level device is based on the Canopen communication protocol. Figure 1 The specific operation procedure is as follows: First, start the assist mode on the host computer and set the drivers of the two motors to position mode. For the entire motion cycle T, the initial step size is set to ΔS = 0.08T, as follows... Figure 2 During assisted training, the robot's initial position is an upright lower limb position with an initial angle of 0°. The robot is programmed to follow a movement trajectory, with the hip joint's forward leg swing direction defined as positive (maximum angle 26°) and backward leg swing direction as negative (minimum angle -16°), resulting in a joint range of motion [-16, 26]. Similarly, the knee joint's bending direction is positive (maximum angle 80°) and extension direction is negative (minimum angle 0°), with a joint range of motion [0, 80]. The sign of the joint torque is determined by the interaction torque relative to the robot's movement direction. In passive assistance, this is primarily achieved by the robot itself. When the interaction torque is forward, it is considered to have a forward tendency relative to the mechanical leg, resulting in a positive joint torque; conversely, a negative value indicates a downward tendency. Then, the interaction torque and motor position acquisition channels are activated to acquire the robot's joint torque signals and drive motor positions in real time. Based on the feedback signals, the target position of the drive unit is sent at set step intervals to achieve interpolation. When the joint torque is below the set threshold torque and opposite to the direction of movement of the robotic leg, passive assistance training is performed with a fixed stride length ΔS; when the joint torque is in the same direction as the gait movement, the stride length is increased to 2ΔS to increase the training speed, achieving compliant assistance. Based on the principle of assisted training control, the speed control logic is drawn as follows: Figure 3 As shown.

Claims

1. A lower limb assisted movement training method based on a treadmill, characterized in that, The aforementioned assisted training method plans a fixed motion trajectory for the lower limb rehabilitation robot, and the treadmill serves as the training platform to provide an appropriate training speed for the robot's motion trajectory. During robot operation, the position information and interaction torque information of the drive unit are collected and compared, and the assisted training method selects a passive assistance or compliant assistance scheme based on this information. The walking machine is a passive walking machine, which relies on friction to drive the moving platform of the walking machine and adjusts its working speed according to the pre-set motion trajectory of the robot. The motion trajectory has a period of 1s under normal conditions. In order to obtain the tracking curve for the motion trajectory, multi-order trigonometric functions are used to fit the discrete data points in the normal trajectory, and the coefficient that minimizes the residual error is found. Finally, the function formula (1) is determined to be used. The operating cycle of the lower limb rehabilitation robot was set to 4 seconds, and the amplitude coefficient in the standard trajectory was adjusted. frequency coefficient , first phase Adjusting the offset k, we obtain the motion trajectory; (1)； in, Represents the vibration coefficient. Represents the frequency coefficient. Represents the initial phase, and k represents the offset distance; The robot will move along a preset trajectory with a fixed step size. Discretize the data, and during runtime, send the discretized trajectory position to the drive unit. The lower limb assistive training method described above is implemented by a lower limb assistive training system based on a treadmill, which includes a passive treadmill and a lower limb rehabilitation robot; wherein the treadmill provides a platform for the robot to move, and the robot provides an assistive solution; In the motion trajectory planning of the robot in assisted training, the motion data is obtained from OpenSim software, and its motion cycle is 1s. In order to obtain the motion tracking curve for training, the data is processed and put into MATLAB. Multi-order trigonometric functions are used to fit the discrete data points in the normal motion curve, and the coefficients that minimize the residual error are found. Finally, the function formula (1) is determined. Considering that the actual motion speed is low, the motion cycle of the lower limb rehabilitation robot is set to 4s, and the amplitude coefficient in the standard trajectory is adjusted. frequency coefficient , first phase Adjust the offset k to obtain a suitable new motion trajectory; Since the robot's pose during the operation phase is periodic and symmetrical, the robot's motion trajectory throughout the entire cycle can be deduced based on the acquired new gait; the kinematic function equations of the hip joint are as shown in equation (2), and the kinematic function equations of the knee joint are as shown in equation (3). (2)； (3)； in, Kinematic functions representing the hip joint The function represents the motion of the knee joint, where t represents the stride length; For assisted training, the training mode is set to assisted training mode in the human-computer interaction interface. After the mode starts, the torque sensor reads the interactive torque information in real time as a torque signal. The lower computer in the control system filters the torque signal and transmits the data to the PC through the serial port. The upper computer executes the robot's control decision, analyzes the current motion trend of the patient relative to the robot, and calculates the target motion angle and speed of each joint based on the interactive torque information. It then hands the data over to the servo motor controller so that it can dynamically adjust the angle output of the servo motor according to the desired angle, and guide the robot to execute the predetermined trajectory. In assisted training mode, set the torque threshold. The torque region is divided into [0, ..., using the threshold as the dividing point. ]and[ The feedback parameters acquired by the robot are used as control signals for training. In this method, the control signal can be interpreted as interactive torque information F. Different assist schemes in the assist mode are triggered based on the relationship between the interactive torque information F and the direction of movement of the mechanical leg. When F is less than the preset value or the direction is opposite to the direction of movement of the mechanical leg, the lower limb robot will execute a passive motion trajectory. When F is greater than the preset value and the direction of the force is consistent with the direction of movement of the mechanical leg, the control system commands the robot to adjust the joint motors to accelerate rotation in the direction of the force. Based on the control principle of assisted training, the initial velocity In assisted training methods, this is understood as the initial stride length; to avoid disturbance, the hip joint threshold is... The threshold for the knee joint is set at 2.5 Nm. The initial step length is set to 0.7 Nm. Communication between the PC and the lower-level machine is based on the Canopen communication protocol. The specific operation process is as follows: First, start the assist mode on the upper-level machine and set the drivers of the two motors to position mode. For the entire motion cycle T, the initial step length is set to... =0.08T, During assisted training, the robot's initial position is an upright lower limb with an initial angle of 0º. The robot is programmed to follow a motion trajectory, with the hip joint's forward leg swing direction defined as positive and a maximum angle of 26º, and its backward leg swing direction as negative and a minimum angle of -16º, resulting in a joint range of motion [-16, 26]. Similarly, the knee joint's bending direction is positive and a maximum angle of 80º, while its extension direction is negative and a minimum angle of 0º, with a joint range of motion [0, 80]. The sign of the joint torque is determined by the interaction torque information relative to the robot's motion direction. In passive assistance, the robot primarily assists in this process. When the interaction torque information is forward, it is assumed to have a forward motion relative to the robotic leg. If the movement trend is positive, the interaction torque information is positive; otherwise, it is negative. Then, the interaction torque information and motor position acquisition channel is activated to acquire the robot's interaction torque information and drive motor positions in real time. The motor positions are the position information of the drive units. Kinematic functions for the hip and knee joints are established, representing the robot's motion trajectory. During movement, the robot discretizes the set motion trajectory to the motion position of each drive unit based on the step length. Based on the kinematic functions, the target position of the drive units is updated in real time at certain step intervals, i.e., the target position is sent to the driver of each servo motor at fixed step intervals to achieve interpolation. When the interaction torque information is lower than the set threshold torque or the mechanical leg moves in the opposite direction, a fixed step length is used. Perform passive assist training; when the interaction torque information is greater than the set threshold torque and is in the same direction as the gait movement, the stride length will be increased by 2. Increase training speed and achieve adaptive assistance.

Images