System and method for gait phase and terrain recognition for lower extremity assist exoskeletons

Abstract

The application discloses a system and method for gait phase and terrain recognition of lower limb assisting exoskeleton, and belongs to the technical field of lower limb assisting exoskeleton robot perception. The system comprises an inertia detection module, a foot pressure sensor module, a data acquisition and processing module and a central controller. The method divides a gait cycle by using collected inertia data, verifies whether the gait cycle is correct by using collected foot pressure data, then calculates hip joint and knee joint angles of a subject by using corresponding inertia data in the correct gait cycle, and finally determines an angle offset of each joint angle and calibrates by the minimum cost function according to defined kinematic constraints and a cost function, thereby completing gait phase and terrain recognition of the lower limb assisting exoskeleton. The application can correctly recognize gait phases and different terrains, and is used for providing the lower limb exoskeleton, thereby improving coordination and flexibility of device limb movement.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of lower limb assistive exoskeleton robot perception technology, and particularly relates to a system and method for gait phase and terrain recognition of a lower limb assistive exoskeleton. BACKGROUND

[0002] The lower limb assistive exoskeleton is a wearable mechanical device for the human body, which is connected in parallel to the lower limbs of the human body, can be used to enhance or restore the movement ability of the human body, and can also be used for gait rehabilitation of patients with knee joint dysfunction. It can follow the limbs according to the movement intention of the human body and provide assistance or support for the lower limbs.

[0003] The lower limb assistive exoskeleton robot is a wearable lower limb walking bionic mechanical leg, which is centered on the human body, collects the trend of human movement through sensors, and gives joint assistance in the same gait direction as the human body in the aspect of assistance. The current lower limb assistive exoskeleton robot generally collects bioelectric signals through bioelectromyography sensors, electroencephalography sensors, etc., predicts the movement intention of the human body, and then inversely deduces the movement state of the human body through terrain changes to realize the prediction of the movement mode. However, this method has the problems of recognition lag and poor anti-interference ability. Therefore, it is necessary to study a system and method for gait phase and terrain recognition of a lower limb assistive exoskeleton robot to solve the above problems. SUMMARY

[0004] In view of the deficiencies in the prior art, the present application provides a system and method for gait phase and terrain recognition of a lower limb assistive exoskeleton.

[0005] The technical scheme of the present application is as follows:

[0006] A system for gait phase and terrain recognition of a lower limb assistive exoskeleton, comprising an inertia detection module, a foot pressure sensor module, a data acquisition and processing module, and a central controller.

[0007] The inertia detection module is used to detect the inertia data of the thighs, lower legs and tailbones of the wearer in real time, and the inertia data includes angle data and angular velocity data.

[0008] The foot pressure sensor module is used to detect the foot pressure data of the wearer in real time.

[0009] The data acquisition and processing module comprises a microcontroller and a bus communication module. The microcontroller is used to collect and denoise the inertia data and the foot pressure data, then divide the gait cycle according to the preprocessed inertia data, verify whether the gait cycle is correct by using the collected foot pressure data, calculate the hip joint and knee joint angles of the subject by using the corresponding inertia data in the correct gait cycle, and transmit the results to the central controller through the bus communication module.

[0010] The central controller is used for receiving the angles of the hip joint and the knee joint of the subject sent by the data acquisition and processing module; according to the kinematic constraint and the cost function, the angle offset of each joint angle is determined and calibrated by the minimum cost function, and the gait phase and terrain recognition of the lower limb assisting exoskeleton are completed.

[0011] Further, the inertial detection module is used for detecting the angle data and the angular velocity data of the thigh, the shank and the coccyx in real time, and comprises: an inertial sensor A, an inertial sensor B, an inertial sensor C, an inertial sensor D and an inertial sensor E; wherein the inertial sensor A and the inertial sensor B are respectively installed on the left shank outside and the right shank outside of the exoskeleton robot, the inertial sensor C and the inertial sensor D are respectively installed on the left thigh outside and the right thigh outside of the exoskeleton robot, and the inertial sensor E is installed on the coccyx of the exoskeleton robot; the X-axis positive direction of each inertial sensor is vertically upward, and the Y-axis positive direction points to the inside.

[0012] The foot pressure sensor module adopts a foot button sensor, and comprises a foot button sensor E located on the instep and a foot button sensor F located on the heel, and is used for measuring the foot pressure data of the wearer in real time.

[0013] Further, the inertial sensor adopts an XSENS micro-capacitive three-axis angle and angular velocity sensor.

[0014] Based on the above system, the application further provides a method for gait phase and terrain recognition of a lower limb assisting exoskeleton, comprising the following steps:

[0015] Step 1: the wearer wears the system for gait phase and terrain recognition of a lower limb assisting exoskeleton, the inertial data of the wearer's legs and coccyx are detected in real time by the inertial detection module, and the foot pressure data of the wearer are detected in real time by the foot pressure sensor module.

[0016] Step 2: the data acquisition and processing module acquires the inertial data and the foot pressure data; according to the inertial data, the gait cycle is divided, and whether the gait cycle is correct is verified by using the acquired foot pressure data; then the angles of the hip joint and the knee joint of the subject are calculated by using the corresponding inertial data in the correct gait cycle; the specific process is as follows:

[0017] Step 2.1: the data acquisition and processing module carries out denoising pretreatment on the acquired inertial data and foot pressure data.

[0018] Step 2.2: According to the pre-processed inertial data, the angular acceleration data at the corresponding time is calculated from the angle data and the angular velocity data; the first extreme point in the angular acceleration data is extracted as the beginning of a gait cycle, and the second extreme point is extracted as the end of the gait cycle and the beginning of the next gait cycle, so as to divide the gait cycle.

[0019] In order to avoid false positives, the foot pressure data is used to verify whether the gait cycle is accurate; specifically, when the pressure data of the instep or heel of one foot is greater than a threshold value, it is determined that the foot is in a support state, otherwise it is in a swing state; whether the error between the time when the foot enters the support state from the swing state and the beginning time of the gait cycle is within a threshold value is calculated; if it is within the threshold value, the gait cycle division is correct and the next step is performed, otherwise the data in the gait cycle is discarded.

[0020] Step 2.3: The hip joint angle and the knee joint angle of the wearer are calculated according to the inertial data.

[0021] Step 2.4: The hip joint angle and the knee joint angle at each time are sent to the central controller.

[0022] Step 3: The central controller determines the angle offset of each joint angle and calibrates according to the kinematic constraint and the cost function, and completes the gait phase and terrain recognition of the lower limb assisting exoskeleton.

[0023] The pre-processing process in the step 2.1 is filtering by using a Butterworth filter or single threshold multi-threshold activity segment detection.

[0024] The system of the present application comprises an inertial detection module, a foot pressure sensor module, a data acquisition and processing module, and a central controller. The method of the present application divides the gait cycle by using the collected inertial data, verifies whether the gait cycle is correct by using the collected foot pressure data, then calculates the hip joint and knee joint angles of the subject by using the corresponding inertial data in the correct gait cycle, and finally determines the angle offset of each joint angle and calibrates according to the defined kinematic constraint and cost function, and completes the gait phase and terrain recognition of the lower limb assisting exoskeleton. The present application can correctly identify the gait phase and different terrains, and is used to provide the lower limb exoskeleton, thereby improving the coordination and flexibility of the device limb movement. BRIEF DESCRIPTION OF DRAWINGS

[0025] Figure 1 The system structure diagram for the gait phase and terrain recognition of the lower limb assisting exoskeleton in the specific embodiment of the present application;

[0026] Figure 2 The specific flowchart for the gait phase and terrain recognition of the lower limb assisting exoskeleton in the specific embodiment of the present application;

[0027] Figure 3 A phase division diagram for a right foot in a gait cycle in the specific embodiment of the present application;

[0028] Figure 4 A position diagram of the IMU placed on the large and small legs in the specific embodiment of the present application;

[0029] Figure 5 A diagram of the motion trajectory of the thigh and the small leg and the angle change of the knee joint obtained in the specific embodiment of the present application;

[0030] Figure 6 A diagram of the sole pressure sensor module in the specific embodiment of the present application;

[0031] Figure 7 A diagram of the definition of the local coordinate system of the human joint in the specific embodiment of the present application;

[0032] Figure 8 A diagram of the definition of the absolute angle and the relative angle of the human joint in the specific embodiment of the present application. DETAILED DESCRIPTION

[0033] The specific embodiment of the present application will be described in detail below with reference to the accompanying drawings.

[0034] A system for gait phase and terrain recognition of a lower limb assisting exoskeleton, the system structure of which is shown in Figure 1 , and includes an inertia detection module, a sole pressure sensor module, a data acquisition and processing module, and a central controller.

[0035] The inertia detection module is used to detect the inertia data of the wearer's legs and coccyx in real time, and the inertia data includes angle data and angular velocity data.

[0036] The sole pressure sensor module, as shown in Figure 6 , is composed of two pressure sensors in front and behind the foot, and is used to detect the sole pressure data of the wearer in real time.

[0037] The data acquisition and processing module includes a microcontroller and a bus communication module; wherein the microcontroller collects the inertia data and the sole pressure data and pre-processes them by using a Butterworth filter for noise reduction, then divides the gait cycle according to the pre-processed inertia data, verifies whether the gait cycle is correct by using the collected sole pressure data, calculates the hip joint and knee joint angles of the subject by using the corresponding inertia data in the correct gait cycle, and transmits the results to the central controller through the bus communication module. The inertia detection module and the sole pressure sensor module both communicate with the data acquisition and processing module by Bluetooth, and the data acquisition and processing module communicates with the central controller through an RS232 serial port.

[0038] In this embodiment, the inertial detection module comprises: inertial sensor A, inertial sensor B, inertial sensor C, inertial sensor D and inertial sensor E; wherein the inertial sensor A and the inertial sensor B are respectively installed on the left and right sides of the lower leg of the exoskeleton robot, the inertial sensor C and the inertial sensor D are respectively installed on the left and right sides of the thigh of the exoskeleton robot, and the inertial sensor E is installed on the coccyx of the exoskeleton robot; the X-axis positive direction of each inertial sensor is vertically upward, and the Y-axis positive direction points to the inside, as shown in Figure 4 The angle data and angular velocity data during walking of the person are detected in real time by the inertial sensors arranged at the lower limbs.

[0039] In this embodiment, the inertial sensor adopts the micro-capacitive three-axis angle and angular velocity sensor of the MMA7260QT model in XSENS. The micro-capacitive sensor integrates signal conditioning, single-pole low-pass filter and temperature compensation technology, and provides four measurement ranges: ±1.5g, ±2g, ±4g and ±6g. The MMA7260QT has high sensitivity, and when the measurement range of ±l.5g is selected, the sensitivity reaches 800mV / g. The package is 6mmx6mmx1.45mm QFN, the volume is very small, the zero deviation is only ±50mv, and only a small board space is occupied. It has three-axis detection function, which enables portable devices to intelligently respond to changes in position, orientation and movement. Using silicon MEMS technology, it still maintains excellent performance under severe impact and vibration conditions; the start-up time is very short, not more than 0.5s; the current is very small when working normally, less than 0.5mA, and the sensitivity of the horizontal axis is less than 1%.

[0040] In this embodiment, the central controller adopts the NVIDI A Jetson TK1 development mainboard produced by NVIDIA Company. This component can release the powerful ability of GPU for embedded applications, and the rich functional units provide guarantee for the realization and future upgrade and expansion of system functions.

[0041] Based on the above system, the method flow of gait phase and terrain recognition of the lower limb assisting exoskeleton in this embodiment is as shown in Figure 2 The specific process is as follows:

[0042] Step 1: The wearer wears the system for gait phase and terrain recognition of the lower limb assisting exoskeleton. The inertial detection module and the plantar pressure sensor module detect the inertial data and the plantar pressure data of the wearer when completing different gait actions under different road conditions in real time.

[0043] Step 2: The data acquisition and processing module collects the inertial data and the plantar pressure data through wireless communication; the gait cycle is divided according to the inertial data, and the hip joint angle and the knee joint angle at each time are calculated; the specific process is as follows:

[0044] Step 2.1: The data acquisition and processing module denoises and pre-processes the collected inertial data and plantar pressure data using a Butterworth low-pass filter.

[0045] In this embodiment, a Butterworth low-pass filter with zero phase lag is used to denoise and pre-process the inertial data and plantar pressure data, with a cutoff frequency of 2 Hz. This very low frequency filter is used to eliminate any spikes caused by sensor noise.

[0046] Step 2.2: The data acquisition and processing module calculates the angular acceleration data at the corresponding time from the angle data and angular velocity data based on the pre-processed inertial data; extracts the first extreme point in the angular acceleration data as the start of a gait cycle, and the second extreme point as the end of the gait cycle and the start of the next gait cycle, and divides the gait cycle.

[0047] Since the angular velocity and angular acceleration from the swing phase to the support phase are extreme points in the gait phase curve, the gait cycle is divided using the extreme points.

[0048] To avoid false positives, the plantar pressure data is used to verify whether the gait cycle is accurate; specifically, when the pressure data of the sole or heel of one foot is greater than a threshold value, it is determined that the foot is in the support phase, otherwise it is in the swing phase; calculate whether the error between the time when the foot enters the support phase from the swing phase and the start time of the gait cycle is within a threshold value; if it is within the threshold value, the gait cycle division is correct and the next step is performed, otherwise the data in the gait cycle is discarded.

[0049] Step 2.3: The hip joint angle and knee joint angle of the wearer are calculated based on the inertial data.

[0050] The inertial data is used to generate the center of mass (CoM) trajectory using the linear inverted pendulum model (LIPM), and then the trajectory of the hip center (HC) is obtained based on the generated CoM trajectory:

[0051]

[0052]

[0053] h zc (t)=base z +amp z ×{0.5-0.5 cos(4πf)+0.02 sin(4πf)}(3)

[0054] where h xc (t), h yc (t), and h zc(t) represent the trajectory of the hip center in x, y, z directions respectively, x(0), y(0) represent the x coordinate and y coordinate of the hip position at the initial time of a gait cycle respectively, x ′ (0), y ′ (0) represent the velocity of the hip position in x, y directions at the initial time of a gait cycle respectively, p x , p y and p z are the x, y and z coordinates of the hip position, which are updated every time according to the change of the hip position; base z is the height of the hip; amp z is the amplitude of the hip swing; Tc is the time constant; f is the sensor frequency; t is the time; the time constant Tc is defined in (4), where g is the gravity; LipmZ c is the mass center of the human body.

[0055]

[0056] The trajectories of the thigh and shank of the swing leg are obtained by formulas (5)-(7):

[0057]

[0058]

[0059]

[0060] wherein x s (t), y s (t) and z s (t) represent the coordinates of the swing leg in x, y, z directions at time t respectively; Length, Height and Shift are the striking length, striking height and moving distance of the thigh or shank in a gait cycle respectively; p is the time percentage when the swing leg reaches the highest position; T s is a gait cycle time.

[0061] The joint angles of the two legs are further calculated by using inverse kinematics (IK) model.

[0062] The lower limbs of the human body are regarded as rigid rod members and simplified as a five-bar model; the local coordinate systems of the rod members are determined, and the geometric relationship of the rod members is described by the coordinate systems; wherein the local coordinate system of the human joint is defined as Figure 7 shown, and the absolute angle and the relative angle are defined as Figure 8 shown.

[0063] The absolute angle vector q and the relative angle vector q are:

[0064]

[0065] wherein q represents the included angle of the rod and the vertical direction, q1 is the absolute angle of the shank of the support leg; q2 is the absolute angle of the thigh of the support leg; q3 is the absolute angle of the upper limb; q4 is the absolute angle of the thigh of the swing leg; q5 is the absolute angle of the shank of the swing leg. θ represents the relative angle between two adjacent rods, θ1 is the knee joint angle of the support leg; θ2 is the hip joint angle of the support leg; θ3 is the knee joint angle of the swing leg; θ4 is the hip joint angle of the swing leg; θ5 is the posture angle of the upper limb.

[0066] 2 vector spaces satisfy the relationship:

[0067] q T = M θq ·θ T (9)

[0068]

[0069] wherein the superscript T represents transposing the rectangle and converting the absolute angle into the relative angle; in combination with equations (1)-(10), the hip joint angle and the knee joint angle corresponding to the time t are finally obtained.

[0070] Step 2.4: the hip joint angle and the knee joint angle corresponding to the time t are sent to the central controller.

[0071] Step 3: the central controller determines the angle offset of each joint angle and calibrates according to the kinematic constraint and the cost function, and completes the gait phase and terrain recognition of the lower limb assisting exoskeleton according to the minimum cost function.

[0072] The above examples only illustrate the technical concept and characteristics of the present application, and the purpose is to enable those skilled in the art to understand the content of the present application and to implement it, and cannot limit the protection scope of the present application. Any equivalent changes or modifications made according to the spirit and essence of the present application shall be covered within the protection scope of the present application.

Claims

1. A gait phase and terrain recognition system for lower extremity assistive exoskeleton, comprising an inertial detection module, a foot pressure sensor module, a data acquisition and processing module, and a central controller; The inertial detection module is used to detect the inertial data of the wearer's thighs, shanks and tailbone in real time, and the inertial data includes angle data and angular velocity data; The foot pressure sensor module is used to detect the foot pressure data of the wearer in real time; The data acquisition and processing module comprises a microcontroller and a bus communication module; wherein The microcontroller is used to collect and denoise the inertial data and foot pressure data, then divide the gait cycle according to the preprocessed inertial data, and verify whether the gait cycle is correct by using the collected foot pressure data; Then calculate the hip and knee angles of the subject using the corresponding inertial data in the correct gait cycle, and transmit the results to the central controller through the bus communication module; The central controller is used to receive the hip and knee angles of the subject sent by the data acquisition and processing module; According to the kinematic constraint and the cost function, determine the angle offset of each joint angle and calibrate by the minimum cost function, complete the gait phase and terrain recognition of the lower extremity assistive exoskeleton.

2. A gait phase and terrain recognition system for a lower extremity powered exoskeleton as claimed in claim 1, wherein, The inertial detection module is used to detect the angle data and angular velocity data of the thighs, shanks and tailbone in real time, which includes inertial sensor A, inertial sensor B, inertial sensor C, inertial sensor D and inertial sensor E; Wherein, the inertial sensor A and the inertial sensor B are respectively installed on the left shank and the right shank of the exoskeleton robot, the inertial sensor C and the inertial sensor D are respectively installed on the left thigh and the right thigh of the exoskeleton robot, and the inertial sensor E is installed on the tailbone of the exoskeleton robot; The X axis positive direction of each inertial sensor is vertically upward, and the Y axis positive direction points to the inside; The foot pressure sensor module adopts foot button sensor, including foot button sensor E located on the instep and foot button sensor F located on the heel, which is used to measure the foot pressure data of the wearer in real time.

3. A gait phase and terrain recognition system for a lower extremity power assist exoskeleton as claimed in claim 2, wherein, The inertial sensor adopts XSENS micro-capacitive three-axis angle and angular velocity sensor.

4. A gait phase and terrain recognition method of a lower extremity assistive exoskeleton, characterized by, The steps include: Step 1: the wearer wears the gait phase and terrain recognition system for lower extremity assistive exoskeleton, detects the inertial data of the wearer's legs and tailbone in real time through the inertial detection module, and detects the foot pressure data of the wearer through the foot pressure sensor module; Step 2: the data acquisition and processing module collects the inertial data and foot pressure data; According to the inertial data, divide the gait cycle, and verify whether the gait cycle is correct by using the collected foot pressure data; Then calculate the hip and knee angles of the subject using the corresponding inertial data in the correct gait cycle; The specific process is as follows: Step 2.1: the data acquisition and processing module denoises and preprocesses the collected inertial data and foot pressure data; Step 2.2: According to the pre-processed inertial data, the angular acceleration data at the corresponding time is calculated from the angle data and the angular velocity data; the first extreme point in the angular acceleration data is extracted as the beginning of a gait cycle, and the second extreme point is extracted as the end of the gait cycle and the beginning of the next gait cycle, so as to divide the gait cycle; In order to avoid false positives, the foot pressure data is used to verify whether the gait cycle is accurate; specifically, when the pressure data of the sole or heel of one foot is greater than a threshold value, it is determined that the foot is in a support state, otherwise it is in a swing state; whether the error between the time when the foot enters the support state from the swing state and the beginning time of the gait cycle is within a threshold value is calculated; If it is within the threshold value, the gait cycle division is correct and the next step is performed, otherwise the data in the gait cycle is discarded; Step 2.3: The hip joint angle and the knee joint angle of the wearer are calculated according to the inertial data; Step 2.4: The hip joint angle and the knee joint angle at each time are sent to the central controller; Step 3: The central controller determines the angular offset of each joint angle and calibrates according to the kinematic constraint and the cost function, and completes the gait phase and terrain recognition of the lower limb assistive exoskeleton by the minimum cost function.

5. The gait phase and terrain recognition method of a lower extremity assistive exoskeleton according to claim 4, wherein, The pre-processing process in step 2.1 is filtering by using a Butterworth filter or single-threshold multi-threshold activity segment detection.

Images