A multi-dimensional lower limb parameter detection method and system fusing space-time features

Abstract

This application discloses a method and system for detecting multidimensional lower limb parameters by fusing spatiotemporal features. The method includes the following steps: using a wearable device to collect the wearer's lower limb motion information; the lower limb motion information includes position information, posture information, and multidimensional plantar force information; estimating lower limb spatiotemporal parameters and gait events based on the lower limb biomechanical model and the lower limb motion information; and obtaining multidimensional plantar force data with fused spatiotemporal features based on the multidimensional plantar force measurement model and the lower limb motion information. This method can accurately estimate lower limb spatiotemporal parameters and human trajectory, and realize the solution and mapping of three-dimensional plantar force space.

Description

Technical Field

[0001] This application relates to the field of physiological parameter detection technology, and in particular to a method and system for detecting multidimensional lower limb parameters that integrates spatiotemporal features. Background Technology

[0002] like Figure 1 As shown, human motion can be described using planes and axes, consisting of the coronal plane, sagittal plane, and horizontal plane. These planes intersect at the body's center of mass. Each plane rotates around an axis perpendicular to it, the direction of which is determined by the right-hand rule: when the four fingers are bent along the direction of motion, the direction the thumb points indicates the axis of rotation. Based on their respective planes, the axes of rotation can be categorized as the coronal axis (left-right direction), the sagittal axis (front-back direction), and the vertical axis (up-down direction).

[0003] Human movement can be described as movement around a parallel axis in a plane parallel to the aforementioned planes and axes. Movement in the coronal plane revolves around a sagittal axis perpendicular to the coronal plane; common somatic movements include arm and leg abduction. Sagittal plane movement occurs around a coronal axis perpendicular to the sagittal plane; common somatic movements include forward and backward arm swinging and flexion and extension of the lower limb joints. Horizontal plane movement occurs around a vertical axis perpendicular to the horizontal plane; common somatic movements include head shaking and internal / external rotation of the legs.

[0004] Gait cycles are the natural cyclical movements of the human body during walking, describing the complete process from the heel of one foot striking the ground to the foot striking the ground again. The main motion characteristics of the lower limbs during walking are concentrated in the sagittal plane. Common biomechanical models of lower limb movement include the seven-bar and four-bar models. Although the seven-bar and four-bar models based on human lower limb models can describe the lower limb motion state relatively well, for lower limb kinematic sensing acquisition, these two models rely on kinematic information from multiple limbs. Lower limb position and attitude estimation based on inertial sensors suffers from accumulated integral errors and measurement errors caused by magnetic field interference. Summary of the Invention

[0005] The main technical problem addressed in this application is to provide a multidimensional lower limb parameter detection method and system that integrates spatiotemporal features, thereby solving the measurement errors caused by the accumulation of integral errors and magnetic field interference in existing lower limb position and attitude estimation based on inertial sensors.

[0006] To address the aforementioned technical problems, this application provides a method for detecting multidimensional lower limb parameters by fusing spatiotemporal features. The method includes the following steps: using a wearable device to collect the wearer's lower limb movement information; the lower limb movement information includes position information, posture information, and multidimensional plantar force information; estimating lower limb spatiotemporal parameters and gait events based on a lower limb biomechanical model and the lower limb movement information; and obtaining multidimensional plantar force data fusing spatiotemporal features based on a multidimensional plantar force measurement model and the lower limb movement information.

[0007] This application also provides a multidimensional lower limb parameter detection system that integrates spatiotemporal features. The system includes: an information acquisition unit for acquiring the wearer's lower limb movement information, including position information, posture information, and multidimensional plantar force information; a first analysis unit for estimating lower limb spatiotemporal parameters and gait events based on a lower limb biomechanical model and the lower limb movement information; and a second analysis unit for obtaining multidimensional plantar force data that integrates spatiotemporal features based on a multidimensional plantar force measurement model and the lower limb movement information.

[0008] The beneficial effects of this application are as follows: This invention discloses a method and system for detecting multidimensional lower limb parameters by fusing spatiotemporal features. The method includes the following steps: using a wearable device to collect the wearer's lower limb motion information; the lower limb motion information includes position information, posture information, and multidimensional force information of the sole; estimating the spatiotemporal parameters and gait events of the lower limb based on the lower limb motion biomechanics model and the lower limb motion information; and obtaining multidimensional force data of the sole by fusing spatiotemporal features based on the multidimensional force measurement model of the sole and the lower limb motion information. This method can accurately estimate the spatiotemporal parameters of the lower limb and the human body trajectory, and realize the solution and mapping of the three-dimensional force space of the sole. Attached Figure Description

[0009] Figure 1 It is a planar diagram of human anatomy that describes human movement;

[0010] Figure 2 This is a flowchart of an embodiment of a multidimensional lower limb parameter detection method that integrates spatiotemporal features according to this application;

[0011] Figure 3 This is a schematic diagram of the composition of a multidimensional force sensing system on the sole of the foot in one embodiment of a multidimensional lower limb parameter detection method that integrates spatiotemporal features according to this application;

[0012] Figure 4 This is a schematic diagram of multiple coordinate systems involved in a multi-dimensional lower limb parameter detection method that integrates spatiotemporal features in one embodiment of the lower limb multi-sensor coordinate system of this application;

[0013] Figure 5 This is a schematic diagram of the measurement model of the lower limb spatiotemporal parameter sensing system in one embodiment of a multidimensional lower limb parameter detection method that integrates spatiotemporal features according to this application;

[0014] Figure 6 This is a schematic diagram of a two-dimensional linear inverted pendulum model in one embodiment of a multidimensional lower limb parameter detection method that integrates spatiotemporal features according to this application;

[0015] Figure 7 This is a schematic diagram of the state of the lower limb supporting foot at the moment of switching in one embodiment of a multidimensional lower limb parameter detection method that integrates spatiotemporal features according to this application;

[0016] Figure 8 This is a schematic diagram of the walking gait with the lower limb supporting foot switching twice in one embodiment of a multidimensional lower limb parameter detection method that integrates spatiotemporal features according to this application;

[0017] Figure 9 This is a schematic diagram of a lower limb biomechanical model in one embodiment of a multidimensional lower limb parameter detection method that integrates spatiotemporal features according to this application;

[0018] Figure 10 This is a schematic diagram of the features of two gait events in an embodiment of a multidimensional lower limb parameter detection method that integrates spatiotemporal features according to this application;

[0019] Figure 11 This is a gait cycle segmentation and gait parameter estimation diagram in one embodiment of a multidimensional lower limb parameter detection method that integrates spatiotemporal features according to this application;

[0020] Figure 12 This is a schematic diagram illustrating the relationship between gait cycle and foot position and speed in an embodiment of a multidimensional lower limb parameter detection method that integrates spatiotemporal features according to this application;

[0021] Figure 13 This is a schematic diagram of gait parameter estimation within one gait cycle in an embodiment of a multidimensional lower limb parameter detection method that integrates spatiotemporal features according to this application;

[0022] Figure 14 This is a schematic diagram of the features of two gait events in an embodiment of a multidimensional lower limb parameter detection method that integrates spatiotemporal features according to this application;

[0023] Figure 15 This is a diagram showing the relationship between foot position and gait cycle in the XOY plane of the world coordinate system in one embodiment of a multidimensional lower limb parameter detection method that integrates spatiotemporal features according to this application. Detailed Implementation

[0024] To facilitate understanding of this application, a more detailed description is provided below with reference to the accompanying drawings and specific embodiments. Preferred embodiments of this application are shown in the drawings. However, this application can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the disclosure of this application.

[0025] It should be noted that, unless otherwise defined, all technical and scientific terms used in this specification have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the scope of this application. The term "and / or" as used in this specification includes any and all combinations of one or more of the associated listed items.

[0026] Figure 2 An embodiment of the multidimensional lower limb parameter detection method integrating spatiotemporal features of this application is shown, including:

[0027] Step S1: Use wearable devices to collect the wearer's lower limb movement information.

[0028] It should be noted that lower limb movement information includes position information, posture information, and multi-dimensional force information on the sole of the foot.

[0029] Furthermore, the wearable device includes a lower limb spatiotemporal parameter sensing system (Shank-mounted Ranging-inertial Odometry, Shank-RIO), which includes a lidar ranging sensor and an inertial sensor. The lidar ranging sensor is used for distance measurement, and the inertial sensor is used for attitude detection.

[0030] like Figure 3As shown, the wearable device also includes a multi-dimensional force sensing system for the sole of the foot. This system comprises two flexible array sensors, each including multiple sensing units arranged in an array. From top to bottom, each sensing unit includes a loading block 11, a first electrode plate 12, a dielectric layer 13, a second electrode plate 14, and a substrate 15. The loading block 11 is tightly fitted to the first electrode plate 12 to transmit external force to it, causing deformation of the dielectric layer 13. The first electrode plate 12 and the second electrode plate 14 are parallel to each other. The first electrode plate 12 includes a plurality of first working electrodes 21 symmetrically distributed in an array, with each first working electrode 21 in the first direction connected in series. The second electrode plate 14 includes a plurality of second working electrodes 41 symmetrically distributed in an array, with each second working electrode 41 in the second direction connected in series, and the first and second directions being orthogonal. The center positions of each first working electrode 21 and its corresponding second working electrode 41 coincide. The dielectric layer 13 is used to separate the first electrode plate 12 and the second electrode plate 14. The substrate 15 is tightly attached to the second electrode plate 14. Two flexible array sensors are fixedly connected through corresponding substrates 15. One flexible array sensor (foot-shoe interface) is used to contact the sole of the foot to detect the multi-dimensional pressure of the sole area, with the measurement unit being kPa. The other flexible array sensor (shoe-ground interface) is used to contact the ground to detect the overall three-dimensional ground reaction force, with the measurement unit being N.

[0031] It should be noted that the foot is partially stationary during the support phase, a state known as Zero Velocity State (ZVS). Theoretically, the sensor velocity should be zero in ZVS. However, due to sensor noise and errors, the directly measured velocity value may not be precisely zero, leading to significant errors in position estimation. Furthermore, the foot experiences a large impact during the heel-to-ground contact event at the start of the support phase, which can interfere with the accelerometer located on the foot, further influencing position estimation. The lower leg, however, is in a swinging state throughout the support phase, not in a ZVS state. Compared to the foot, the lower leg is more suitable for sensor placement for position measurement. Additionally, optical rangefinders placed on the foot may fail to measure over the relatively short distance during foot contact with the ground. Therefore, for the acquisition scheme of the lower limb spatiotemporal parameter sensing system, placing the system on the lower leg is more reasonable and better improves the accuracy of lower limb position and attitude estimation.

[0032] Therefore, before step S1, the steps include: wearing the lower limb spatiotemporal parameter sensing system on the wearer's calf and wearing the foot multidimensional force sensing system on the wearer's foot.

[0033] like Figure 4As shown, this application's lower limb multi-sensor coordinate system utilizes multiple coordinate systems. World coordinate system {O} G} is an absolute coordinate system, {O G The starting point of} is the initial position of the lower limb multi-sensor coordinate system. The coordinates of the lower limb multi-sensor coordinate system are all based on {O}. G The origin of the coordinate system {O} determines their respective positions. B} is the Frenet coordinate system, used to describe the position of the lower limb multi-sensor system relative to the body center. Lower limb spatiotemporal parameter sensor system coordinate system {O D Located on the lower leg, belonging to {O} B}. Combining Figure 15 Foot coordinate system {O F1} and {O F2} are used to describe the multi-dimensional force sensing systems of the left and right soles of the feet, respectively. F1} and {O F2 The origins of} are located at {O} B The foot positions on both sides in the lower limb biomechanics model of}. F1} and {O F2} Each includes its own foot-shoe interface coordinate system and shoe-ground interface coordinate system, where {O F1 The foot-shoe interface coordinate system and the shoe-ground interface coordinate system are {O}, respectively. inole1} and {O sole1}, {O F2 The foot-shoe interface coordinate system and the shoe-ground interface coordinate system are {O}, respectively. inole2} and {O sole2}, and the foot-shoe interface coordinate system {O inole1} / {O insole2} and shoe-ground interface coordinate system {O sole1} / {O sole2 The coordinate system {O} may also include the corresponding sensing unit coordinate system. U}

[0034] like Figure 5 As shown, for the measurement model of the lower limb spatiotemporal parameter sensing system, the inertial sensor Rotation[R]∈R is used. 3×3 The attitude information is used to perform vector decomposition on the distance information of the lidar ranging sensor Distance[d].

[0035] Considering that both the inertial sensor and the lidar ranging sensor are rigidly fixed, it can be assumed that the two sensors in Shank-RIO possess the same attitude information. The attitude information detected by the inertial sensor can be solved using the Mahony algorithm. The Mahony algorithm utilizes the calculation error after converting the output values ​​of the magnetometer and accelerometer to the reference physical values ​​(gravitational acceleration and magnetic field strength) in the world coordinate system. It calculates the attitude data by taking the outer product of the accelerometer output value and the theoretical value, and vice versa. After obtaining the control errors of the accelerometer and magnetometer, the feedback controller compensates for the gyroscope error, and the attitude data is obtained by integrating the gyroscope data.

[0036] First, the quaternion q0 is initialized to [1, 0, 0, 0]. T Integral error e int =[0, 0, 1] T Then, the accelerometer error e is calculated. a The output values ​​of the accelerometers are a, a, and a, respectively. x a y a z The values ​​of gravitational acceleration after transforming from the world coordinate system to the body coordinate system with the inertial sensor as the origin are v. x v y v z :

[0037] ;

[0038] in, Let g be the rotation matrix from the world coordinate system to the body coordinate system. The gravitational acceleration g in the body coordinate system after coordinate transformation is g_t. b .

[0039] ;

[0040] a b This is a vector representation of the accelerometer output value, expressed by g. b and a b Perform the cross product of vectors to obtain the cross product e. a This refers to the error control term of the accelerometer:

[0041] .

[0042] Furthermore, the output values ​​of the magnetometer in the body coordinate system are m x m y m z Transform it into a numerical value h in the world coordinate system. x h y h zSince the x-axis of the world coordinate system is aligned with true north, the projection of the magnetometer onto the XOY plane is transformed to the x-axis:

[0043] ;

[0044] ;

[0045] ;

[0046] Where b is the reference value of the magnetometer in the world coordinate system, which needs to be transformed to the body coordinate system and multiplied by the magnetometer's output value to form a vector outer product:

[0047] ;

[0048] Among them, w x w y w z This is the reference output value of the magnetometer. Magnetometer control error term e m The calculation formula is as follows:

[0049] .

[0050] Then add the accelerometer error e a With magnetometer error e m After merging, the total error is:

[0051] .

[0052] After calculating the total error e, the Mahony algorithm will calculate the angular velocity correction value δ through the PI controller, as shown below:

[0053] ;

[0054] .

[0055] After obtaining the angular velocity correction value δ, the corrected angular velocity value ω can be obtained by the following formula:

[0056] ;

[0057] This allows for the updating of the quaternion q:

[0058] .

[0059] Attitude calculation based on inertial sensors yields the quaternion quat = [x, y, z, w], and the rotation matrix R of the lower limb spatiotemporal parameter sensing system can be calculated.

[0060] .

[0061] In this system, the distance data Range obtained by the lidar ranging sensor represents the distance between the lower limb spatiotemporal parameter sensing system and the ground. To decompose the scalar Range into XYZ directions, a reference vector V is established in the lower limb spatiotemporal parameter sensing system. ref =[0, 0, 1] T V ref Direction and {O D The Z-axis direction is consistent with that of the lower limb spatiotemporal parameter sensing system relative to V. ref When an attitude change occurs, rotation matrices R and V can be used. ref The product directly decomposes the range into Ranges. xyz =[R x , R y ,R z Range xyz The calculation method is as follows:

[0062] .

[0063] In this embodiment, the attitude information calculated by the inertial sensor is used to perform vector decomposition on the position information collected by the lidar ranging sensor. The three-dimensional distance after vector decomposition can be represented as the spatial position of the lower limb spatiotemporal parameter sensing system, which can avoid the accumulation of integration error of the inertial sensor.

[0064] Step S2: Estimate the spatiotemporal parameters and gait events of the lower limbs based on the lower limb biomechanical model and lower limb motion information.

[0065] It should be noted that in the Linear Inverted Pendulum Model (LIPM), the lower limbs are simplified to a problem of a single center of mass moving on an inverted pendulum. When the lower limbs are in a single support phase, the dynamic model of the lower limbs can be simplified using an inverted pendulum. LIPM assumes that the mass of the legs and feet is zero, and all the mass is concentrated at a single center of mass in the hip and waist area. During walking, the motion of this center of mass can be regarded as the motion of an inverted pendulum.

[0066] like Figure 6 As shown, in the sagittal plane, the linear inverted pendulum model consists of a center of mass (CoM) and a stretchable, lightweight leg. The inputs to the inverted pendulum include the torque acting at the fulcrum and the stretching force at the joint along the leg link. The dynamics of the inverted pendulum can be described by the following two differential equations:

[0067] .

[0068] Assuming the input torque τ=0, the LIPM is in an unstable state. The center of mass CoM is kept at a constant height by controlling the expansion force f. The definition of the expansion force f is as follows:

[0069] .

[0070] The center of mass CoM moves horizontally under the action of the stretching force f. After the vertical component of the stretching force f is balanced by gravity, its horizontal component still exists. This component causes the center of mass CoM to accelerate horizontally. The corresponding equation of motion can be expressed as:

[0071] .

[0072] Formula Substitution have to:

[0073] ;

[0074] Where x and z are the coordinates of the center of mass CoM of the inverted pendulum on the x-axis and z-axis, respectively, and the differential equation of the horizontal motion of the center of mass CoM is:

[0075] .

[0076] When z is a constant, the differential equation is solved to obtain x:

[0077] ;

[0078] ;

[0079] ;

[0080] in, The constant is determined by the height of the center of mass and the acceleration due to gravity. and Let represent the position and acceleration of the center of mass at time zero, respectively. Given the initial conditions and the target state at a certain time, from the formula... and The equation relating the two states above is obtained as follows:

[0081] ;

[0082] ;

[0083] Where τ is the time of motion of the centroid between the two states, denoted by T. c Multiply by formula and with formula Summing yields:

[0084] .

[0085] Furthermore, we can obtain τ:

[0086] .

[0087] Similarly, using the formula Multiply by T c and with formula The difference can be obtained by taking the subtraction:

[0088] .

[0089] With a sufficiently large initial velocity, the direction of motion of the center of mass CoM remains unchanged. In this case, the velocity of the center of mass CoM is minimized at the potential energy vertex, at which point the position of the center of mass CoM is directly above the fulcrum. Now, let's consider the formula... Multiplying both sides by x and integrating, we get:

[0090] ;

[0091] This result can be expressed as:

[0092] ;

[0093] In the formula on the left, the first term is kinetic energy, the second term is assumed potential energy, and the total energy E is the orbital energy. This demonstrates that the orbital energy of the linear inverted pendulum is conserved during its motion.

[0094] like Figure 7 As shown, the relationship between foot switching and inverted pendulum walking motion is explored by utilizing the instantaneous state of the supporting foot during the switching process. In the figure, s represents the stride length, and x... f v represents the horizontal position of the center of mass relative to the previous support point. f Let v be the velocity of the center of mass during the support phase transition. To simplify calculations, the final velocity of the support phase at the previous moment and the initial velocity of the support phase at the current moment are both v. f The orbital energies before and after the switching are E1 and E2, respectively.

[0095] ;

[0096] .

[0097] The final velocity of the previous support phase and the initial velocity of the new support phase are both v. f Under the premise that the formula and Eliminate v together f Tangent conditions can be obtained:

[0098] ;

[0099] This allows us to obtain the switching speed v. f for:

[0100] .

[0101] like Figure 8 As shown, in this mode, the supporting foot switches twice to complete one step of walking. At this time, three orbital energies are required. At the beginning of the walking phase, the orbital energy E0 can be obtained based on the initial position of the centroid CoM:

[0102] .

[0103] During the motion between the first supporting foot and the second supporting foot, the orbital energy E1 is determined by the velocity v1 of the center of mass CoM when it passes above the support point:

[0104] .

[0105] During the final stage of the walk, the orbital energy E2 is generated by the final stopping position x of the center of mass CoM. e Decision made:

[0106] .

[0107] Supporting the conditions for the first switch x f0 Based on E0 and E1, the condition x that supports the second switch can be obtained. f1 As can be obtained from E1 and E2, walking motion can be achieved by controlling the movement of the non-supporting foot to meet the switching conditions.

[0108] Furthermore, the capture point is determined to assess and control the stability of the human lower limb gait. The capture point refers to the position where the swing leg should land at the end of the current gait cycle to achieve the desired orbital energy at the start of the next gait cycle. The capture point is calculated based on the current orbital energy E1 and the target orbital energy E2. Orbital energy is a parameter describing the motion state of the human lower limb's center of mass; it is conserved in the absence of external disturbances. By controlling the switching time of the supporting leg and the landing point of the swing leg, the orbital energy can be controlled, thereby controlling the motion state of the human lower limb. The capture point x is calculated. cp The formula can be expressed as:

[0109] ;

[0110] Where, x t It is the current position of the center of mass. Here, x is the velocity of the center of mass, z is the height of the center of mass, g is the acceleration due to gravity, and E2 is the target orbital energy. When E2 > 0, the initial orbital energy is large enough that the LIPM can continue moving beyond the point of maximum potential energy; when E2 < 0, the initial orbital energy is insufficient, and the LIPM's velocity drops to 0 before reaching the point of maximum potential energy, causing it to begin moving in the opposite direction; when E2 = 0, the calculated x... cp This will cause the inverted pendulum model to have a center of mass velocity of 0 at the point of highest potential energy after taking a step, and it will no longer move forward or backward, that is, it will reach a state of rest and equilibrium after taking a step.

[0111] Therefore, under the premise that the human lower limbs have the required momentum to maintain stability and forward movement, i.e., E2 ≥ 0, the human lower limbs can complete continuous forward walking. In continuous forward walking, the heel contact event represents the start of the supporting foot switching process. Therefore, under the condition that E2 ≥ 0, x... cp This corresponds to the heel strike event, which also represents the maximum value of the foot in the forward direction. In this state, the human lower limbs can smoothly complete the switching of the supporting foot for continuous forward walking.

[0112] In the lower limb dual-sensor acquisition scheme of this application, LIPM can serve as a kinematic model applicable to this acquisition scheme. It can estimate and detect lower limb kinematic parameters and gait events by acquiring positional information of the lower leg and foot. Compared to the thigh, which is affected by knee joint movement, the kinematic information of the lower leg and foot can more directly represent lower limb movement and is more consistent with the model assumptions of LIPM.

[0113] It should be noted that the LIPM-based dual-sensor acquisition scheme for the lower limbs has the problem of inaccurate measurement in the AP direction. The rotation center of the lower limb in LIPM is the centroid CoM, while the actual rotation center of the human lower limb is the greater trochanter, and the greater trochanter is a certain distance away from the centroid CoM. This will introduce measurement errors in the estimation of the foot's position in the AP direction.

[0114] like Figure 9 As shown, based on the walking characteristics of the human lower limbs, this application designs a lower limb kinematics model based on LIPM. The lower limb kinematics model includes a model center of mass C, a first rotation axis, a second rotation axis, a first retractable rigid rod, and a second retractable rigid rod. The model center of mass C is the center of the first and second rotation axes, and is also the geometric center of the body. During the wearer's walking process, the model center of mass C will move and is not always located at the center of the body. The first rotation axis is connected to the first retractable rigid rod, and the second rotation axis is connected to the second retractable rigid rod. The first rotation axis corresponds to the wearer's left greater trochanter, and the second rotation axis corresponds to the wearer's right greater trochanter. The first retractable rigid rod corresponds to the wearer's left leg (blue), and the second retractable rigid rod corresponds to the wearer's right leg (red). Wherein, {O BThe origin of {O} is located at the center between the left and right greater trochanters. D dL and dR under} are transformed into {O B}middle B pL and B pR is calculated using the following formula:

[0115] ;

[0116] Wherein, H represents the height of the left / right greater trochanter from the ground, and p0 represents the distance between the left / right greater trochanter and the geometric center of the body. Both of these parameters need to be measured on the wearer before use.

[0117] Therefore, this method also includes the steps of measuring the height of the wearer's left / right greater trochanter from the ground, and measuring the distance between the left / right greater trochanter and the geometric center of the body.

[0118] Furthermore, the location information includes the foot position, and the posture information includes the foot velocity. The gait events in this application are generated using a lower limb biomechanics model, from {O} B The position of the two feet in} and The corresponding features are extracted and identified.

[0119] like Figure 10 and Figure 12 As shown, taking the right leg as an example, the characteristics of two gait events are illustrated. Among them, Figure 10 (a) in the diagram shows the state of {O} B The foot in the} is at its maximum position in the forward direction. The corresponding heel strike (HS) event; Figure 10 (b) shows the result in {O} B The foot in the} has the maximum backward velocity The corresponding toe-off (TO) event. The HS event marks the beginning of the support phase, characterized by the foot's position relative to {O}. B At its furthest point in the forward direction, the swing phase has been completed and it is preparing to enter the support phase. Reaching its maximum value. The TO event marks the end of the support phase, occurring when the foot is relative to {O}. B The maximum speed is in the rearward direction because the supporting leg needs to accelerate backward to reach its maximum in order to push off the foot and complete the toes leaving the ground.

[0120] It should be noted that the HS event and TO event are the most basic events in gait periodization. In addition to the HS event and TO event, this application further extracts the RLA eight-phase method. and Features of [the text].

[0121] like Figure 11 and Figure 12 As shown, the following gait events are also included. First, the event marking the end of the initial support phase (Loadresponse, LR) indicates that the foot is in full contact with the ground, signifying the beginning of full weight-bearing. This also corresponds to the occurrence of the contralateral foot TO event, marking the end of the double support phase and the beginning of the single support phase. Second, the event marking the end of the middle support phase (Heel off, HO) indicates that the supporting leg is about to enter the final support phase. The middle indicates the maximum speed of the support foot. At this point, the supporting foot completes forward support and the heel begins to leave the ground. Furthermore, the event marking the end of the initial swing phase (Feet adjacent, FA) represents the point where the swinging foot is closest to the opposite supporting foot. The middle represents the maximum speed of the swinging foot. Finally, the event marking the end of the mid-cycle (Tibia vertical, TV) indicates that the swing foot is in a vertical position and is about to begin deceleration and braking, entering the end-cycle phase. The middle represents the maximum speed of the swinging foot. The first derivative is negative.

[0122] In summary, this application uses and Feature values ​​are used to segment RLA gait cycles. Therefore, it is possible to identify and This enables the detection of gait events.

[0123] In this embodiment, the time interval between any two HS events is defined as the single-step time (T). step The time interval between two HS events on a single foot is the stride time (T). stride The number of heels per minute (HS) is expressed as step frequency, and the longitudinal straight-line distance between the two points where the left and right heels successively touch the ground during walking is the step length (Range). step The longitudinal straight-line distance between two consecutive heel strikes on the same side is the stride length. stride (Stride), also known as stride length, is equivalent to the sum of the left and right strides.

[0124] like Figure 13 As shown, the step size (Range) step The range is represented by the displacement in the AP direction between two HS events of the left and right feet. Therefore, the step size (Range) can be calculated by the displacement between the HS events of the same-side foot and the HS events of the opposite-side foot. stepAt this point, the HS event of the opposite foot corresponds to the SW-Pre (initial swing) event of the same side foot. The right step length corresponding to the right foot can be obtained by subtracting the position of the HS event from the position of the SW-Pre event of the right foot.

[0125] .

[0126] Combination Figure 11 Step length (Range) stride The definition of ) is the distance in the AP direction between two consecutive HS events of the same side foot, which is equivalent to the sum of the left and right foot lengths, i.e., the left foot length. and right step length The sum of:

[0127] .

[0128] In addition, stride length (Range) stride It can also be obtained by observing the movement of the ipsilateral foot in the AP direction. In calculating the stride length (Range)... stride When the event occurs, the displacement from the HS event to the TO event and from the SW-F (swinging foot crosses the opposite supporting foot) event to the next HS event needs to be accumulated, using the following formula:

[0129] ;

[0130] Here, the HS+1 event represents the next HS event on the same side of the foot. After the toes leave the ground at the end of the support phase, the foot will swing backward a certain distance due to inertia. Therefore, the foot displacement from the TO event to the SW-F event is not considered forward motion.

[0131] Step speed can be measured by stride length (range). stride (Range step ) and step time (single step time) T stride (T step The ratio of ) is calculated and can be expressed as:

[0132] .

[0133] Combination Figure 1 and Figure 11 Step width refers to the distance in the ML direction between the HS events of the left and right feet in two consecutive steps. In other words, step width can be expressed as:

[0134] .

[0135] Step S3: Based on the multidimensional force measurement model of the foot and lower limb motion information, obtain multidimensional force data of the foot that integrates spatiotemporal features.

[0136] In this embodiment, a lower limb positioning model also needs to be established, which includes a linear motion model and a turning motion model.

[0137] like Figure 14 As shown, the lower limb biomechanical model proposed in this application is located in {O}. B Let FRENITION be the Frenet coordinate system. Within the Frenet coordinate system, the motion of an object relative to the origin of the Frenet coordinate system is linear motion. However, the rotation of the Frenet coordinate system relative to the world coordinate system requires modeling the moving object within the Frenet coordinate system to obtain its rotation matrix relative to the world coordinate system.

[0138] like Figure 14 As shown in (a), the linear motion model includes the step length of the motion trajectory after the alternating push of both feet.

[0139] Combination Figure 1 and Figure 4 , {O B} relative to {O G There are no pitch and roll angles, only yaw angles. The following describes a step-size-based steering motion model. In {O B In the process, the body's displacement in the AP direction is determined by the bilateral step length. According to the proposed lower limb kinematics model, after each support phase, the model's center of mass O... B Move in the direction of AP, the distance of which is the step length. By alternately pushing with both feet, the body can achieve forward movement in the direction of AP.

[0140] like Figure 14 As shown in (b), during turning and walking, the degree of yaw in the body's trajectory can be described by curvature. Curvature represents the degree of bending of a curve at a certain point. According to the formula describing the radius of curvature... Therefore, Δs is the trajectory of the body in {O}. G The arc length generated when the body rotates is Δα, which is the angle of rotation of the body's trajectory around the center of rotation. The forward movement of the body in the AP direction is accomplished by the alternating push of both feet. When the stride lengths of the two sides are inconsistent, the forward movement of the body in the AP direction will be deflected to maintain walking balance.

[0141] It should be noted that the curvature generated when an object bends can be obtained by the ratio of the difference in deformation between the upper and lower surfaces to their thickness. Applied to this application, the curvature of the wearer's body's forward motion trajectory in the AP direction is... This can be obtained by the ratio of the difference in step length between the two sides to the step width. Therefore, the turning motion model includes the trajectory curvature, average rotation angle, rotation matrix of the body coordinate system relative to the world coordinate system, and the positions of the model's center of mass, left foot, and right foot in the world coordinate system after the alternating push of the two feet.

[0142] Specifically, the curvature of the wearer's forward motion trajectory in the sagittal axis is obtained by using the ratio of the difference in stride length on both sides to the stride width. The trajectory curvature can be expressed as:

[0143] ;

[0144] in, Indicates the stride length of the left foot; Indicates the stride length of the right foot; This represents the stride width between the left and right feet. The average rotation angle is the product of the average stride length of the left and right feet and the trajectory curvature, which can be expressed as:

[0145] ;

[0146] {O B} relative to {O G The rotation matrix of} can be represented as:

[0147] ;

[0148] The position of the model's centroid in the world coordinate system can be represented as:

[0149] ;

[0150] in, This represents the coordinates of the model's centroid in the world coordinate system; Indicates stride length; H indicates the height of the wearer's left / right greater trochanter from the ground;

[0151] The positions of the left and right feet in the world coordinate system can be represented as follows:

[0152] ;

[0153] ;

[0154] in, and These represent the coordinates of the left and right feet in the world coordinate system, respectively. , and These represent the straight-line distances between the left foot and the model's center of mass along the x, y, and z axes of the body coordinate system, respectively. , and These represent the straight-line distances between the right foot and the model's center of mass along the x, y, and z axes of the body coordinate system, respectively.

[0155] like Figure 14 and Figure 15As shown, a multi-dimensional force measurement model of the foot, incorporating spatiotemporal features, is established based on a multi-dimensional force measurement system of the foot and a lower limb positioning model. The multi-dimensional force measurement model includes the yaw angle of the foot coordinate system relative to the world coordinate system for the multi-dimensional force measurement systems worn on the left and right feet, the position of the origin of the foot coordinate system in the world coordinate system, the rotation matrix of each sensing unit in the multi-dimensional force measurement system relative to the world coordinate system, and the position of each sensing unit in the world coordinate system.

[0156] like Figure 15 As shown, {O F1} and {O F2} in {O B Yaw rotation angle will be generated in} and Yaw rotation angle and The occurrence of this phenomenon stems from the foot angle during the HS event, and can be determined by... and ( and The arctangent of the ratio is obtained by finding the arctangent of the ratio of ).

[0157] .

[0158] Because {O B} is the Frenet coordinate system, therefore the yaw rotation angle ψ F1 and ψ F2 In {O G} can be represented as:

[0159] .

[0160] in, and These represent the yaw angles between the foot coordinate system and the world coordinate system for the left and right feet, respectively. and These represent the yaw rotation angles between the foot coordinate system and the body coordinate system corresponding to the left and right feet, respectively.

[0161] Each sensing unit relative to {O G The rotation matrix of} can be represented as:

[0162] ;

[0163] .

[0164] in, This indicates the various sensing units of the multi-dimensional force sensing system worn on the left foot. Rotation matrix relative to the world coordinate system; This indicates the various sensing units of the multi-dimensional force sensing system worn on the right foot. Rotation matrix relative to the world coordinate system.

[0165] according to Figure 3 The lower limb multi-sensor coordinate system shown is consistent with the lower limb kinematics model proposed in this application, {O F1} and {O F2 The origin of {O} is the same as the position of the foot in the lower limb biomechanics model. Therefore, {O} F1} and {O F2 The origin of} is in {O G The position of} can be represented as:

[0166] ;

[0167] in, and These represent the coordinates of the origin of the foot coordinate system corresponding to the multi-dimensional force sensing system worn on the left and right feet in the world coordinate system, respectively. and These represent the coordinates of the left and right feet in the world coordinate system, respectively; the coordinates of the origin of the multi-dimensional force sensing system of the soles of both feet in the world coordinate system are the same as the coordinates of the left and right feet in the world coordinate system.

[0168] In this embodiment, both the foot-shoe interface and the shoe-ground interface include 64 sensing units arranged in 8 rows and 8 columns. The position of each sensing unit in the world coordinate system of the multi-dimensional force sensing system of the foot can be represented as follows:

[0169]

[0170]

[0171] in, This refers to the sensing unit at the foot-shoe interface of the multi-dimensional force sensing system worn on the left foot. Its position in the world coordinate system; This refers to the sensing unit at the shoe-ground interface of the multi-dimensional force sensing system worn on the left foot. Its position in the world coordinate system; This refers to the sensing unit at the foot-shoe interface of the multi-dimensional force sensing system worn on the right foot. Its position in the world coordinate system; This refers to the sensing unit at the shoe-ground interface of the multi-dimensional force sensing system worn on the right foot. Position in the world coordinate system; r represents row; c represents column.

[0172] In this embodiment, the multidimensional force information of the sole includes the multidimensional forces of the sole measured by each sensing unit. Since the multidimensional forces of the sole measured by the two flexible array sensors corresponding to both feet are within {O G The XOY plane in} also has a yaw angle. and Therefore, combining the multidimensional force measurement model of the foot, the multidimensional force of the foot measured by each sensing unit in the world coordinate system can be expressed as:

[0173]

[0174] in, , and These represent the forces in the x, y, and z axes of the foot-shoe interface of the multi-dimensional force sensing system worn on the left foot, respectively. , and These represent the forces exerted by the sensing unit at the foot-shoe interface of the multi-dimensional force sensing system worn on the left foot in the x, y, and z axes of the foot coordinate system. , and These represent the forces in the x, y, and z axes of the shoe-ground interface of the multi-dimensional force sensing system worn on the left foot, respectively. , and These represent the forces in the x, y, and z axes of the foot coordinate system at the shoe-ground interface of the multi-dimensional force sensing system worn on the left foot. , and These represent the forces in the x, y, and z axes of the foot-shoe interface of the multi-dimensional force sensing system worn on the right foot, respectively. , and These represent the forces in the x, y, and z axes of the foot coordinate system at the foot-shoe interface of the multi-dimensional force sensing system worn on the right foot. , and These represent the forces in the x, y, and z axes of the shoe-ground interface of the multi-dimensional force sensing system worn on the right foot, respectively. , and These represent the forces exerted by the sensing unit at the shoe-ground interface of the multi-dimensional force sensing system worn on the right foot in the x, y, and z axes of the foot coordinate system.

[0175] In summary, the multidimensional plantar force data integrating spatiotemporal features can be represented as:

[0176] .

[0177] Spatiotemporal parameter measurements and foot multidimensional force measurement applications were also conducted for the method proposed in this application. The results show that within the walking speed range of 4~8km / h, the root mean square error of step length is <0.025m, and the relative error is <3%; the root mean square error of single step time is <0.0072s, and the relative error is <1%; the root mean square error of walking speed is <0.0249m / s, and the relative error is <1.5%; the positioning accuracy and the maximum positioning error for 1000-meter straight walking are 0.988% and 0.416%, respectively.

[0178] Based on the same inventive concept, this invention also provides a multidimensional lower limb parameter detection system that integrates spatiotemporal features, the system comprising:

[0179] The information acquisition unit is used to collect the wearer's lower limb movement information; the lower limb movement information includes position information, posture information and multi-dimensional force information on the sole of the foot.

[0180] The first analysis unit is used to estimate the spatiotemporal parameters and gait events of the lower limbs based on the lower limb biomechanics model and the lower limb motion information.

[0181] The second analysis unit is used to obtain multidimensional plantar force data with spatiotemporal features based on the multidimensional plantar force measurement model and the lower limb motion information.

[0182] In this application, the other technical features of the multidimensional lower limb parameter detection system that integrates spatiotemporal features are the same as those disclosed in the above method embodiments, and will not be repeated here.

[0183] Therefore, this invention discloses a method and system for detecting multidimensional lower limb parameters by integrating spatiotemporal features. The method includes the following steps: using a wearable device to collect the wearer's lower limb motion information; the lower limb motion information includes position information, posture information, and multidimensional plantar force information; estimating lower limb spatiotemporal parameters and gait events based on the lower limb biomechanical model and the lower limb motion information; and obtaining multidimensional plantar force data integrating spatiotemporal features based on the multidimensional plantar force measurement model and the lower limb motion information. This method can accurately estimate lower limb spatiotemporal parameters and human trajectory, and realize the solution and mapping of three-dimensional plantar force space.

[0184] The above are merely embodiments of this application and do not limit the scope of this patent application. Any equivalent structural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of this application.

Claims

1. A method for detecting multidimensional lower limb parameters by integrating spatiotemporal features, characterized in that, The method includes the following steps: Wearable devices are used to collect the wearer's lower limb movement information; the lower limb movement information includes position information, posture information, and multi-dimensional force information on the sole of the foot; Based on the lower limb biomechanics model and the lower limb motion information, estimate the spatiotemporal parameters of the lower limb and gait events; Determine the capture point, which is used to assess and control the stability of the human lower limb gait. The capture point refers to the position where the wearer's swing leg lands at the end of the current gait cycle in order to achieve the desired orbital energy at the start of the next gait cycle. The capture point x cp The formula is expressed as: ； Where, x t It is the current position of the center of mass. Let x be the velocity of the center of mass, z be the height of the center of mass, g be the acceleration due to gravity, and E2 be the target orbital energy. When E2 ≥ 0, the human lower limbs complete continuous forward walking. In continuous forward walking, the heel contact event represents the start of the supporting foot switching process. Under the condition that E2 ≥ 0, x... cp This corresponds to the heel strike event; The gait events are detected by identifying the maximum forward position of the foot and the maximum backward velocity of the foot; the gait events include events marking the end of the initial support phase, events marking the end of the middle support phase, events marking the end of the initial swing phase, and events marking the end of the middle swing phase. The event marking the end of the initial support phase indicates that the foot is in full contact with the ground, and also marks the beginning of full weight-bearing on the foot. At this time, the toes of the opposite foot also leave the ground. At this time, the double support phase in the support phase ends and the single support phase begins. The event that marks the end of the middle of the support phase indicates that the supporting leg is about to enter the end of the support phase, at which point the supporting foot completes forward support and the heel begins to leave the ground; The event that marks the end of the initial phase of the swing indicates the point where the swinging foot is closest to the opposite supporting foot; The event marking the end of the mid-swing indicates that the swinging foot is in a vertical position and is about to begin deceleration and braking, entering the final stage of the swing. Based on the multidimensional force measurement model of the foot and the lower limb motion information, multidimensional force data of the foot with spatiotemporal features are obtained.

2. The multidimensional lower limb parameter detection method according to claim 1, characterized in that, The wearable device includes a lower limb spatiotemporal parameter sensing system and a foot multidimensional force sensing system; The lower limb spatiotemporal parameter sensing system includes a lidar ranging sensor and an inertial sensor. The lidar ranging sensor is used for distance measurement, and the inertial sensor is used for attitude detection. The foot multidimensional force sensing system includes two flexible array sensors, which are fixedly connected by multiple corresponding substrates. One of the flexible array sensors is used to contact the sole of the foot to detect multidimensional pressure in the sole area; the other flexible array sensor is used to contact the ground to detect the overall three-dimensional ground reaction force.

3. The multidimensional lower limb parameter detection method according to claim 2, characterized in that, The method further includes the following steps: The lower limb spatiotemporal parameter sensing system is worn on the wearer's calf, and the foot multidimensional force sensing system is worn on the wearer's foot.

4. The multidimensional lower limb parameter detection method according to claim 3, characterized in that, The lower limb biomechanical model includes a model center of mass, a first rotation axis, a second rotation axis, a first extendable rigid rod, and a second extendable rigid rod. The model's center of mass is the center of the first and second rotation axes, and the model's center of mass is the geometric center of the body. The model's center of mass will move during the wearer's walking process. The first rotating shaft is connected to the first retractable rigid rod, and the second rotating shaft is connected to the second retractable rigid rod. The first rotating shaft corresponds to the wearer's left leg major trochanter, the second rotating shaft corresponds to the wearer's right leg major trochanter, the first retractable rigid rod corresponds to the wearer's left leg, and the second retractable rigid rod corresponds to the wearer's right leg.

5. The multidimensional lower limb parameter detection method according to claim 4, characterized in that, The method further includes the following steps: The height of the wearer's left / right greater trochanter from the ground was measured, as well as the distance between the left / right greater trochanter and the geometric center of the body was measured.

6. The multidimensional lower limb parameter detection method according to claim 5, characterized in that, The estimation of lower limb spatiotemporal parameters and gait events based on the lower limb biomechanical model and the lower limb motion information includes: Calculate the step length based on the displacement between the heel strike event of the same-side foot and the heel strike event of the opposite-side foot; The stride length is calculated by accumulating the displacement of the ipsilateral foot from the heel-to-toe-toe-off event to the swing-off event and from the next heel-to-toe-off event. Calculate the step speed based on the ratio of the step length to the step time; The stride width is calculated based on the distance in the coronal axis direction between the heel strike events of the left and right feet in two consecutive steps.

7. The multidimensional lower limb parameter detection method according to claim 6, characterized in that, The method further includes the following steps: A lower limb positioning model is established, comprising a linear motion model and a turning motion model. The linear motion model includes the step length of the movement trajectory after alternating foot pushes. The turning motion model includes the trajectory curvature, average rotation angle, rotation matrix of the body coordinate system relative to the world coordinate system, and the positions of the model's centroid, left foot, and right foot in the world coordinate system. The trajectory curvature is the ratio of the difference in step length between the left and right feet to the step width, and the average rotation angle is the product of the average step length of the left and right feet and the trajectory curvature. The rotation matrix of the body coordinate system relative to the world coordinate system can be expressed as: ； in, This represents the average rotation angle.

8. The method for detecting multidimensional lower limb parameters according to claim 7, characterized in that, The method further includes: Based on the aforementioned multi-dimensional force sensing system and the lower limb positioning model, a multi-dimensional force sensing model of the foot is established that integrates spatiotemporal features. The multi-dimensional force sensing model of the foot includes the yaw angle of the foot coordinate system relative to the world coordinate system corresponding to the multi-dimensional force sensing system worn on the left foot and the right foot, the position of the origin of the foot coordinate system in the world coordinate system, the rotation matrix of each sensing unit in the multi-dimensional force sensing system of the foot relative to the world coordinate system, and the position of each sensing unit in the world coordinate system. The yaw angle is calculated by the yaw rotation angle between the foot coordinate system and the body coordinate system. The rotation matrix of each sensing unit in the foot multidimensional force sensing system relative to the world coordinate system can be expressed as: ； ； in, This represents the rotation matrix of each sensing unit of the multi-dimensional force sensing system worn on the left foot relative to the world coordinate system; This represents the rotation matrix of each sensing unit of the multi-dimensional force sensing system worn on the right foot relative to the world coordinate system; and These represent the yaw angles between the foot coordinate system and the world coordinate system corresponding to the left and right feet, respectively.

9. A multidimensional lower limb parameter detection system integrating spatiotemporal features, characterized in that, The system includes: The information acquisition unit is used to collect the wearer's lower limb movement information; the lower limb movement information includes position information, posture information, and multi-dimensional force information on the sole of the foot; The first analysis unit is used to estimate the spatiotemporal parameters and gait events of the lower limbs based on the lower limb biomechanics model and the lower limb motion information. Determine the capture point, which is used to assess and control the stability of the human lower limb gait. The capture point refers to the position where the wearer's swing leg lands at the end of the current gait cycle in order to achieve the desired orbital energy at the start of the next gait cycle. The capture point x cp The formula is expressed as: ； Where, x t It is the current position of the center of mass. Z is the velocity of the center of mass, z is the height of the center of mass, g is the gravitational acceleration, and E2 is the target orbital energy. When E2≥0, the human lower limbs complete continuous forward walking. In continuous forward walking, the heel strike event represents the start of the supporting foot switching process; x satisfies the condition E2≥0. cp This corresponds to the heel strike event; The gait events are detected by identifying the maximum forward position of the foot and the maximum backward velocity of the foot; the gait events include events marking the end of the initial support phase, events marking the end of the middle support phase, events marking the end of the initial swing phase, and events marking the end of the middle swing phase. The event marking the end of the initial support phase indicates that the foot is in full contact with the ground, and also marks the beginning of full weight-bearing on the foot. At this time, the toes of the opposite foot also leave the ground. At this time, the double support phase in the support phase ends and the single support phase begins. The event that marks the end of the middle of the support phase indicates that the supporting leg is about to enter the end of the support phase, at which point the supporting foot completes forward support and the heel begins to leave the ground; The event that marks the end of the initial phase of the swing indicates the point where the swinging foot is closest to the opposite supporting foot; The event marking the end of the mid-swing indicates that the swinging foot is in a vertical position and is about to begin deceleration and braking, entering the final stage of the swing. The second analysis unit is used to obtain multidimensional plantar force data with spatiotemporal features based on the multidimensional plantar force measurement model and the lower limb motion information.

Images