Intelligent walking aid control method for hind limbs of pet

Abstract

The invention provides a pet hind limb intelligent walking assisting control method which is realized through a power assisting device and comprises the following steps: S1, acquiring the angular velocity, acceleration and attitude angle of a dog in real time through an intelligent measuring mechanism; s2, fusing the data in the S1 according to a complementary filtering algorithm to obtain a corrected attitude angle; s3, the walking state of the dog is judged according to the pressure, the attitude angular rate and the acceleration modulus length collected by the intelligent measuring mechanism; and S4, constructing the intelligent measurement mechanism into a two-dimensional pressure distribution matrix, and adjusting the output pressure value through a PID algorithm according to the target supporting torque. The reliability is good, and the adjustment is convenient.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of pet assistance, in particular to a hind limb intelligent walking aid control method for pets. BACKGROUND

[0002] With the development of the times, in order to meet the special needs of old dogs and dogs with movement disorders, auxiliary assistance devices will be set up for them. The existing auxiliary assistance devices cannot dynamically adjust according to the individual differences, gait changes and health conditions of the dogs. Since this equipment cannot be adjusted in real time according to the health and mobility of the dog, long-term wearing may cause discomfort to the dog, and even affect its health; That is, the existing auxiliary assistance devices can only output a preset torque, and cannot adjust the assistance in real time according to the gait of the dog, which may cause the dog's hind legs to be suddenly pushed by too strong a force as soon as they leave the ground, and then suddenly lose support at the moment of landing, thereby causing the joint to be further worn.

[0003] As disclosed in the patent document with patent application number 202010904604.4 and publication date of November 27, 2020, a walking aid for dogs with hind limb disabilities includes a support for supporting the hind limbs of the dog, a plurality of intelligent transmission modules for acquiring muscle signals of the dog, predicting behaviors of the dog, and adjusting the support to assist the dog to complete the predicted behaviors according to the muscle signals. It can support the dog with hind limb disabilities to complete the actions required for daily activities.

[0004] The above document controls the predicted behavior of the dog through the intelligent transmission module to adjust the support, so that the support assists the dog to complete the predicted behavior; the intelligent transmission module can only predict the behavior to output fixed assistance to change the shape of the support, but it cannot real-time fine-tune the size of the assistance and the attitude angle during the dog's walking process. If the pet is in a supporting state due to fatigue or lowering of the leg lifting height, and the intelligent transmission module directly outputs the assistance size and rotation angle in the preset supporting state, the hind limbs may be forcibly pulled, and the joint may be sprained. SUMMARY

[0005] The present application provides a hind limb intelligent walking aid control method for pets, which has good assistance effect and good control effect.

[0006] To achieve the above purpose, the technical scheme of the present application is: a hind limb intelligent walking aid control method for pets, which is realized by setting a walking aid device on the hind limbs of the pet. The walking aid device includes an intelligent measurement mechanism and a driving device, the driving device is electrically connected with the intelligent measurement mechanism, and the specific steps include: S1 real-time acquisition of the angular velocity of the pet through the intelligent measurement mechanism , acceleration , roll angle and yaw angle ; S2 fusing the data in S1 according to a complementary filter algorithm to obtain a corrected attitude angle ; and constructing a state vector according to the corrected attitude angle, acceleration, roll angle and yaw angle; adjusting the output attitude angle according to the target attitude angle and the corrected attitude angle by using a PID algorithm; S3 judging the walking state of the pet according to the pressure , attitude angle rate and acceleration module length collected by the intelligent measuring mechanism; if it is in the support phase, entering step S4; S4 constructing the intelligent measuring mechanism into a two-dimensional pressure distribution matrix of the pet's foot bottom , i is the setting position in the horizontal direction, and j is the setting position in the vertical direction; determining a two-dimensional distribution torque according to the two-dimensional pressure distribution matrix; and determining a target support torque according to the preset expected pitch angle and the preset target angular velocity, and adjusting the output torque according to the target support torque and the corresponding two-dimensional distribution torque by using a PID algorithm.

[0007] S5 adjusting by the driving device according to the output torque and the output attitude angle.

[0008] The above setting realizes real-time monitoring of the pet's attitude angle and pressure state through the intelligent measuring mechanism, then the attitude angle of the intelligent measuring mechanism is corrected by the complementary filter algorithm, then the error between the corrected attitude angle and the target attitude angle is dynamically adjusted by the PID algorithm to output the attitude angle, and when the pet's hind limbs are in the support state, the pet's state is identified and judged, the target torque is determined according to the attitude angle, and then the error value between the target torque is adjusted according to the detected two-dimensional pressure value matrix according to the PID algorithm to output the pressure value, so as to ensure that the pet's hind limbs are adjusted in real time according to the detected attitude angle and the pressure value of the foot bottom to output the attitude angle and the pressure value, rather than simply adjusting by the preset value, thereby preventing the damage of the hind limbs caused by the instantaneous straightening of the hind limbs, so as to have good and reliable power assisting effect and good control effect.

[0009] Further, step S3 further comprises: S31 if the power assisting device is in a static state, correcting the acceleration by using static gravity data; the correction formula is .

[0010] The above setting assigns zero value to the acceleration when the power assisting device is in a static state, thereby ensuring the reliability of the initial state.

[0011] Further, the intelligent measuring mechanism comprises an IMU unit, and the IMU unit comprises a three-axis gyroscope and a three-axis acceleration sensor, and in step S1, the angular velocity , wherein is a rotation angular velocity of the pet around the X axis, is a rotation angular velocity of the pet around the Y axis, is a rotation angular velocity of the pet around the Z axis, acceleration wherein is an acceleration of the pet around the X axis, is an acceleration of the pet around the Y axis, is an acceleration of the pet around the Z axis.

[0012] The above settings are used to collect the motion state of the pet in real time through an IMU (Inertial Measurement Unit) unit including a three-axis gyroscope and a three-axis acceleration sensor, so as to ensure the correctness of the collected data.

[0013] Further, in step S2, the complementary filtering algorithm includes a formula , t is a current sampling time, Δt is a sampling period time, g is a gravitational acceleration, is a preset parameter; In step S21, a state vector is constructed , wherein Xt represents an attitude angle and a linear acceleration state at time t.

[0014] The above settings are used for the complementary filtering algorithm to fuse the data in S1, so as to obtain a corrected attitude angle , reduce errors, and construct a state vector through the corrected attitude angle and the acceleration so as to determine the attitude angle and the linear acceleration subsequently.

[0015] Further, in step S3, the walking state of the pet is determined, including: a preset pressure threshold , if , it indicates that the pet lands; a preset disturbance threshold , if , it indicates that the attitude of the pet is stable; the acceleration module length is determined, if the acceleration module length is approximately equal to the gravitational acceleration g, it indicates that the pet motion tends to be stationary; in step S42, if all the above conditions are met, it indicates that the pet enters the support phase, otherwise, it is the swing phase.

[0016] The above settings are used to determine the walking state of the pet through the pressure , the attitude angular rate , and the acceleration module length . Only when the above three independent physical variables are met at the same time, the state of the pet can be determined as the support phase, so as to reduce the risk of misjudgment.

[0017] Further, the intelligent measuring mechanism comprises a pressure sensor, and the target support torque M is determined according to the preset expected pitch angle and the preset target angular velocity in step S4, , θref is a set expected attitude angle, ωref is a set target angular velocity, kp is a proportional gain, kd is a differential gain coefficient, θ is a currently measured attitude angle, and ω is a currently measured angular velocity.

[0018] The above setting determines the target support torque through the attitude angle, then determines a two-dimensional pressure matrix through the pressure sensor, and determines the torque through the product of the pressure and the force arm, and the reliability is good.

[0019] Further, the PID algorithm in step S2 comprises: , wherein kp is a proportional gain, Kp is a control gain, kd is a differential gain, u(t) is an output attitude angle, e(t) is an error value of the attitude angle, , θref is a set expected attitude angle at a current time, θ is a current attitude angle.

[0020] The above setting adjusts the output control driving device dynamically to adjust the attitude angle through the PID regulation.

[0021] Further, S3 further comprises: S32 if a high-frequency changing angular velocity is detected, the differential gain in the PID algorithm is dynamically adjusted, and the differential gain formula in the PID algorithm is dynamically adjusted: , L is a lower limit of the high-frequency loop, U is an upper limit of the high-frequency loop, ω is a current system angular velocity change rate, ΔKd is a preset maximum amplitude of the gain that can be improved, K d0 is a preset differential gain under a static condition.

[0022] The above setting proves that the assistance device is in a high-frequency disturbance condition when the high-frequency angular velocity is detected, and the differential gain is adjusted according to the angular velocity in different ranges at this time, so that the differential gain can more accurately adjust the attitude angle and the pressure value.

[0023] Further, step S3 further comprises S33 when the acceleration direction and the angular velocity are opposite , is a first derivative of the acceleration in the x direction; is an angular velocity in the x direction, and the control buffer is started.

[0024] When the acceleration direction and the angular velocity direction are reversed, the rotation speed of the motor or the output force is actively slowed down or reduced, so as to avoid imbalance or falling of the pet caused by sudden adjustment, thereby making the support adjustment process more stable and safe.

[0025] Further, the walking aid device further comprises a fixed support, the fixed support is provided with a first bandage, the fixed support is connected with the hind leg thigh of the pet through the first bandage, and the two sides of the fixed support are respectively provided with driving devices.

[0026] The fixed support is connected with the hind leg thigh of the dog through the first bandage, the fixed support is connected with the two sides of the moving support through the driving devices, the moving support is connected with the hind leg calf of the pet through the second bandage, the driving devices drive the moving support to rotate relative to the fixed support, and the wearing of the walking aid device is completed. BRIEF DESCRIPTION OF DRAWINGS

[0027] Figure 1 The working principle flowchart of the present application is shown.

[0028] Figure 2 The installation schematic diagram of the power assisting device on the hind limbs of the pet in an embodiment of the present application is shown.

[0029] The fixed support is connected with the hind leg thigh of the dog through the first bandage, the fixed support is connected with the two sides of the moving support through the driving devices, the moving support is connected with the hind leg calf of the pet through the second bandage, the driving devices drive the moving support to rotate relative to the fixed support, and the wearing of the walking aid device is completed. DETAILED DESCRIPTION

[0030] Embodiment one.

[0031] As Figures 1-2As shown, a method for controlling intelligent hind limb walking assistance for a pet is implemented by providing a walking assistance device on the pet's hind limb. The walking assistance device includes an intelligent measuring mechanism, a fixed bracket 1 and a movable bracket 3. A driving device 2 is provided between the fixed bracket 1 and the movable bracket 3. The driving device 2 is electrically connected to the intelligent measuring mechanism. The specific steps include: S1 collects the pet's angular velocity in real time through an intelligent measurement mechanism , acceleration , roll angle and yaw angle ; S2 fuses the data in S1 according to the complementary filtering algorithm to obtain the corrected attitude angle ; And construct the state vector according to the corrected attitude angle, acceleration, roll angle and yaw angle; Use PID algorithm to adjust the output attitude angle according to the target attitude angle and the corrected attitude angle; S3 collects pressure based on intelligent measuring mechanism , attitude angular rate and acceleration modulus , judge the pet's walking status; S4 constructs the intelligent measurement mechanism into a two-dimensional pressure distribution matrix , i is the setting position in the lateral direction, j is the setting position in the vertical direction; the two-dimensional distributed torque is determined according to the two-dimensional pressure distribution matrix; and the target support torque is determined according to the preset desired pitch angle and the preset target angular velocity, and the output torque is adjusted according to the target support torque and the corresponding two-dimensional distributed torque using the PID algorithm; S5 is adjusted by the drive device according to the output torque and output attitude angle.

[0032] like Figure 1 As shown, the intelligent measurement mechanism includes an IMU unit 6 (Inertial Measurement Unit), which includes a three-axis gyroscope and a three-axis acceleration sensor. The fixing bracket 1 is provided with a mounting slot for mounting the IMU unit 6 (not shown in the figure). In step S1, the angular velocity ,in is the pet's angular velocity around the X axis, is the pet's rotational angular velocity around the Y axis, The pet's angular velocity and acceleration around the Z axis ,in is the pet's acceleration around the X axis, is the pet's acceleration around the Y axis, The acceleration of the pet around the Z axis is collected by an IMU (Inertial Measurement Unit) unit including a three-axis gyroscope and a three-axis acceleration sensor in real time to ensure the correctness of the collected data.

[0033] In an embodiment, in step S2, the complementary filtering algorithm includes a formula , wherein t is a current sampling time, At is a sampling period time, and g is a gravity acceleration. is a preset parameter.

[0034] In step S2, a state vector is constructed , wherein Xt represents an attitude angle and a linear acceleration state at time t, and the specific construction of the state vector Xt is performed by sequentially arranging a plurality of parameters to form a matrix. The complementary filtering algorithm fuses the data in S1 to obtain a corrected attitude angle , thereby reducing errors.

[0035] In an embodiment, in step S3, the walking state of the pet is determined by: presetting a pressure threshold , if , it indicates that the pet lands; presetting a disturbance threshold , if , it indicates that the attitude of the pet is stable; determining an acceleration module length , if the acceleration module length is approximately equal to the gravity acceleration g, it indicates that the pet motion tends to be stationary; and in step S42, if all the above conditions are met, it indicates that the pet enters a support phase, otherwise it is a swing phase. , the attitude angle rate , and the acceleration module length are used to determine the walking state of the pet.

[0036] In step S2, the PID algorithm includes: , wherein is a proportional gain; Kp is a control gain; Kd is a differential gain, u(t) is an output attitude angle, and e(t) is an error value of the attitude angle. , is a set expected attitude angle at a current time, is an attitude angle at a current time.

[0037] In step S4, the PID algorithm formula is still ; In this formula, v(t) is the adjusted output torque, e(t) is the error between the target support torque and the detected two-dimensional distributed torque, ki is the proportional gain; kp is the control gain; kd is the differential gain.

[0038] In this embodiment, by calculating the two-dimensional distribution moment M _actual = ∑ P ij · r ij , r ij is the vertical distance from the i,jth sensor to the rotation center, i is the setting position of the pressure sensor in the lateral direction, and j is the setting position of the pressure sensor in the vertical direction. Then the error value e(t) = ΔM = M–M is obtained. _actual , and then use the PID algorithm to adjust the output torque according to the target support torque and the corresponding two-dimensional distribution torque, and then according to the output torque M1= ∑ P ij · r ij , due to r ij is the vertical distance from the i,jth sensor to the center of the drive device.

[0039] In one embodiment, step S01 is further included before step S1; when the driving device 2 stops working, step S01 performs zero offset correction. ; Step S2 also includes steps S11-S12; if S11 detects a high-frequency angular velocity change, the differential gain in the PID algorithm is dynamically adjusted; the differential gain formula in the PID algorithm is dynamically adjusted: , is the lower limit of high-frequency bypass, is the upper limit of high-frequency bypass, is the current system angular velocity change rate, ΔKd is the maximum amplitude that the preset gain can increase, K d0 The preset derivative gain under static conditions.

[0040] S12 detects that the acceleration direction is opposite to the angular velocity , start control buffer, is the first derivative of the acceleration in the x direction; The angular velocity in the x-direction is controlled by actively slowing down the motor's rotation speed or reducing the output force to prevent the pet from losing balance or falling due to sudden adjustments, thereby making the support adjustment process smoother and safer.

[0041] In the situation of fast movement of the pet, uneven ground, sensor drift, etc., the system may have a short time error or response delay, so the zero point offset correction of step S01 can be performed, and when the acceleration direction and the angular velocity direction are opposite in step S12, the rotation speed of the motor is actively slowed down or the output intensity is reduced, so as to avoid the pet from losing balance or falling due to sudden adjustment, thereby making the support adjustment process more stable and safe.

[0042] As shown in Figures 1-2 The fixed support 1 is provided with a first binding band 4, the fixed support 1 is connected with the pet's hind leg thigh part through the first binding band 4, and the fixed support 1 is provided with a driving device 2 on each side. The two driving devices 2 are connected with the two sides of the moving support 3 respectively, the moving support 3 is provided with a second binding band 5, the moving support 3 is connected with the pet's hind leg calf part through the second binding band 5, the driving device 2 drives the moving support 3 to rotate relative to the fixed support 1, the driving device 2 includes a driving mounting seat provided with a driving motor, the output shaft of the driving motor is fixedly connected with the moving support 3, the driving mounting seat is connected with the fixed support 1, the fixed support 1 is connected with the pet's hind leg thigh part through the first binding band 4, the fixed support 1 is connected with the two sides of the moving support 3 through the driving device 2, the moving support 3 is connected with the pet's hind leg calf part through the second binding band 5, the driving device 2 drives the moving support 3 to rotate relative to the fixed support 1, and the wearing of the walking aid device is completed. Because the driving device 2 is arranged on each side of the fixed support 1, the two independent driving devices 2 ensure symmetrical assistance, prevent lateral load of the joint caused by unilateral driving device 2 assistance, and when the driving device 2 drives the moving support 3 to move, the moving direction of the moving support 3 is coaxial with the natural swing direction of the pet's hind leg calf part, that is, the rotation axis of the moving support 3 is coincident with the instantaneous rotation axis of the dog's knee joint, so as to ensure that the force direction provided by the driving device 2 is consistent with the movement direction of the pet's hind leg calf part. The first binding band 4 and the second binding band 5 are elastic binding bands.

[0043] In another embodiment, the driving device of the walking aid device can also be rotated through the walking aid device in Chinese patent application No. 202010904604.4.

[0044] The working principle of the present application is that the real-time monitoring of the pet posture angle and pressure state is realized through the intelligent measuring mechanism, then the posture angle is corrected through the complementary filtering algorithm, then the error between the corrected posture angle and the target posture angle is dynamically adjusted according to the PID algorithm to output the posture angle, and when the pet hind limbs are in the supporting state, the pet state is identified and judged, the target torque is determined according to the posture angle, then the error value between the target torque is adjusted according to the detected two-dimensional pressure value matrix according to the PID algorithm to output the pressure value, so that the pet hind limbs can be adjusted and output in real time according to the detected posture angle and the pressure value of the sole, and the adjustment is not simply through the preset value, so as to prevent the damage of the hind limbs caused by the instantaneous straightening of the hind limbs, thereby achieving good power assisting effect and strong reliability.

Claims

1. A method for controlling the intelligent walking assistance of a pet's hind limbs, which is achieved by providing a walking assistance device on the pet's hind limbs, characterized in that: The walking aid device includes an intelligent measuring mechanism and a driving device, wherein the driving device is electrically connected to the intelligent measuring mechanism. The specific steps include: S1 collects the pet's angular velocity in real time through an intelligent measurement mechanism , acceleration , attitude angle θ, roll angle and yaw angle ; S2 fuses the data in S1 according to the complementary filtering algorithm to obtain the corrected attitude angle ; And construct the state vector according to the corrected attitude angle, acceleration, roll angle and yaw angle; Use PID algorithm to adjust the output attitude angle according to the target attitude angle and the corrected attitude angle; S3 collects pressure based on intelligent measuring mechanism , attitude angular rate and acceleration modulus , judge the walking state of the pet; if it is in the supporting phase, go to step S4; S4 uses intelligent measurement mechanism to construct a two-dimensional pressure distribution matrix on the pet's sole , i is the setting position in the lateral direction, j is the setting position in the vertical direction; the two-dimensional distributed torque is determined according to the two-dimensional pressure distribution matrix; and the target support torque is determined according to the preset desired pitch angle and the preset target angular velocity, and the output torque is adjusted according to the target support torque and the corresponding two-dimensional distributed torque using the PID algorithm; S5 is adjusted by the drive device according to the output torque and output attitude angle.

2. The method for controlling the intelligent hind limbs of a pet according to claim 1, characterized in that: Before step S1, the following steps are also included: S01: if the power assist device is in a stationary state, the acceleration is corrected by using the static gravity data; the correction formula is: .

3. The method for controlling the intelligent hind limbs of a pet according to claim 1, characterized in that: The intelligent measurement mechanism includes an IMU unit, and the IMU unit includes a three-axis gyroscope and a three-axis acceleration sensor. In step S1, the angular velocity ,in is the pet's angular velocity around the X axis, is the pet's angular velocity around the Y axis, The pet's angular velocity and acceleration around the Z axis ,in is the pet's acceleration around the X axis, is the pet's acceleration around the Y axis, The pet's acceleration around the Z axis.

4. The method for controlling the intelligent hind limb walking of a pet according to claim 2, characterized in that: In step S2, the complementary filtering algorithm includes the formula , t is the current sampling time, Δt is the sampling period time, g is the acceleration of gravity, are preset parameters; In step S2, construct the state vector , Xt represents the attitude angle and linear acceleration state at time t.

5. The method for controlling the intelligent hind limb walking of a pet according to claim 1, characterized in that: The step S3 of judging the pet's walking state includes: presetting the pressure threshold ,like , it means the pet has landed; preset disturbance threshold ,like , it means the pet's posture is stable; determine the acceleration modulus , if the acceleration modulus If it is approximately equal to the acceleration due to gravity g, it means that the pet's movement is tending to stillness; if the three conditions are met, it means that the pet has entered the support phase, otherwise it is in the swing phase.

6. The method for controlling the intelligent hind limb walking of a pet according to claim 1, characterized in that: The intelligent measuring mechanism includes a pressure sensor. In step S4, the target support torque M is determined according to the preset desired pitch angle and the preset target angular velocity. , θref is the desired attitude angle; ωref is the target angular velocity; kp is the proportional gain, kd is the differential gain coefficient; θ is the currently measured attitude angle and ω is the currently measured angular velocity.

7. The method for controlling the intelligent hind limb walking of a pet according to claim 1, characterized in that: The PID algorithm in step S2 includes: ,in is the proportional gain; Kp is the control gain; Kd is the differential gain, u(t) is the output attitude angle, e(t) is the error value of the attitude angle, , Set the desired attitude angle for the current moment, is the attitude angle at the current moment.

8. The method for controlling the intelligent hind limb walking of a pet according to claim 1, characterized in that: S1 also includes: S11 dynamically adjusts the differential gain in the PID algorithm if a high-frequency angular velocity change is detected; the differential gain formula in the dynamic adjustment PID algorithm is: , is the lower limit of high-frequency bypass, is the upper limit of high-frequency bypass, is the current system angular velocity change rate, ΔKd is the maximum amplitude that the preset gain can increase, K d0 The preset derivative gain under static conditions.

9. The method for controlling the intelligent hind limb walking of a pet according to claim 1, characterized in that: Step S1 also includes step S12: when it is detected that the acceleration direction is opposite to the angular velocity , is the first derivative of the acceleration in the x direction; The angular velocity in the x direction starts the control buffer.

10. The method for controlling the intelligent hind limb walking of a pet according to claim 1, characterized in that: The walking aid device also includes a fixed bracket, which is provided with a first strap, and the fixed bracket is connected to the pet's hind thigh through the first strap. Driving devices are respectively provided on both sides of the fixed bracket, and the two driving devices are respectively connected to the two sides of the movable bracket. The movable bracket is provided with a second strap, and the movable bracket is connected to the pet's hind leg through the second strap. The driving device drives the movable bracket to rotate relative to the fixed bracket, and the driving device includes a driving mounting seat provided with a driving motor, the driving motor output shaft is connected to the movable bracket, and the driving mounting seat is connected to the fixed bracket.

Images