Self-adaptive collaborative motion planning method for exoskeleton and safety support

Abstract

The invention relates to the field of robot motion planning, in particular to a self-adaptive collaborative motion planning method of an exoskeleton and a safety support, which takes an expected foothold position of a lower limb exoskeleton as input, and uses a space quantization dynamics gait method based on a cosine method to generate a straight knee gait mode of the lower limb exoskeleton in a self-adaptive manner. And then dynamic motion of the mass center at different walking speeds is modeled by using a reduced-order kinetic model, and optimal solutions of the self-adaptive foothold position and the walking speed are solved by constructing a target function, so that cooperative motion control of the exoskeleton and the walking bracket is realized. The problems of cooperative motion control of a human-exoskeleton-walking support system and self-adaption of the walking support to different walking speeds are solved.

Description

Technical Field

[0001] The present invention relates to the field of robot motion planning, and in particular to an adaptive collaborative motion planning method for an exoskeleton and a safety bracket. Background Art

[0002] Paraplegic patients who use lower limb exoskeletons to walk usually need to use crutches to maintain balance, which is extremely difficult for patients with weak upper limb strength, so a movable safety bracket is generally used to help patients maintain balance during gait training. However, if there is no proper coordinated motion planning between the exoskeleton and the walking bracket, the safety bracket will push or pull the human-exoskeleton robot system, causing the system to present an abnormal gait pattern, thereby affecting the rehabilitation effect of gait training.

[0003] In the early stages of rehabilitation, for patients with weak muscle strength, an external support is needed to maintain balance, either a fixed frame or a removable walking frame. For a fixed frame, gait training at different walking speeds is usually performed on a treadmill. For example, Lokomat is a treadmill-based body weight-supported robotic gait training device that provides lateral movement for the patient's center of mass transfer and generates a human-like straight-knee gait training pattern, and also allows gait training at different walking speeds. Lopes is also a similar device with a gravity support system and treadmill, as well as an exoskeleton leg with two actuated joints, which provides more degrees of freedom for gait training due to the use of cable-driven elastic actuators. Although the exoskeleton robot with a fixed frame and gravity support system is beneficial to the patient's gait recovery, gait training on a treadmill does not correspond to real-life ground walking. Therefore, for patients with weak muscle strength, a removable walking frame for ground gait training in the early stages of rehabilitation is essential.

[0004] Some mobile walking supports, such as Andago and other mobile robot platforms, can be connected to patients with ropes, saddles, and some other soft materials. Part of the body weight is supported by the mobile walking support, effectively reducing the burden on the legs. CPWalker is designed for the rehabilitation training of children with cerebral palsy. It is a passive mobile walking support that allows patients to walk freely with body weight support. AiWalker is a passive mobile walking support used in combination with an exoskeleton to evaluate the walking efficiency of spinal cord injury patients during gait training. Usually, the connection between the exoskeleton and the walking support is a hard connection without any drive. In addition, these walking supports do not provide active control, which limits the vertical and horizontal movement of the center of mass of the patient during walking and is not conducive to the recovery of normal gait. Most importantly, combining wearable lower limb exoskeleton robots with mobile walking supports for ground gait training has great potential for the rehabilitation of gait disorder patients. The biggest challenge is to control the coordinated movement of the exoskeleton robot and the walking support and perform natural walking assistance. Summary of the Invention

[0005] The purpose of the present invention is to provide an adaptive coordinated motion planning method for an exoskeleton and a safety support. This method generates a straight-knee anthropomorphic gait pattern of the exoskeleton by generating the expected landing point position and the expected walking speed, and realizes the coordinated control of the lower limb exoskeleton and the walking support.

[0006] To achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0007] An adaptive coordinated motion planning method for an exoskeleton and a safety support includes the following steps:

[0008] 1. An adaptive coordinated motion planning method for an exoskeleton and a safety support, which is applicable to a human-lower limb exoskeleton-support system. The support in the human-lower limb exoskeleton-support system includes a driving wheel and a linear joint for connecting and supporting the human-exoskeleton. The characteristics are that the implementation steps of this method include:

[0009] Step 1: Use the spatial quantization dynamics gait generation method based on the cosine method to generate the center of mass trajectory of the lower limb exoskeleton;

[0010] Step 2: In the swing phase: Based on the expected landing point, use the center of mass trajectory generated in Step 1, and use the trajectory generation method to generate the gait trajectory of the swing leg of the lower limb exoskeleton; Calculate all joint angles of the lower limb exoskeleton according to the position of the supporting leg, the center of mass trajectory, and the gait trajectory of the swing leg in the swing phase;

[0011] Step 3: Perform spatial discretization processing on the center of mass trajectory using the preset uniform increment length of the center of mass horizontal displacement, and perform spatial discretization processing on the joint angles using the preset center of mass horizontal displacement;

[0012] Step 4: Establish a cart-pendulum dynamic model for the human-exoskeleton system, construct its optimization equation, convert the centroid trajectory and joint angles in the spatial domain to the time domain, and generate a variable-speed straight-knee anthropomorphic gait trajectory:

[0013] 4.1. Based on the reduced-order dynamic model, establish a cart-pendulum dynamic model for the human-exoskeleton-bracket system, and use the cart-pendulum dynamic model to calculate the horizontal displacement of the centroid varying with space transformation;

[0014] 4.2. Construct an optimization problem equation for the cart-pendulum dynamic model, use the desired walking speed as the input, and use the optimization problem equation to calculate the horizontal displacement of the centroid varying with time;

[0015] 4.3. Determine the horizontal x-coordinate of the centroid varying with time at the point where the horizontal displacement of the centroid varying with space coincides with the horizontal displacement of the centroid varying with time;

[0016] 4.4. Calculate the horizontal displacement time per unit length of the centroid according to the centroid horizontal speed. According to the horizontal displacement time per unit length and the horizontal x-coordinate of the centroid varying with time, convert the gait of the lower-limb exoskeleton in the spatial domain to the time domain, and generate a variable-speed straight-knee anthropomorphic gait trajectory;

[0017] Step 5: Use the centroid trajectory obtained in Step 1 as the reference trajectory, calculate the reference positions of the driving wheels and linear joints of the bracket according to the reference trajectory, and input them into the controllers of the driving wheels and linear joints of the bracket, so as to realize the coordinated motion control of the exoskeleton and the bracket, and thus complete the gait planning.

[0018] An adaptive coordinated motion planning method for an exoskeleton and a safety bracket provided by the present invention takes the desired landing point position of the lower-limb exoskeleton as the input, uses the spatial quantization dynamic gait method based on the cosine method to adaptively generate the gait pattern of the lower-limb exoskeleton, then uses the reduced-order dynamic model to model the dynamic motion of the centroid at different walking speeds, and obtains the optimal solutions of the adaptive landing point position and walking speed by constructing an objective function, so as to realize the coordinated motion control of the exoskeleton and the walking bracket. It solves the problems of coordinated motion control of the human-exoskeleton-walking bracket system and the adaptability of the walking bracket to different walking speeds. Description of the Drawings

[0019] Figure 1 is the overall framework diagram of the adaptive coordinated motion planning method for the exoskeleton and the safety bracket in the embodiment;

[0020] Figure 2 is the straight-knee kinematic schematic diagram based on the cosine method in the embodiment;

[0021] Figure 3Schematic diagram of calculating the ankle position from the landing point position in the embodiment;

[0022] Figure 4 Schematic diagram of the trolley swing model in the embodiment;

[0023] Figure 5 Schematic diagram of the human-exoskeleton-walking support model in the embodiment;

[0024] Figure 6 Schematic diagram of the walking support joint in the embodiment;

[0025] Description of the drawings: 1 is the first wheel, 2 is the second wheel, 3 is the third wheel, 4 is the fourth wheel, and 5 is the linear joint.

[0026] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0027] As Figure 1 shown, an adaptive collaborative motion planning method for an exoskeleton and a safety support according to an embodiment of the present invention includes the following steps:

[0028] Step 1, generating the centroid trajectory of the lower limb exoskeleton:

[0029] The front and rear leg knee joints are fully extended during the toe-off, heel-strike, and transition phases between the two phases. Based on this feature, the cosine method is used to model the legs of the lower limb exoskeleton. As Figure 2 shown: Let a be the length of the front leg of the lower limb exoskeleton, b be the length of the rear leg of the lower limb exoskeleton, and c be the connecting line between the ankle of the front leg and the ankle of the rear leg of the lower limb exoskeleton, that is, the relative distance between the front leg ankle joint and the rear leg ankle joint. When the knee joint is fully flexed and extended, a = b, and the three sides a, b, and c form a triangle. α is the angle formed by side a and side c, β is the angle formed by side c and the horizontal line passing through the ankle of the front leg, and θ 1 and are the specified angles between the foot of the front leg and the ground, and θ 2 is the specified angle between the foot of the rear leg and the ground; when in the heel-strike phase, θ 1 = 0, and θ 2 > 0; when in the toe-off phase, θ 1 > 0 and θ 2 = 0; and when transitioning from the heel-strike phase to the toe-off phase, θ 1 > 0 and θ 2 > 0. The relationship between angle α and a, b, and c is described based on the cosine method as:

[0030]

[0031] Then the angle α can be obtained through calculation:

[0032]

[0033] On the other hand, assume that in the Cartesian space, the ankle position of the hind leg of the lower limb exoskeleton is P 1 = [x 1 y 1 z 1 T , and the ankle position of the front leg of the lower limb exoskeleton is P 2 = [x 2 y 2 z 2 T , so the following formula can be obtained:

[0034]

[0035] It should be noted that: the ankle positions of the front and hind legs both change with θ 1 and θ 2 , so the heel contact phase, toe off phase, and the transition period from heel contact to toe off can all be realized through the changes of θ 1 and θ 2 . Therefore, the hip joint position P hip is:

[0036] P hip = P S + [-acos(α + β) 0 asin(α + β)] T (4)

[0037] Among them, P S is the ankle position of the supporting leg during the transition from heel contact to toe off phase, and this position is set artificially. According to the corresponding relationship between the center of mass and the hip joint movement, the center of mass is set at the middle position of the two hip joints of the lower limb exoskeleton, and then according to the ankle position P 1 of the hind leg of the lower limb exoskeleton, the ankle position P 2 of the front leg of the lower limb exoskeleton, and the hip joint position P hip , the center of mass trajectory of the entire lower limb exoskeleton during walking can be calculated.

[0038] ​​As is well known, in the gait planning method of straight-knee walking, in terms of its swing phase, the position of the lower limb exoskeleton foot is generated by a trajectory generation method based on the established landing position of each step of the foot, leg length, and angles α and β. The joint angles of both legs are calculated by inverse kinematics. That is to say, the gait pattern of the lower limb exoskeleton is generated based on the given landing point position (such as the vertical projection of the ankle joint on the ground). Therefore, after generating the centroid trajectory of the lower limb exoskeleton in step 1 of this embodiment, it is also necessary to complete the conversion from the desired landing point position to the ankle joint position through step 2 to generate the gait trajectory of the swing leg of the lower limb exoskeleton and calculate all joint angles.

[0039] Step 2: Generate all joint angles of the lower limb exoskeleton

[0040] The schematic diagram of calculating the ankle joint position from the desired landing point position is as Figure 3 shown. Figure 3 In it, P 1 ′ is the ankle joint at the toe-off phase, P 2 ′ is the ankle joint at the heel-strike phase, A is the preset landing point position, B is the position of the toe at the toe-off phase or the position of the heel at the heel-strike phase, C is the vertical projection of the ankle joint on the foot, D is the connection point of the line between the ankle joint and its vertical projection point on the ground and the foot. Therefore, we get:

[0041] d CD = htanθ 1 (5)

[0042] Where d CD is the distance between C and D, and h is the height of the ankle.

[0043] When at the toe-off phase as Figure 3 (a), the position of the ankle joint P 1 ′ is calculated as follows:

[0044]

[0045] Where P A is the position of point A, d AB is the distance between A and B, and d AB = d BC ;

[0046] When at the heel-strike phase as Figure 3 (b), the position of the ankle joint P 2 ′ is calculated as follows:

[0047]

[0048] Similarly, d AB = d BCand d CD = htanθ 2 (8)

[0049] The joint angles during the transition stage (stance phase) from the heel strike phase to the toe off phase are determined according to the specified landing point position and θ 1 and θ 2 are determined; on this basis, the joint angles during the transition stage (swing phase) from the toe off phase to the heel strike phase are determined by using the trajectory generation method, so as to generate the gait trajectory of the swing leg of the lower limb exoskeleton. Then, all joint angles of the lower limb exoskeleton are calculated according to the position of the supporting leg, the centroid trajectory, and the gait trajectory of the swing leg during the swing phase.

[0050] During the generation process of the gait trajectory of the swing leg above, the combination of complex motion patterns and adaptability in the gait is allowed, enhancing the anthropomorphism of the movement.

[0051] Step 3: Perform spatial discretization on the centroid trajectory and joint angles:

[0052] The centroid trajectory is spatially discretized to form the centroid trajectory in the airspace. In this embodiment, the centroid trajectory is spatially discretized by using the uniform incremental length of the centroid horizontal displacement. The uniform incremental length of the centroid horizontal displacement Δx = 0.0001 m, and the discretization formula is:

[0053] x i = Δx·i (9)

[0054] where i = 0, 1, 2,..., N represents the discrete index, and N represents the total number of data points corresponding to multiple steps.

[0055] In essence, the joint angle can be characterized as a function of the horizontal displacement of the centroid in the spatial domain, which is called spatial quantization gait (SQG). Therefore, the joint angles in this embodiment can also be discretized in the spatial domain according to the set centroid horizontal displacement.

[0056] Step 4: Establish the trolley pendulum dynamic model of the human-exoskeleton system, construct its optimization equation, convert the centroid trajectory and joint angles in the airspace into the time domain, and generate a variable-speed straight-knee anthropomorphic gait trajectory:

[0057] In the foregoing, since the joint angle is expressed as a function of the horizontal displacement of the center of mass in the spatial domain, it is necessary to convert the generated joint angle from the spatial domain to the time domain. The joint angle after the time domain conversion is used as the control input of the lower limb exoskeleton, and the position of the corresponding center of mass changing with time determines the control of the walking support. This integrated method enables the exoskeleton robot and the walking support to perform cooperative motion control and adapt to various walking speeds. In order to achieve a natural walking posture and select the optimal horizontal displacement, this embodiment adopts a reduced-order dynamic model to model the human-exoskeleton system. Specifically:

[0058] As Figure 4 shown, regarding the assisting force of the walking support as an external force, the human-exoskeleton system is modeled as a cart-pendulum model, and its center of mass is located at the midpoint between the two hip joints of the lower limb exoskeleton. The dynamic description of the cart-pendulum model established based on the reduced-order dynamic model is:

[0059]

[0060] where z is the average height of the center of mass position during walking, p is the position of the landing point, and x represents the center of mass position.

[0061] The above formula can be solved in the time domain. Let i be the discretized position index, and the dynamic process of the cart-pendulum model is discretized using a constant horizontal displacement Δx′ of the center of mass. Then the i-th center of mass position is:

[0062] x i = Δx′·i (i = 0, 1, 2, …) (11)

[0063] where Δx′ = 0.001 m. Assuming that the velocity v i of the center of mass is known, then the time Δt i from x i+1 to x i is expressed as:

[0064]

[0065] On this basis, the velocity of the center of mass at the next index is calculated using formula (13), that is, formula (13) is used to determine the velocity of the center of mass in the subsequent spatial steps:

[0066]

[0067] To prevent singularities, the constraint condition for v i is:

[0068] ‖v i ‖> ∈ (14)

[0069] Among them, ∈ is a positive constant. To ensure numerical stability and accuracy in the calculation, in this embodiment, ∈ = 0.001 m / s.

[0070] To satisfy Spatially Quantized Dynamics (SQG) and track the reference trajectory generated by SQG, this embodiment formulates the optimization problem of the cart-pendulum model dynamics as follows:

[0071]

[0072] Where the ideal walking speed, is the reference position of the zero moment point, that is, the landing position of the desired zero moment point; x i is the displacement of the center of mass; v i and p i are both the optimized output results; v i represents the reference walking speed at index i, generated by formula (15); p i represents the landing position at index i; λ is a scalar coefficient used to adjust the balance and coordination relationship between the accuracy of the zero moment point and the speed curve.

[0073] Based on the generated reference walking speed, given the landing position, the time of the entire walking process is calculated as follows:

[0074]

[0075] So far, the discrete center-of-mass trajectory and joint angles are converted from the spatial domain to the temporal domain. When the center-of-mass trajectory and joint angles are successfully converted from the spatial domain to the temporal domain, it means that a straight-knee anthropomorphic gait pattern that can adapt to different walking speeds is generated accordingly.

[0076] Step 5: Use the center-of-mass trajectory obtained in Step 1 as the reference trajectory, and calculate the following positions of the bracket wheels and linear joints according to the reference trajectory to achieve the coordinated motion control of the exoskeleton and the bracket:

[0077] Figure 5 is a schematic diagram of the human-exoskeleton-walking bracket model in the embodiment, Figure 6 is a schematic diagram of the joints of the walking bracket in the embodiment. As Figure 5 - Figure 6 shown, in the human-exoskeleton-bracket system: each leg of the lower-limb exoskeleton is equipped with three rotary joints, and the three rotary joints correspond to the hip joint, knee joint, and ankle joint respectively;

[0078] The support includes four wheels and a linear joint for connecting and supporting the human-exoskeleton. The four wheels are the first wheel 1, the second wheel 2, the third wheel 3, and the fourth wheel 4 respectively. Each of the four wheels is provided with a rotary joint that rotates around the Y-axis. Among them, the first wheel 1 and the third wheel 3 are the front wheels, and the second wheel 2 and the fourth wheel 4 are the rear wheels. The linear joint 5 can reciprocate along the z-axis. The four wheels can either move passively or be actively controlled collaboratively, enabling the support to follow the ground movement of the human-exoskeleton system. The linear joint 5 is fixed to the backpack of the lower-limb exoskeleton through a cantilever beam to support part of the weight of the human-exoskeleton system and make the vertical displacement of the center of mass of the lower-limb exoskeleton smoother during movement.

[0079] Let H be the vertical offset between the linear joint and the midpoint of the line connecting the centers of the two rear wheels, L represent the horizontal distance from the linear joint to the center of mass of the human-exoskeleton system, R be the radius of the wheel, and D 1 be the horizontal distance from the linear joint to the front wheel, and D 2 be the horizontal distance from the linear joint to the rear wheel. Assume the position of the center of mass is:

[0080] P COM = [x, y, z] T (17)

[0081] Determine the position of the linear joint according to the position of the center of mass:

[0082] P pj = P COM + [-L, 0, 0] T (18)

[0083] Based on the synchronous change of the vertical position of the linear joint and the position of the center of mass of the human-exoskeleton system, calculate the positions of the four wheels as follows:

[0084]

[0085] Among them, P w1 is the position of the first wheel 1, P w2 is the position of the second wheel 2, P w3 is the position of the third wheel 3, and P w4 is the position of the fourth wheel 4. W is the width of the support, that is, the distance between the centers of the two front wheels.

[0086] Similarly, the joint angles of each wheel are spatially discretized by the horizontal displacement change Δx, and the calculation method is:

[0087]

[0088] Based on the centroid trajectory of the human-exoskeleton system in the airspace, the corresponding wheel reference joint angles are calculated. The reference trajectories of the centroid, linear joints, and four-wheel joints of the human-exoskeleton system are formulated from the joint angles collected from healthy subjects and the specified desired walking speed.

[0089] Thus, the exoskeleton straight-knee anthropomorphic gait pattern can be generated according to the desired landing point position and the desired walking speed, and the coordinated control of the lower-limb exoskeleton and the walking support is achieved.

[0090] Finally, it should be noted that the trajectory generation method adopted in this embodiment can be an interpolation method, an imitation learning method, or other trajectory generation methods.

Claims

1. An adaptive collaborative motion planning method for an exoskeleton and a safety support, the method being applicable to a human-lower limb exoskeleton-support system, wherein the support in the human-lower limb exoskeleton-support system comprises a driving wheel and a linear joint for connecting and supporting the human-exoskeleton, and characterized in that: The implementation steps of this method include: Step 1: Generate the center of mass trajectory of the lower limb exoskeleton using a spatial quantized dynamic gait generation method based on the cosine method; Step 2, in the swing phase: based on the expected foothold, using the center of mass trajectory generated in step 1, a trajectory generation method is used to generate a gait trajectory of the swing leg of the lower limb exoskeleton; all joint angles of the lower limb exoskeleton are calculated according to the support leg position, the center of mass trajectory, and the gait trajectory of the swing leg in the swing phase; Step 3, using the preset uniform increment length of the center of mass horizontal displacement to perform spatial discretization on the center of mass trajectory, and using the preset center of mass horizontal displacement to perform spatial discretization on the joint angle; Step 4: Establish the pendulum dynamics model of the human-exoskeleton system and construct its optimization equation, convert the center of mass trajectory and joint angle in the airspace into the time domain, and generate a variable-speed straight-knee anthropomorphic gait trajectory: 4.

1. Based on the reduced-order dynamics model, a trolley pendulum dynamics model for the human-exoskeleton-support system is established, and the horizontal displacement of the center of mass with spatial transformation is calculated using the trolley pendulum dynamics model; 4.

2. Construct the optimization problem equation of the trolley pendulum dynamics model, take the expected walking speed as input, and use the optimization problem equation to calculate the horizontal displacement of the center of mass over time; 4.

3. The point where the horizontal displacement of the center of mass that varies with space and the horizontal displacement of the center of mass that varies with time overlap is determined as the horizontal x-coordinate of the center of mass that varies with time; 4.

4. Calculate the horizontal displacement time per unit length of the center of mass according to the horizontal velocity of the center of mass, convert the gait of the lower limb exoskeleton in the airspace into the time domain according to the horizontal displacement time per unit length and the horizontal x-coordinate of the center of mass that changes with time, and generate a variable-speed straight-knee anthropomorphic gait trajectory; Step 5: Use the center of mass trajectory obtained in step 1 as the reference trajectory, calculate the reference position of the bracket driving wheel and the linear joint according to the reference trajectory, and input it into the controller of the bracket driving wheel and the linear joint to achieve coordinated motion control of the exoskeleton and the bracket.

2. The method for adaptive collaborative motion planning of an exoskeleton and a safety support according to claim 1, characterized in that: The implementation method of step 1 is: Step 1.1, based on the characteristics that the front leg knee joint and the rear leg knee joint are in a fully extended state when the toes leave the ground, the heels touch the ground, and the transition stage between the two phases, the cosine method is used to calculate the ankle joint position and hip joint position of the front and rear legs of the lower limb exoskeleton robot; Step 1.2: According to the corresponding relationship between the center of mass and the movement of the hip joint, the center of mass is set at the middle position of the two hip joints of the lower limb exoskeleton. The position of the ankle joints and the hip joints of the front and rear legs of the lower limb exoskeleton are used to generate the center of mass trajectory of the entire lower limb exoskeleton during walking.

3. The method for adaptive collaborative motion planning of an exoskeleton and a safety support according to claim 2, characterized in that: The implementation method of generating the swing leg gait trajectory in step 2 is: Assume that P1′ is the ankle joint at the toe-off phase, P2′ is the ankle joint at the heel-touching phase, A is the preset foothold position, B is the position of the toe at the toe-off phase or the position of the heel at the heel-touching phase, C is the vertical projection of the ankle joint on the foot, and D is the connection point between the line connecting the ankle joint and its vertical projection point on the ground and the foot. Then the distance between C and D is: d CD =htanθ1 Among them, d CD is the distance between C and D, h is the height of the ankle; When in the toe-off phase, the ankle joint P1′ position is calculated as follows: Where P A is the position of point A, d AB is the distance between A and B, and d AB =d BC ; When in the heel-strike phase, the ankle joint P2′ position is: d AB = d BC and d CD = htanθ2 The joint angles during the transition from the heel strike phase to the toe lift phase are determined based on the specified footfall position and θ1 and θ2; According to the ankle joint P1′, ankle joint P2′ and the joint angles in the transition phase from the heel-touching phase to the toe-off phase, the gait trajectory of the swing leg can be generated by using the trajectory generation method; All joint angles of the lower limb exoskeleton are calculated based on the support leg position, center of mass trajectory, and gait trajectory of the swing leg in the swing phase.

4. According to the method for adaptive collaborative motion planning of an exoskeleton and a safety support according to claim 3, the implementation method of step 5 comprises: The walking frame includes four wheels and one linear joint, two of which are front wheels and the other two are rear wheels; let H be the vertical offset between the linear joint of the walking frame and the midpoint of the line connecting the two rear wheel centers, L is the horizontal distance from the linear joint of the walking frame to the center of mass of the human-exoskeleton system, R is the radius of the four wheels, D1 is the horizontal distance from the linear joint of the walking frame to the front wheel of the walking frame, and D2 is the horizontal distance from the linear joint of the walking frame to the rear wheel; suppose the center of mass position of the human-exoskeleton system is: P COM =[x,y,z] T Determine the position of the linear joint of the walking frame according to the position of the center of mass: P pj =P COM +[-L,0,0] T Based on the synchronous change of the vertical position of the linear joint of the walking frame and the center of mass of the human-exoskeleton system, the positions of the four active wheels of the walking frame are calculated as follows: Among them, P w1 is the position of the first wheel, P w2 is the position of the second wheel, P w3 is the position of the third wheel, P w4 is the position of the fourth wheel; W is the width of the bracket, that is, the wheel center distance of the two front wheels; The joint angles of each wheel are spatially discretized through the horizontal displacement change Δx, and the calculation method is: Based on the center of mass trajectory of the human-exoskeleton system in the airspace, the corresponding wheel reference joint angles are calculated; the reference trajectories of the center of mass, linear joints, and four-wheel joints of the human-exoskeleton system are set based on the joint angles collected from healthy subjects and the specified expected walking speed.

Images