Design and analysis of a lightweight lower extremity exoskeleton with novel compliant ankle joints

Abstract

BACKGROUND:

The exoskeleton for lower limb rehabilitation is an uprising field of robot technology. However, since it is difficult to achieve all the optimal design values at the same time, each lower extremity exoskeleton has its own focus.

OBJECTIVE:

This study aims to develop a modular lightweight lower extremity exoskeleton (MOLLEE) with novel compliant ankle joints, and evaluate the movement performance through kinematics analysis.

METHODS:

The overall structure of the exoskeleton was proposed and the adjustable frames, active joint modules, and compliant ankle joints were designed. The forward and inverse kinematics models were established based on the geometric method. The theoretical models were validated by numerical simulations in ADAMS, and the kinematic performance was demonstrated through walking experiments.

RESULTS:

The proposed lower extremity offers six degrees of freedom (DoF). The exoskeleton frame was designed adjustable to fit wearers with a height between 1.55 m and 1.80 m, and waist width from 37 cm to 45 cm. The joint modules can provide maximum torque at 107 Nm for adequate knee and hip joint motion forces. The compliant ankle can bear large flexible deformation, and the relationship between its angular deformation and the contact force can be fitted with a quadratic polynomial function. The kinematics models were established and verified through numerical simulations, and the walking experiments in different action states have shown the expected kinematic characteristics of the designed exoskeleton.

CONCLUSIONS:

The proposed MOLLEE exoskeleton is adjustable, modular, and compliant. The designed adjustable frame and compliant ankle can ensure comfort and safety for different wearers. In addition, the kinematics characteristics of the exoskeleton can meet the needs of daily rehabilitation activities.

Keywords

Lower extremity exoskeleton walking rehabilitation lightweight structure modular design compliant joint kinematics analysis

1. Introduction

As stroke and spinal cord injury (SCI) patients increase each year, the demand for physical rehabilitation services is growing progressively, but there is a serious lack of physiotherapists [1]. To help the paralyzed individuals restore walking ability as much as possible, a variety of wearable medical and rehabilitation lower extremity exoskeletons have been developed [2]. Due to the advantages of standard and repetitive robotic gait training over manual rehabilitation therapy, various medical exoskeletons have been widely used in hospitals and rehabilitation centers in recent years [3, 4].

Over the past two decades, research in the field of lower limb rehabilitation exoskeletons has grown rapidly [5]. The well-known commercial lower extremity exoskeleton, like the ReWalk [6], Ekso [7], Indego [8], and HAL [9], were designed for those with mobility disabilities, such as paraplegia, quadriplegia, and hemiplegia. Other similar exoskeletons, like the Mindwalker [10], Mina [11], ExoAtlet [12], and AIDER [13], have also been developed for post-stroke, paraplegic, and SCI patients. However, some of the technical issues, such as weight constraints, maximum torque, and structure adjustability, have limited the applicability of the existing lower extremity exoskeletons [14]. To solve the above problems, some exoskeletons have been designed with off-board motors and flexible transmissions to reduce the device weight [15, 16]. To increase the output torque of the exoskeleton joints, some studies developed the hydraulic actuating system [17, 18] and series elastic actuator [19, 20]. To fit wearers with different body sizes and shapes, the length of the waist, thigh, and shank segments are manually adjustable on many of the lower extremity exoskeletons [21, 22].

Generally, to obtain a lightweight and small shape lower extremity exoskeleton, actuators are mounted to the hips and knees for the active flexion/extension movements, while the ankles use passive or sprung joints [23]. Compared to the conventional rigid hinge joint, the flexure joint is a flexible structure that can produce movement through its elastic deformation, which has advantages of fewer components, small size, clearance-free, frictionless motion, reduced weight and inertia [24, 25]. Therefore, the lower extremity exoskeleton with compliant ankle joints can minimize the reaction force due to instant impact over the ground, thus improving the safety of human-exoskeleton interaction.

The main objective of this work is to design a lightweight lower extremity exoskeleton with compliant ankle joints from the bionic perspective. In this paper, considering the design requirements of applicability and modularity, the adjustable exoskeleton frames and joint modules were proposed, and a novel compliant ankle joint was designed with flexible deformation. Then, the forward and inverse kinematics analysis was presented to determine the pose and moving joint angles of the exoskeleton, respectively. Further, numerical simulations and practical experiments were conducted to verify the validity of theoretical kinematics models and illustrate the characteristics of the compliant ankle joint.

2. Exoskeleton design methods

2.1 Design requirements

In order to provide multi-rehabilitation training functions for different patient groups, we have proposed the concept of a modular lightweight lower extremity exoskeleton (MOLLEE). Before starting the exoskeleton design, the specified design requirements are as follows:

(1)
Structure adjustability: To improve the applicability, on the basis of Chinese Standard ‘GB/T 10000-1988 Human Dimensions of Chinese Adults’, the exoskeleton structure should be adjustable for wearers with a height between 1.55 m and 1.80 m and the weight under 100 kg.
(2)
Modular design: For a simplified design and easy maintenance, the exoskeleton joint units need to be modular, and the structure should be easy to (dis)assemble.
(3)
Joint torques: To ensure adequate torques for rehabilitation motion assistance, the minimum torque of actuated joints was defined as 80 Nm, which is based on the data of a healthy adult with 80 kg weight and 1.8 m height [26].
(4)
Range of motion (ROM): For the complex motion rehabilitation tasks, a sufficient ROM of each movement joint is needed regarding different actions, such as sit, stance, and walking.

Figure 1.
Structure of the lower extremity exoskeleton: (a) Prototype, (b) Exploded view of the CAD model.

(5)
Walking speed: To achieve a natural walking gait, the exoskeleton is expected to generate bipedal walking over the ground with a maximum speed of 0.75 m/s [27].

2.2 Overall structure

According to design requirements, the MOLLEE exoskeleton was developed with one active DoF at the hip, one active at the knee, and one passive at the ankle per leg (Fig. 1a). Specifically, to fully support the lower limbs of paralyzed patients, the designed exoskeleton structure is rigid and parallel to the human legs, which mainly includes the backpack, frame links, joint units, binding belts, and footplates (Fig. 1b). The concept design contains four actuated units at the knee joints and hip joints, which can provide active flexion/extension motion assistance in the sagittal plane. In addition, frame lengths of the waist, thigh, and shank links can be adjusted to fit various body sizes and shapes of different wearers.

Table 1
Comparison of technical data between the existing lower extremity exoskeletons and expected design of MOLLEE

Exoskeleton	DoF	Actuator	Max walking speed	Size restrictions	Weight
ReWalk 6.0	6	4 rotary	0.6 m/s	100 kg, 160–190 cm	23 kg
Ekso GT	6	4 linear	0.55 m/s	100 kg, 158–188 cm	20 kg
Indego	4	4 rotary	0.55 m/s	113 kg, 155–191 cm	12 kg
HAL-5	6	4 rotary	0.55 m/s	100 kg, 150–190 cm	14 kg
MOLLEE	6	4 rotary	0.75 m/s	100 kg, 155–180 cm	16 kg

Table 2

Parameters of the modular actuator unit for rotation output

Continuous torque	Peak torque	Max speed	Gear ratio	Diameter	Axial length	Mass	Communication bus
52 Nm	107 Nm	30 rpm	100	90 mm	102 mm	2 kg	EtherCAT

The comparison of technical data between the existing lower extremity exoskeletons and expected design MOLLEE are listed in Table 1. Moreover, the exoskeleton frame uses carbon fiber and 7075 aluminum alloy materials for a lightweight structure.

2.3 Adjustable frame

The exoskeleton frame can connect the hip, knee, and ankle joints, and it will transfer motion and force to the human lower limb through binding belts. However, because the body sizes and shapes of different wears vary greatly, the length of the waist, thigh, and shank frames should be adjustable to align the exoskeleton joints with human lower limb joints. Therefore, the adjustable frame of the exoskeleton was designed with a telescopic structure, and the adjusted step length was engraved on the telescopic rod (Fig. 2a).

Figure 2.

Adjustable frame design: (a) Telescopic structure, (b) Clamp mechanism.

Figure 3.

Adjustable dimensions of the exoskeleton frames: (a) Side view, (b) Rear view.

The clamp mechanism of the adjustable frame consists of the outer rod, location pin, connector, lock nut, and inner rod (Fig. 2b), which is simple in structure and reliable to adjust the frame length. In addition, a series of small holes were provided on the surface of the rod for positioning through a location pin. Consequently, the exoskeleton frame can be adjusted to accommodate wearers with a body height between 1.55 m and 1.80 m, and the waist width between 37 cm and 45 cm (Fig. 3).

Figure 4.

Joint modules design: (a) Hip joint module, (b) Knee joint module.

2.4 Joint module

The joint module is an electric actuator that can provide rotational torque to the connection joint. For a streamlined installation and maintenance of the exoskeleton, the hip and knee joint module units (Fig. 4) have been developed based on commercial rotary actuators. Further, to achieve the required joint torque of 80 Nm, a modular sensor-actuator unit for rotation output was applied, which includes e.g. a brushless direct current (BLDC) motor with friction brake, a harmonic drive gear unit, a servo driver, incremental and absolute position encoders, and EtherCAT communication bus.

The adopted rotary actuator can bring large joint torque, and its performance parameters are listed in Table 2. In addition, for the human-exoskeleton interaction security, the limit block was adopted in the joint module for mechanical position limitation. The corresponding ROM of hip and knee joint modules are $-$ 90 ${}^{\circ}$ (Flexion) $\sim$ 25 ${}^{\circ}$ (Extension) and 0 ${}^{\circ}$ (Extension) $\sim$ 90 ${}^{\circ}$ (Flexion), here the positive and negative senses of rotation are assumed by the anatomical convention [11], as shown in Fig. 3.

2.5 Compliant ankle

Compliance is very critical for a safe interaction between the exoskeleton, humans, and the environment. In this work, to achieve a passive ankle dorsi/plantar-flexion movement, the single-DoF compliant ankle joint that relies on the elastic deformation of its flexible beams was proposed (Fig. 5a). The compliant ankle can undergo large deflections and minimize the ground impact to the exoskeleton, which differs from a traditional sprung ankle joint.

Figure 5.

Compliant ankle design: (a) CAD model, (b) Structure parameters.

The concept of a single-axis compliant ankle joint was constructed by two curved beams, and its rotation axis was approximately located in the center of the cross-section at the position of minimum thickness, with a large range of motion (Fig. 5b). To study the mechanical properties of the compliant joint, such as rotation capacity, maximum stress and strain, the structure parameters need to be determined, including the height $h$ , width $w$ , curved beam thickness $t$ , and the minimum distance between two beams $l$ . The material used here is the spring steel of 60Si2MnA.

Figure 6.

Displacement comparisons of the compliant ankle joint with variable parameters using finite element analysis: (a) $l_{1}=$ 1 mm, (b) $l_{2}=$ 3 mm, (c) $h_{1}=$ 56 mm, (d) $h_{2}=$ 76 mm, (e) $t_{1}=$ 1.5 mm, (f) $t_{2}=$ 2.5 mm, (g) $w_{1}=$ 15 mm, (h) $w_{2}=$ 25mm.

In order to predict the deformation behavior of the designed compliant ankle joint, we use finite element analysis (FEA) to compare the characteristics with a series of variable structure parameters (Fig. 6). The lower rigid support of the compliant ankle joint was fixed, meanwhile, the upper rigid support was subjected to a force perpendicular to the cylindrical surface with $F=$ 300 N. The initial structural parameters were set as: $h_{0}=$ 66 mm, $l_{0}=$ 1 mm, $t_{0}=$ 2 mm, $w_{0}=$ 20 mm. During the FEA simulation, we have changed one of the structural parameters individually to see the variation of maximum deformation $d_{\textit{max}}$ of the upper rigid support. The loads and displacements simulation results indicate that the parameters of $h$ and $w$ are proportional to the value of $d_{max}$ , while the parameters of $t$ and $l$ are inversely proportional to the value of $d_{max}$ . Besides that, $t$ and $h$ can affect the stiffness of the compliant ankle joint significantly. Therefore, based on the FEA method, the designed compliant ankle joint has shown low bending stiffness and high rotational deflection.

Figure 7.

Kinematics model of the exoskeleton: (a) Simplified 5-link model, (b) Geometric constraints of the knee joint.

3. Kinematics analysis

The kinematics model of the proposed exoskeleton can be simplified as a 5-link model (Fig. 7a). The coordinates are defined as: $O_{w}-x_{w}y_{w}z_{w}$ is the world coordinate frame, $O_{c}-x_{c}y_{c}z_{c}$ is the waist coordinate frame. The thigh length of the exoskeleton is $L_{t}$ , the shank length is $L_{s}$ , and the distance between the hip joint and waist center is $L_{w}$ . In addition, the rotation angles of the active hip and knee joints are represented as $\theta_{1}\sim\theta_{4}$ .

3.1 Forward kinematical modeling

The purpose of the forward kinematical modeling is to obtain the left foot position P ${}_{L}$ and right foot position P ${}_{R}$ in the world frame when the waist position P ${}_{c}$ and attitude R ${}_{c}$ , and the moving joint angles $\theta_{1}\sim\theta_{4}$ are given. For the left leg of the exoskeleton (Fig. 7b), the left foot position ${}^{c}{\bf P}_{L}$ in the waist frame can be written as:

$\displaystyle\left\{{\begin{array}[]{l}{}^{c}{{\bf P}}_{Lx}=L_{t}\sin\left({-% \theta_{2}}\right)-L_{s}\cos\left({\theta_{1}+\theta_{2}}\right)\\ {}^{c}{{\bf P}}_{Ly}=\;L_{w}\\ {}^{c}{{\bf P}}_{Lz}=-L_{t}\cos\left({-\theta_{2}}\right)-L_{s}\cos\left({% \theta_{1}+\theta_{2}}\right)\\ \end{array}}\right.$ (1)

where $\theta_{1}\in\left[{0^{\circ}\sim 90^{\circ}}\right]$ , $\theta_{2}\in\left[{-90^{\circ}\sim 25^{\circ}}\right]$ .

Then, the left foot position P ${}_{L}$ in the world frame can be calculated by:

$\displaystyle{\bf P}_{L}={\bf P}_{c}+{\bf R}_{c}\cdot^{c}{\bf P}_{L}$ (2)

Because the left and right legs are symmetrically distributed, the right foot position P ${}_{R}$ in the world frame can also be calculated by the same method as follows:

$\displaystyle\left\{{\begin{array}[]{l}{}^{c}{{\bf P}}_{Rx}=L_{t}\sin\left({-% \theta_{3}}\right)-L_{s}\cos\left({\theta_{3}+\theta_{4}}\right)\\ {}^{c}{{\bf P}}_{Ry}=\;-L_{w}\\ {}^{c}{{\bf P}}_{Rz}=-L_{t}\cos\left({-\theta_{3}}\right)-L_{s}\cos\left({% \theta_{3}+\theta_{4}}\right)\\ \end{array}}\right.$ (3)

where $\theta_{3}\in\left[{-90^{\circ}\sim 25^{\circ}}\right]$ , $\theta_{4}\in\left[{0^{\circ}\sim 90^{\circ}}\right]$ .

Similarly, the right foot position P ${}_{R}$ in the world frame can be obtained by:

$\displaystyle{\bf P}_{R}={\bf P}_{c}+{\bf R}_{c}\cdot^{c}{\bf P}_{R}$ (4)

3.2 Inverse kinematical modeling

The purpose of the inverse kinematical modeling is to obtain the moving joint angles $\theta_{1}\sim\theta_{4}$ when the left and right foot positions of P ${}_{L}$ and P ${}_{R}$ , and the waist position P ${}_{c}$ and attitude R ${}_{c}$ are given. For the left leg of the exoskeleton, the foot position ${}^{c}{\bf P}_{L}$ in the waist frame can be calculated as:

${}^{c}{\bf P}_{L}=R_{c}^{-1}\left({\bf P}_{L}-{\bf P}_{c}\right)$ (5)

According to the geometric constraints of the knee joint (Fig. 7b), the angles $\gamma$ and $\beta$ can be calculated based on the law of cosines:

$\displaystyle\left\{{\begin{array}[]{l}\gamma=\arccos\left({\frac{L_{t}^{2}+L_% {s}^{2}-^{c}{\bf P}_{Lx}^{2}-^{c}{\bf P}_{Lz}^{2}}{2L_{t}L_{s}}}\right)\\ \beta=\arccos\left(\frac{{}^{c}{\bf P}_{Lx}^{2}+^{c}{\bf P}_{Lz}^{2}+L_{t}^{2}% -L_{s}^{2}}{2L_{t}\sqrt{{}^{c}{\bf P}_{Lx}^{2}+^{c}{\bf P}_{Lz}^{2}}}\right)\\ \end{array}}\right.$ (6)

Then, the knee joint angle $\theta_{1}$ and hip joint angle $\theta_{2}$ can be obtained by:

$\displaystyle\left\{{\begin{array}[]{l}\theta_{1}=\pi-\gamma\\ \theta_{2}=\varphi-\beta\\ \end{array}}\right.$ (7)

where $\varphi=\arctan\left({\raise 3.01pt\hbox{${{}^{c}{\bf P}_{Lx}^{2}}$}\!\mathord% {\left/{\vphantom{{{}^{c}{\bf P}_{Lx}^{2}}{{}^{c}{\bf P}_{Lz}^{2}}}}\right.% \kern-1.2pt}\!\lower 3.01pt\hbox{${{}^{c}{\bf P}_{Lz}^{2}}$}}\right)$ .

Similarly, the hip joint angle $\theta_{3}$ and knee joint angle $\theta_{4}$ of the right leg can also be obtained by the same method as follows:

$\displaystyle\left\{{\begin{array}[]{l}\theta_{3}=\arctan\left({\raise 3.01pt% \hbox{${{}^{c}{\bf P}_{Rx}^{2}}$}\!\mathord{\left/{\vphantom{{{}^{c}{\bf P}_{% Rx}^{2}}{{}^{c}{\bf P}_{Rz}^{2}}}}\right.\kern-1.2pt}\!\lower 3.01pt\hbox{${{}% ^{c}{\bf P}_{Rz}^{2}}$}}\right)-\arccos\left({\frac{{}^{c}{\bf P}_{Rx}^{2}+{}^% {c}{\bf P}_{Rz}^{2}+L_{t}^{2}-L_{s}^{2}}{2L_{t}\sqrt{{}^{c}{\bf P}_{Rx}^{2}+^{% c}{\bf P}_{Rz}^{2}}}}\right)\\ \theta_{4}=\pi-\arccos\left({\frac{L_{t}^{2}+L_{s}^{2}-^{c}{\bf P}_{Rx}^{2}-^{% c}{\bf P}_{Rz}^{2}}{2L_{t}L_{s}}}\right)\\ \end{array}}\right.$ (8)

4. Simulation and experiment results

In this section, the theoretical models and kinematics performance of the designed exoskeleton are evaluated by numerical simulations and prototype experiments.

Figure 8.

Compliant ankle joint experiment illustration: (a) Experiment setup, (b) Experimental measurement, (c) Relationship between the compliant ankle joint angles and the contact forces.

4.1 Compliant ankle joint experiment illustration

In order to illustrate the characteristics of the compliant ankle joint, an experiment is conducted to explore the relationship between the joint angle and the contact force. As Fig. 8a shows, the lower part of the exoskeleton leg, containing the foot, the passive ankle and the shank link, is fixed on a table. The experimenter applies a series of pulling forces $F$ , which can be tested with a forcemeter, at the vertex of the shank link to mimic the contact force during wearing walking, then the angular deformation $\phi$ of the passive ankle can be recorded, shown as Fig. 8b. To determine the structural parameters, a series of ankle joints with different size parameters are designed and manufactured. After that, these ankle joints are analyzed with simulation and experiments in different walking environments, such as level walking, walking upstairs and downstairs. The structural parameters are determined by the lightest weight on the premise of satisfying the requirements of strength and deformation angle. Finally, the structural parameters of the compliant ankle are set as: $h_{0}=$ 66 mm, $l_{0}=$ 1 mm, $t_{0}=$ 2 mm, $w_{0}=$ 20 mm. The pulling force $F$ increases from 0 N to 100 N at intervals of 10 N, and the caused angular deformation $\phi$ are recorded as the red dotted line in Fig. 8c. The experiment result indicates that the relationship between the joint angle $\phi$ and the force $F$ is not linear, and it can be fitted with a quadratic polynomial function, shown as Eq. (9) and its curve is the blue solid line in Fig. 8c.

$\displaystyle\phi\left(F\right)=p_{1}\cdot F^{2}+p_{2}\cdot F+p_{3}$ (9)

where, the coefficients $p_{1}=$ 0.818, $p_{2}=$ 3.141, and $p_{3}=$ 2.933.

Figure 9.

Kinematics model validation: (a) Given joint angles of the left and right legs of the exoskeleton, (b) Calculated foot trajectories by the forward kinematics model, (c) Imported foot trajectories of the exoskeleton and kinematics simulation in ADAMS, (d) Output moving joint angles by the inverse kinematics model.

4.2 Exoskeleton walking experiment illustration

In order to validate the established kinematics models, the numerical simulation of the exoskeleton is conducted in ADAMS. Firstly, the moving joint angles of the left and right legs were given (Fig. 9a), and the corresponding foot trajectories can be calculated through the forward kinematics model (Fig. 9b). Then, the calculated foot trajectories were imported into ADAMS for the kinematics simulation (Fig. 9c), and the output moving joint angles can be obtained by the inverse kinematics model (Fig. 9d). Finally, as can be seen from the consistent results between given joint angles (Fig. 9a) and calculated output joint angles (Fig. 9d), the established theoretical forward and inverse kinematics models have been proved to be feasible.

Figure 10.

The control framework of the exoskeleton.

To illustrate the practical kinematics performance of the proposed exoskeleton, we have conducted the wearing walking experiments on a healthy person. The control framework of the exoskeleton in practical application is shown in Fig. 10. Firstly, the reference trajectories for driving the exoskeleton are generated by a personalized gait trajectory generator, whose detailed description can be seen in our previous work [28], based on the human body parameters of the wearer, such as body height, thigh length, shank length, and foot length. Then the generated reference trajectories can be divided into various sub-trajectories used in different states, including sitting, standing and walking at different speeds. The sub-trajectories are selected with finite state machines, which can be controlled by the human with the triggers placed on the crutches. The exoskeleton is controlled by a traditional PID controller, and all the motor drivers operate in position mode.

Figure 11.

Kinematics performance illustrations of the exoskeleton: (a) Sit, (b) Stance, (c) Lean forward, (d) Local view of compliant ankle deformation, (e) Left thigh flexion with $-$ 45 ${}^{\circ}$ , (f) Left shank flexion with 50 ${}^{\circ}$ , (g) Right thigh flexion with $-$ 45 ${}^{\circ}$ , (f) Right shank flexion with 50 ${}^{\circ}$ .

Based on the above structure design and kinematics analysis, the prototype of MOLEE was built for practical experiments. The experiment trials mainly include sit, stance, lean forward, left leg extension/flexion, and right leg extension/flexion phases (Fig. 11). During the initial walking process, the body will lean forward about 8 degrees, and we can observe a flexible deformation of the compliant ankle joint (Fig. 11d). The human-exoskeleton walking experiment results illustrate that the exoskeleton can meet the rehabilitation tasks for daily activities.

Figure 12.

Comparison of simulated joint angles $\theta_{i}$ and experimental joint angles $M_{i}$ and their angular deviations $e_{i}$ in motion trials: (a) Left thigh flexion, (b) Left shank flexion, (c) Right thigh flexion, (d) Right shank flexion.

The joint angles for driving the exoskeleton to realize left and right leg extension/flexion motions in Fig. 11(e $\sim$ h) are shown in Fig. 12(a $\sim$ d). The dotted lines $\theta$ ${}_{i}$ ( $i=$ 1 $\sim$ 4) are the calculated joint angles trajectories by the kinematics model, the solid lines $M_{i}$ are the recorded ones from the motor drivers, and $e_{i}$ are the angular deviations between them. The comparison results illustrate that $\theta_{i}$ are almost consistent with $M_{i}$ , and the maximum of the deviations is about 0.5 ${}^{\circ}$ . Thus, the performance of the driving motors can meet the use requirements of the exoskeleton. The joint torques, which are calculated through the motor current and joint parameters, for driving the exoskeleton to realize left and right leg extension/flexion motions in Fig. 11(e $\sim$ h) are shown in Fig. 13 (a $\sim$ d). The results indicate that the torque of each joint varies from $-$ 40 Nm to 40 Nm, and the joint torques of the swing leg are larger than those of the support leg.

Figure 13.

Experimental joint torques $T_{i}$ in motion trials: (a) Left thigh flexion, (b) Left shank flexion, (c) Right thigh flexion, (d) Right shank flexion.

In conclusion, since the flexible structure of compliant ankle joints, the exoskeleton has the characteristics of compliance during walking, which can minimize ground contact impact and improve wearing comfort. Compared to other lower extremity exoskeletons, MOLEE has smaller foot inertia, because the total weight of the ankle and footplate is only 0.4 kg. Therefore, the designed lower extremity exoskeleton has a fast dynamic response during fast walking.

5. Discussion and conclusions

This work presented the modular design and kinematic analysis of a lightweight lower extremity exoskeleton with novel compliant ankle joints. The prototype design was implemented according to the specific requirements of adjustability, feasibility, and lightweight. The adjustable exoskeleton frames allow different wearers with a body height between 1.55 m and 1.80 m, and hip width between 37 cm and 45 cm, as well as waist thickness between 15 cm and 19 cm. Based on the designed hip and knee joint modules, the exoskeleton structure was ensured to simplify the design. The novel compliant ankle joints can reduce leg inertia and minimize ground impact during the walking process. The designed joint module is able to provide a maximum torque of 107 Nm for adequate motion assistance. To analyze the kinematic characteristics of the proposed lower extremity exoskeleton, the closed-form solutions for both forward and inverse kinematics models were obtained. The validity of the theoretical models was verified through numerical simulation research. The experimental validation indicates the proposed exoskeleton is suitable for different movement activities in daily life.

In our future work, we will focus on two aspects of research. Firstly, the exoskeleton with four active joints and two passive joints is an underactuated system, and the angles of the passive ankle joints cannot be planned by the algorithm based on the kinematical model. It is necessary to study the dynamic characteristics of the exoskeleton to obtain the relationship between the passive joint angles and the interaction force during walking. Secondly, as Fig. 10 shows, the finite state machines are controlled by pressing the triggers on the crutches, which leads to delayed and unnatural gait transitions. To improve the naturalness of the walking gait, we can study the method to obtain the motion intention through the sEMG signals of the upper limb muscle.

Footnotes

Acknowledgments

The work reported in this paper is partially supported by the International Science & Technology Cooperation Program of China (number: 2018YFE0125600), in part by the NSFC-Shenzhen Robotics Research Center Project (U2013207), and in part by the Science and Technology Service Network Initiatives (Grant No. KFJ-STS-QYZX-095).

Conflict of interest

No conflict of interest exits in the submission of this manuscript. The final manuscript was approved for publication by all authors.

References

Saunders

Sanderson

Hayes

, et al. Physical fitness training for stroke patients. Cochrane Database of Systematic Reviews. 2020; 3: CD003316.

Tan

Koyama

Sakurai

, et al. Wearable robotic exoskeleton for gait reconstruction in patients with spinal cord injury: A literature review. Journal of Orthopaedic Translation. 2021; 28: 55-64.

Louie

Mortenson

Durocher

, et al. Exoskeleton for post-stroke recovery of ambulation (ExStRA): study protocol for a mixed-methods study investigating the efficacy and acceptance of an exoskeleton-based physical therapy program during stroke inpatient rehabilitation. BMC Neurology. 2020; 20: 35.

Mao

, et al. The effects of gait training using powered lower limb exoskeleton robot on individuals with complete spinal cord injury. Journal of Neuroengineering and Rehabilitation. 2018; 15: 14.

Shi

Zhang

, et al. A review on lower limb rehabilitation exoskeleton robots. Chinese Journal of Mechanical Engineering. 2019; 32: 74.

Talaty

Esquenazi

Briceno

. Differentiating ability in users of the ReWalk TM powered exoskeleton: An analysis of walking kinematics. 2013 IEEE 13th international conference on rehabilitation robotics (ICORR). Seattle, Washington USA, IEEE, 2013; 1-5.

Kolakowsky-Hayner

Crew

Moran

, et al. Safety and feasibility of using the Ekso™bionic exoskeleton to aid ambulation after spinal cord injury. Journal of Spine. 2013; S4: 003.

Tefertiller

Hays

Jones

, et al. Initial outcomes from a multicenter study utilizing the Indego powered exoskeleton in spinal cord injury. Topics in Spinal Cord Injury Rehabilitation. 2018; 24(1): 78-85.

Sczesny-Kaiser

Höffken

Aach

, et al. HAL®exoskeleton training improves walking parameters and normalizes cortical excitability in primary somatosensory cortex in spinal cord injury patients. Journal of Neuroengineering and Rehabilitation. 2015; 12: 68.

10.

Wang

Meijneke

, et al. Design and control of the MINDWALKER exoskeleton. IEEE Transactions on Neural Systems and Rehabilitation Engineering. 2014; 23(2): 277-286.

11.

Mummolo

Peng

Agarwal

, et al. Stability of Mina v2 for robot-assisted balance and locomotion. Frontiers in Neurorobotics. 2018; 12: 62.

12.

Pais-Vieira

Allahdad

Neves-Amado

, et al. Method for positioning and rehabilitation training with the ExoAtlet® powered exoskeleton. MethodsX. 2020; 7: 100849.

13.

Wang

Cheng

Hou

. c2AIDER: cognitive cloud exoskeleton system and its applications. Cognitive Computation and Systems. 2019; 1(2): 33-39.

14.

Plaza

Hernandez

Puyuelo

, et al. Lower-limb medical and rehabilitation exoskeletons: A review of the current designs. IEEE Reviews in Biomedical Engineering. 2021; doi: 10.1109/RBME.2021.3078001.

15.

Witte

Collins

. Design of lower-limb exoskeletons and emulator systems. Wearable Robotics. Academic Press, 2020; pp. 251-274.

16.

Kuan

Pasch

Herr

. A high-performance cable-drive module for the development of wearable devices. IEEE/ASME Transactions on Mechatronics. 2018; 23(3): 1238-1248.

17.

Wang

Liang

, et al. Design on electrohydraulic servo driving system with walking assisting control for lower limb exoskeleton robot. International Journal of Advanced Robotic Systems. 2021; 18(1): 1-17.

18.

Zhang

Liu

, et al. Biomechanical design of escalading lower limb exoskeleton with novel linkage joints. Technology and Health Care. 2017; 25(S1): 267-273.

19.

Zhang

Tran

Huang

. Design and experimental verification of hip exoskeleton with balance capacities for walking assistance. IEEE/ASME Transactions on Mechatronics. 2018; 23(1): 274-285.

20.

Hyun

Lim

Park

, et al. Singular wire-driven series elastic actuation with force control for a waist assistive exoskeleton, H-WEXv2. IEEE/ASME Transactions on Mechatronics. 2020; 25(2): 1026-1035.

21.

Meijneke

van Oort

Sluiter

, et al. Symbitron exoskeleton: Design, control, and evaluation of a modular exoskeleton for incomplete and complete spinal cord injured individuals. IEEE Transactions on Neural Systems and Rehabilitation Engineering. 2021; 29: 330-339.

22.

Chen

Zhong

Zhao

, et al. A wearable exoskeleton suit for motion assistance to paralysed patients. Journal of Orthopaedic Translation. 2017; 11: 7-18.

23.

Chen

Qin

, et al. Recent developments and challenges of lower extremity exoskeletons. Journal of Orthopaedic Translation. 2016; 5: 26-37.

24.

Hao

Wright

WMD

. Design and modelling of an anti-buckling compliant universal joint with a compact configuration. Mechanism and Machine Theory. 2021; 156: 104162.

25.

Wang

Lee

. A passive gait-based weight-support lower extremity exoskeleton with compliant joints. IEEE Transactions on Robotics. 2016; 32(4): 933-942.

26.

Bartenbach

Gort

Riener

. Concept and design of a modular lower limb exoskeleton. 2016 6th IEEE International Conference on Biomedical Robotics and Biomechatronics (BioRob). UTown, Singapore, IEEE, 2016, pp. 649-654.

27.

Cao

Zhao

, et al. Target of physiological gait: Realization of speed adaptive control for a prosthetic knee during swing flexion. Technology and Health Care. 2018; 26(1): 133-144.

28.

, et al. GC-IGTG: A rehabilitation gait trajectory generation algorithm for lower extremity exoskeleton. 2019 IEEE International Conference on Robotics and Biomimetics (ROBIO). IEEE, 2019, pp. 2031-2036.

Design and analysis of a lightweight lower extremity exoskeleton with novel compliant ankle joints

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSIONS:

Keywords

1. Introduction

2. Exoskeleton design methods

2.1 Design requirements

Table 1 Comparison of technical data between the existing lower extremity exoskeletons and expected design of MOLLEE

2.5 Compliant ankle

3.1 Forward kinematical modeling

Footnotes

Acknowledgments

Conflict of interest

References

Table 1
Comparison of technical data between the existing lower extremity exoskeletons and expected design of MOLLEE