Experimental characterization of the T-FLEX ankle exoskeleton for gait assistance

Highlights

• The static experimental characterization of a variable stiffness exoskeleton is detailed.

• The element’s influence and the configuration for the best actuator’s performance were measured and determined.

• The results evidence the device’s applicability in assistive scenarios during human walking.

• An adaptative stage to counteract the characteristical system times in real scenarios is required.

• Dynamic test and trials with system configuration’s variation are needed in future studies.

Abstract

Designing robotic devices to assist or emulate the ankle is challenging due to the joint’s complexity and the fundamental role in walking. T-FLEX is an ankle exoskeleton based on vsa for rehabilitation and assistance of people with ankle dysfunctions. This device has presented promising motor recovery results for a stroke patient during a rehabilitation program. However, human walking applications require an electromechanical characterization to measure the device’s capabilities and determine the suitable configuration that responds to this complex task. This work presents T-FLEX’s experimental characterization carried out in a test bench structure. The results showed alterations in system times and actuators’ bandwidth because of the tendons’ force levels. Furthermore, this study determined the most appropriate T-FLEX configuration to obtain the best performance. Thus, this work also presents a preliminary validation under that configuration on a healthy subject in gait assistance to assess the device’s response in different velocities and measure the effects on the user. In conclusion, T-FLEX can assist the human gait for gait cycle duration greater than 0.74 s providing torque on the ankle of up to 12 Nm in propulsion and 20 Nm in dorsiflexion. Nevertheless, it should include an adaptable stage in the control architecture to counteract the stabilization time for providing the maximum torque at the right time.

Introduction

The ankle has a fundamental role during the different phases of the human gait and hence in the execution of other Activities of Daily Living (adl) [1]. These joint functions include specific tasks such as weight distribution, shock absorption, foot-clearance, and controlling the contact of the foot with the ground [1]. In this sense, those tasks require precise motor control and an appropriate amount of energy for the subject [2].

Neurological conditions (i.e., cerebral palsy, stroke, spinal cord injury, and traumatic brain injury) can compromise the performance and execution of those ankle functions [3], [4]. Furthermore, the after-effects produce altered motor control, which affects adl’s accomplishment reducing people’s quality of life [5]. Likewise, the ankle dysfunctions trigger an altered gait pattern, resulting in overloading the other lower limb joints [1]. Consequently, the joints generate a high metabolic cost related to compensatory movements trying to supply those lost functions [6], causing dangerous and permanent impacts on the locomotor system.

To overcome the limitations and recover partially motor skills, the patients perform rehabilitation processes [7]. Traditional physical therapy comprehends exercises such as repetitive movements and task-oriented training, aiming to provide the patient the maximum independence as possible for executing ADL [8]. Nevertheless, although the results for conventional therapy are positive [9], robotic technology’s inclusion improves the patient’s recovery capacity [10], [11].

The ankle dysfunctions can remain in the patients after therapy [12]. Thus, to provide stability and prevent injuries due to the lack of ankle control, passive orthotic devices are prescribed [13]. Those devices are a mechanical structure that only allows the dorsi- plantarflexion movements and restricts the other planes of motion. However, although the use of a passive structure helps the patient execute ADL, the user does not improve the gait pattern. Conversely, the locomotion is affected by both the dysfunction and the ankle restriction [6].

Therefore, robotic devices are being developed to support the rehabilitation processes and assist user gait [5], [14]. Specifically, the powered ankle–foot orthoses (pafos) integrate the principles applied in passive orthotic structures, although they incorporate the advantages of robotics (i.e., energy supply using actuators, user monitoring through sensors, programmed functionality profiles, among others) [14]. Those devices have been divided by the state-of-art according to their purpose and the system actuation that was implemented [15], [16]. In terms of the actuation system, the most common actuators used in pafos are (1) stiff, (2) pneumatic, and (3) series elastic [15], [17]. However, there are other actuation principles based on the previous actuators, such as the variable stiffness actuators (vsa) and cable-driven actuators (cda) [17].

vsa uses the same concept as series elastic (i.e., an elastic element with a spring between the actuator and the load), although this principle includes a variable stiffness spring instead of a constant value [17], [18], [19]. The core feature of actuators based on vsa is changing the system output stiffness in different interaction cases with the environment (e.g., constant load and constant position) [18]. Moreover, this principle also presents advantages in shock loads and backdrivability, in the same way as the series elastic actuators [17], [19]. Considering the mentioned advantages, this actuation type is widely recommended for robotic applications where the robot interacts intensively with humans [20], [21]. Specifically, a study evidenced an influence on the reduction of the metabolic cost related to the stiffness level of the actuator [21].

Against this background, T-FLEX is a wearable and portable ankle exoskeleton based on vsa (see Fig. 1) used to assist the dorsi-plantarflexion movements without restricting the other ankle motions [22]. This pafo comprises of two servomotors attached to bioinspired tendons whose mechanical behavior is similar to the human Achilles tendon [23]. Furthermore, T-FLEX includes a bidirectional system of stiff filaments to help assistive movements and correct the foot pathological postures. This device is manually adjustable and usable for both limbs. Moreover, it includes two operational modes: (1) therapy that consists of repetitions following velocity and frequency defined by the user, and (2) gait assistance, which integrates an inertial sensor and a machine-learning algorithm [24] to detect the gait phase and assist the movement.

The T-FLEX exoskeleton (see Fig. 1) comprehends a four-bar mechanism by each actuator, where one of them is a spring with variable stiffness. Despite the advantages in aspects related to the human–machine interaction, the inclusion of elastic materials limits the actuator’s features in terms of bandwidth, supplied torque, response time, among others [25]. The knowledge of those features and the device’s capabilities allows improving aspects such as control strategies, human–robot interaction, and robot performance [17], [26], [27], [28]. For this purpose, static or dynamic benches have been deployed to assess the actuator’s variables (i.e., force, torque, angular or linear velocity) through different input signals such as step signal or following a goal value, chirp signal, between others [17], [26].

In terms of application, a study showed the potential to use the T-FLEX exoskeleton in stationary therapy of stroke patients [29]. Specifically, the study showed that the use of the device in 18 sessions promotes motor recovery (i.e., improvement in range of motion and spatiotemporal parameters), reduces the spasticity level, and improves ankle control in gait (i.e., reduction in toe drag) [29]. However, considering the other application of T-FLEX for assistive scenarios to improve the gait pattern and reduce the risk of falls, an electromechanical characterization is necessary to measure the device’s capabilities and determine the suitable configuration that responses to this complex task.

Several ankle exoskeletons have applied mechanisms similar to the T-FLEX’s actuation principle. Specifically, a cable-driven parallel robot involves a platform composed of linear actuators attached to flexible cables to assist 3 degrees of freedom dof [30]. Thus, this device can support ankle movements and avoid inertial impactions that are common in rehabilitation scenarios. Likewise, another ankle exoskeleton based on a cable-driven architecture is the CABLEankle [31]. This device includes a lightweight foot platform actuated by servomotors to assist the ankle joint in stationary scenarios, where the system’s response varies concerning the tension in the actuated cables. Notwithstanding, although these robots include novel mechanisms to assist the ankle and exhibit potential in the stationary rehabilitation field, their designs preclude a possible application in gait assistance. Moreover, this is one of the principal activities addressed in a rehabilitation program to guarantee independence and improve the quality of life [32].

On the other hand, studies focused on experimental characterizations for ankle exoskeletons are limited [17], [26], [33], [34], [35]. In general terms, those experiments have been aimed at measuring the system response through the high-level controllers immersed in the device’s application. However, some studies determined this response through simulated environments that could differ from a real scenario. In other cases, the experiments intended to determine the device capabilities for ideal conditions. Therefore, few studies have been focused on the actuator performance, modifying both (1) the initial setup and (2) the scenario to estimate several device’s responses [26], [35].

This work presents the experimental characterization of the T-FLEX ankle exoskeleton (see Fig. 1). The main goals are to determine the appropriate device’s configuration for assisting the human gait with the best performance and measure the device’s capabilities. Therefore, this study assessed the different systems incorporated in T-FLEX (i.e., the bidirectional system and the variable stiffness system) for three initial force levels. Moreover, a case study was carried out on a healthy user to assess the T-FLEX’s performance in a real application and the device’s effects on him.