Wearable Robotics: Challenges and Trends

Research conducted over the last decades indicates a necessity of having larger number of EMG sensors in order to extract sufficient information needed for natural control of upper limb prosthetics. Various studies have addressed this issue, though clinical transition and evaluation of such systems on a larger pool of patients is still missing. We propose a specifically designed system which allows users to perform clinically relevant tests in an unobstructed way while handling dexterous prosthesis. Eight electrodes were embedded into customized sockets along with the controllers driving an algorithm recently tested in laboratory conditions that allows simultaneous manipulation of four out of seven prosthetic functions. The fully self-contained system was evaluated on seven amputees conducting the Southampton Hand Assessment Procedure. The scores achieved were compared to those obtained using their own commercial devices. The study shows the necessary steps to validate novel control algorithms in a clinically meaningful context.

The application of robotic exoskeletons in gait therapy during stroke rehabilitation has grown rapidly over the past decade. The purpose of this investigation is to determine the impact of a robotic exoskeleton (RE) on temporal spatial gait parameters as compared to traditional standard of care gait training in a single inpatient after acute stroke. Outcome measures included temporal spatial gait parameters while walking with and without an RE during a single gait training session two weeks post stroke. During gait training in the RE, walking speed, and stride length increased while step width decreased on the affected and unaffected side. Total double support time decreased, and single support and swing time increased on the affected and unaffected limb during gait training in the RE. Gait training in the RE had a positive effect on the patients overall gait which included improved temporal spatial parameters and gait speed. Positive changes in temporal spatial parameters were evident on the affected and unaffected limb. Preliminary rehabilitative improvements with the RE device may include a gait training environment that is more symmetrical and may improve weight transfer and overall gait speed. Further research with a larger sample with various level of impairment utilizing an RE for gait training is needed to determine the potential utility of REs as an alternative to traditional gait training.

Impaired ability to walk independently is a significant consequence of multiple sclerosis (MS) resulting in substantial limitation in mobility and performance of daily activities, thus restricting full participation and home and community re-integration. Gait training/restoration in MS is necessary but often limited due to the progress and severity of the disease and limitations of traditional strengthening exercises Much attention has been emphasized in developing pharmaceutical treatment to stop or delay the progress of the disease, but not in developing rehabilitation strategies to improve quality of life and facilitate community re-integration for this population. Recently, wearable lower extremity robotic exoskeletons have been developed to restore ambulation in paralyzed or weak individuals with spinal cord injuries. Utilizing exoskeleton in MS rehabilitation may provident difference users experience and improve rehabilitation outcomes. The purpose of this paper is to provide preliminary results and early experience in our MS exoskeleton assisted rehabilitation program.

Exoskeletons are devices which have recently gained substantial attention in the field of rehabilitation, especially for use in individuals with neurological disorders (ND). In individuals with ND, mobility limitations and subsequent reduced activity levels contribute to significant reductions in quality of life. Wearable robots or exoskeletons hold much promise to fulfill this unmet need of over-ground mobility and unimpaired gait. However, variable research evidence and clinical efficacy are deterring the impact of these eloquent devices from being introduced into everyday rehabilitation practice. This research summary will highlight some research studies conducted at the Rehabilitation Institute of Chicago (RIC) with various exoskeletons, focusing on their clinical efficacy and personal mobility outcomes.

Lower limb exoskeletons (EXOs) may be devised as an ambulation device, as a rehabilitation tool or may be aimed at allowing both objectives. Regarding rehabilitation, it is worth noticing that EXOs also provide a perfect environment for precise assessing of rehabilitation protocols and effects. Different is the case of EXO for mobility, in this area the old wheelchair is still largely winning the challenge. In all functional areas: velocity, safety, portability, acceptance as well as autonomy in the ADL none of today EXOs can compete with the performances of an average wheelchair. EXO usage requires learning, and brain changes associated with tool usage introduce the human in the loop concept, a key aspect of clinical relevance for EXO usage. At present, interesting data on the biological mechanisms and rehabilitation relevance of embodiment are providing hints for guiding rehabilitation. These challenges will be presented from a clinical rehabilitation perspective and expectations and critics discussed.

More and more robots are designed to help or substitute humans both in daily activities and dangerous scenarios. These robots should be able to cope with humans and with other robots and to move in houses, factories, hospitals and uncertain outdoor terrains.

In this extended abstract we present a novel architecture of an electric-mechanical actuator suitable for application to systems with periodic motion and here presented as an alternative to traditional actuators employed in wearable robots for lower limbs: the Flywheel - Infinitely Variable Transmission (F-IVT).

This paper presents the design of a compliant, lightweight and adaptable active ankle foot orthosis (AAFO) and preliminary test of its ankle actuator. The ankle actuator is designed to keep its weight as low as possible. The adaptability of the AAFO allows adjusting the device to different patients, without the need of customized versions.

The design of an innovative spherical mechanism with three degrees of freedom for a shoulder joint exoskeleton is presented in this paper. The spherical mechanism is designed with a double parallelogram linkage, which connects two revolute joints to implement the motion as a spherical joint, while maintaining the remote center of rotation. The design has several new features compared to the current state-of-the-art: (1) a relative large range of motion free of singularity, (2) high overall stiffness, (3) lightweight and (4) compact, which make it suitable for assistive exoskeletons.

In this paper, we present the combination of the Soft-SixthFinger, a wearable robotic extra-finger designed to be used by chronic stroke patients to compensate for the missing hand function, with a robotic arm that is used as an assistive device to support the patient arm. The extra-finger is a tendon-driven modular structure worn at the paretic forearm. The robotic extra-finger is used jointly with the paretic hand/arm to grasp an object similarly to the two parts of a robotic gripper. The flexion/extension of the robotic finger is controlled by the patient using an Electromyography (EMG) interface embedded in a cap. The robotic arm is controlled to partially compensate for the weight of the paretic arm, while not interfering with the user arm motion. The system has been designed as a tool that can be used by chronic stroke patients to compensate for grasping in many Activities of Daily Living (ADL). We performed a pilot test to demonstrate that the proposed system can significantly improve the performance and the autonomy in ADL.

A pneumatic exoskeleton intended to decrease the muscle activity of the knee extensors during walking on a negative slope is presented. The device consists of an air spring that can be engaged and disengaged via a solenoid valve. When engaged, the air spring resists knee flexion. A preliminary evaluation of the device was conducted with a single healthy subject. During testing, the EMG activity of the rectus femoris decreased by 15 %, while the EMG activity of the vastus medialis increased by 8 %.

Orientation assessment based on wearable sensors is becoming crucial for providing feedback information in wearable robotics and sport monitoring. Gravitational acceleration and Earth’s magnetic field are commonly used as a reference vectors for orientation estimation. This paper presents a novel sensory fusion algorithm for assessing the orientations of human body segments in long-term human walking, and enhance performance in environment with magnetic disturbance. The proposed system was experimentally validated. The results show accurate joint angle measurements (error median below $${5}{^\circ }$$) with no expressed drift over time. The incorporated compensation of magnetic disturbances proved assessment with absolute median errors bellow $${2.5}{^\circ }$$.

Aging of population and increased incidence of gait impairments are dominant trends undermining social welfare and healthcare system. Lower-limb wearable robots proved to be a viable solution for recovering mobility of people affected by gait disorders. This work presents the design of the mechatronic architecture of a fully self-contained active pelvis orthosis (APO) for assisting hip flexion/extension movements during daily living activities. The APO could act compliantly with the human biomechanics thanks to series-elastic actuation and to a novel kinematics chain endowed with passive degrees of freedom. The portability and autonomy of the control system have opened the horizon to explore different assistive tasks in out-of-lab scenarios.

XoSoft is an EU project that proposes the development of a modular soft lower-limb exoskeleton to assist people with mobility impairments. It aims to be user friendly and comfortable to wear, with a significant impact on the person’s mobility and health, on their independence and quality of life. Being a modular system, it comprises of ankle, knee and hip elements, which can be used individually or combined and used unilaterally or bilaterally.XoSoft follows a user centered design strategy achieved by involving primary, secondary and tertiary end users as participatory stakeholders in the design and development process. Preliminary findings of the interviews with the different users groups are presented in this paper.Advanced textiles and smart materials are being developed to create sensing, variable stiffness joints and flexible tactile sensors. Control will be through biomimetics to identify the user’s motion and intention and to determine and provide the appropriate level of assistance. Connected health connectivity and analysis will enable the wearer and their clinicians/therapist to review activity information. The concept will be tested extensively in the lab, and subject to trials in clinical settings and home environments.

Treatment intensity has a profound effect on motor recovery following neurological injury. The use of robotics has potential to automate these labor-intensive therapy procedures that are typically performed by physical therapists. Further, the use of wearable robotics offers an aspect of portability that may allow for rehabilitation outside the clinic. The authors have developed a soft, portable, lightweight upper extremity wearable robotic device to provide motor rehabilitation of patients with affected upper limbs due to traumatic brain injury (TBI). A key feature of the device demonstrated in this paper is the isolation of shoulder and elbow movements necessary for effective rehabilitation interventions. Herein is presented a feasibility study with one subject and demonstration of the device’s ability to provide safe, comfortable, and controlled upper extremity movements. Moreover, it is shown that by decoupling shoulder and elbow motions, desired isolated joint actuation can be achieved.

This paper presents the design improvement of a polymer-based tendon-driven wearable robotic hand, Exo-Glove Poly. The wearability and adaptiveness are the key points to design the Exo-Glove Poly in considering the cases of practical use. Thus, magnets are embedded into the wearable part for easy donning and doffing. Also, the tendon length adjustment mechanism is designed to adapt different hand sizes by changing length of the tendons. Through these improvements, it is increased the change to practical use of the Exo-Glove Poly.

The goal in wearable rehabilitation is to restore the lost functionality of the body by rebuilding the sensory-motor link. This may be achieved through a replication, in an artificial or robotic system, of the physiotherapy methods employed by human experts. These methods are typically focused on physical manipulation. We suggest that a lower reliance on manipulation, combined with affective touch stimulation, has the potential to provide effective rehabilitation in lower power and lighter wearable devices. Here we consider affective touch driven by soft actuation and how this may be combined with low power artificial muscle actuators for physical rehabilitation.

This abstract describes the design and experimental evaluation of a force tracking controller for hip extension assistance utilizing a soft exosuit connected to a tethered off-board actuation system. The new controller aims to improve the force profile tracking capability and demonstrate its advantages over our previously reported work. The controller was evaluated by one healthy participant walking on a treadmill at 1.35 m/s. Results showed that the system can deliver a predefined force profile robustly with a 200 N peak force. The measured peak force value using force controller was 198.7 ± 2.9 N, and the root-mean-squared (RMS) error was 3.4 N (1.7 % of desired peak force). These results indicate that the force control reduces peak force variability and improves force profile tracking capability.

We present an online decoding method for controlling a powered lower-limb exoskeleton using endogenously generated electroencephalogram (EEG) signals of human users. By performing a series of binary classifications, users control the exoskeleton in three directions: walk front, turn left and turn right. During the first classification phase, the user’s intention to either walk front or change direction is detected. If the user’s intention to change direction is detected, a subsequent classification for turning left or right is performed. Five subjects were able to successfully complete the 3-way navigation task while mounted in the exoskeleton. We report the improved accuracy of our cascaded protocol over a baseline method.

Natural-quality, independent ambulation is a prerequisite for community use of lower extremity exoskeletons by individuals with disabilities. In general, current exoskeletons generate pre-programmed gait, where the user cannot exercise volitional control necessary to navigate over uneven surfaces and avoid obstacles. This project introduces an intuitive control strategy that allows the user to determine and sense the exoskeleton movement in real time using trajectories produced by the hands. The concept allows neurally defined ambulation control to be expressed through alternative biological articulators. This novel approach uses admittance control to compute each exoskeleton’s foot position from Cartesian forces exerted by the user’s hand on a trekking pole that is connected to foot through a multi-axis load cell. The algorithm has been evaluated by naïve, non-disabled users who walked a 10 degree of freedom, ½ scale biped robot on a treadmill. The results show that the algorithm produced robot-generated gait kinematics that are similar to human gait kinematics. A human-scale exoskeleton has been developed to allow further exploration of this control method.

Real-time electromyography (EMG) driven musculoskeletal (NMS) modeling estimates internal body biomechanical parameters and motor intentions. This is central for understanding the dynamics of user-exoskeleton interaction and for developing closed-loop user-exoskeleton interfaces that are intuitive and effective in promoting neuroplasticity. This abstract, presents methods and results behind the interfacing between a six degree of freedom lower limb exoskeleton (H2 exoskeleton, Technaid S.L., Spain) and a real-time EMG-driven NMS model of the human lower extremity.

We take a multi-faceted whole-system approach towards designing and implementing a neural interface for lower limb exoskeletons. This paper highlights some of the initial steps we have taken, from the development of shared control techniques to the characterization of the exoskeleton itself and the identification of brain signals that could be used in such an interface.

To realize a lower extremity exoskeleton that can provide balance assistance in a natural way, an understanding of human balance control is a necessity. In this study, we investigated how the angle, torque and power of the ankle, knee and hip joints changed in response to balance perturbations during walking. Nine healthy young adults walked on an instrumented treadmill and received pelvis perturbations of various magnitudes and directions at the instance of toe-off right. An open source musculoskeletal modeling package (OpenSim) was used to perform inverse kinematics and inverse dynamics. Subjects modulated the ankle torque in the (left) stance foot with the magnitude and direction of the perturbation. Also in gait phases following foot placement, subjects addressed ankle torques to mitigate the remaining effects of the perturbation. The results presented here support the use of ankle actuation in lower extremity orthoses for natural and cooperative balance control.

This study showcases effect of adding assistive torques to the hip, knee, and ankle joints in the sagittal plane on the total human body metabolic energy expenditure using the AnyBody musculoskeletal modeling system. To this goal, a box-lifting task was targeted and metabolic energy was computed for several cases including when each of the three joints was assisted at a time. Simulation results showed that the hip joint assistance affects the total metabolic energy consumption more than the knee and ankle joints.

Variability is an inherent feature of human movement, but little research has been done in order to measure such a characteristic using inertial sensors attached to person’s body (wearable sensors). Therefore the aim of this preliminary study is to investigate the assessment of human movement variability for dance activities. We asked thirteen participants to repeatedly dance two salsa steps (simple and complex) for 20 s. We then used a technique from nonlinear dynamics (time-delay embedding) to obtain the reconstructed state space for visual assessment of the variability of dancers. Such reconstructed state space is graphically linked with their level of skillfulness of the participants.

This abstract presents the real time computation of Centroidal Momentum (CM) in human walking while addressing its applicability as a stability index to human walking with exoskeleton. To this end, a real time full body motion capture suit solution was employed and it was tested both in steady over ground walking and in walking with tripping events. Results show that observation of CM is able to detect abnormal state of human gait in real time thanks to its inherent physical meaning.

We investigated the capabilities of a reflex-based neuromuscular controller with a knee and hip gait trainer worn by a subject with a complete spinal cord injury. With controller assistance, this subject was able to reach a walking speed of 1.0 m/s. Measured joint torques agreed reasonably well with those of healthy subjects. The controller was also robust, recovering from manual swing foot perturbations. These preliminary results are promising for future implementation of neuromuscular controllers on wearable prototypes for real-world walking conditions.

In the Symbitron Project, one of the main objectives is to develop a safe, bio-inspired, and personalized wearable exoskeleton that enables individuals with a spinal cord injury (SCI) to walk without additional assistance, by complementing their remaining motor function. The first target group of five subjects, have enough hip control to keep themselves upright, but need support around the ankle and/or knee joint. This paper gives an overview of the design features of the newly developed exoskeleton and shares some details about the design process.

Current exoskeletons replay pre-programmed trajectories at the actuated joints. Towards the employment of exoskeletons with more flexible and adaptive behavior, we investigate human balance control during gait. We study human balance control by applying brief force pulses at the pelvis in different directions, with different amplitude, and applied at different phases of the gait phase. The observed changes were dependent on the phase at which the perturbation was applied and the walking velocity. From the results we concluded that foot placement was the dominant strategy in the frontal plane, center of pressure (CoP) modulation in the double support phase was utilized in the sagittal plane, and the duration of the swing and double support phase changed. Without the ability to control the CoP through an ankle torque, humans also used a foot placement strategy in the sagittal plane. The center of pressure with respect to the center of mass at the end of the double support phase was linearly related to velocity of the center of mass at the end of the preceding swing phase, which is in agreement with extrapolated center of mass or capture point based stepping strategies previously applied in simple models.

To assist people with disabilities, exoskeletons must be provided with human-machine interfaces (HMI) capable to identify the user’s intentions and enable cooperative interaction. Electromyographic (EMG) signals could be suitable for this purpose, but their usability and effectiveness for shared control schemes in assistive devices is currently unclear. Here we developed advanced machine learning (ML) algorithms for detecting the user’s motion intention and decoding the intended movement direction, and discuss their applicability to the control of an upper-limb exoskeleton used as an assistive device for people with severe arm disabilities.

Wearable Robots, i.e. exoskeletons, are currently entering the market. The specific procedure to allow devices on the market, concerning their product safety depends on the application domain they are sold for (medical—industrial—personal care). Safety Standards, such as produced by ISO and IEC are important tools to demonstrate safety for specific devices, but at this moment there are only very few specific standards available and no specific testing methods in place for Wearable Robots. Standardized testing methods that do not require human subject testing can make the road to the market easier and better affordable. Such test methods need to be based on validated experimental data, and may require the development and definition of specific targeted test benches or equipment.

Worldwide, a significant interest in wearable robots or exoskeletons does exist, also from an industrial background. This paper provides an overview of assistive exoskeletons that have specifically been developed for industrial purposes. It discusses their potential in increasing performance and flexibility on one hand and in reducing the mechanical loads on workers involved in manual work on the other. From this it is concluded that exoskeletons have the potential to increase performance under specific conditions and to reduce physical loads significantly. However, several technical issues hinder mainstay practical use of exoskeletons in industry until now. One main issue concerns the human-machine interaction which stands in the way of acceptance. This issue and its linkage to ethics and standardization will be discussed during the conference.

The paper provides a complete overview of the legal issues involving wearable robots, in particular ranging from their definition and applicable regulation at the European level, civil liability issues of producers and users as well as a brief approach to issues of human enhancement.

Physical stress hazards are identified as a part of the safety requirements for personal care robots as described in ISO 13482. We conducted a safety verification study to clarify an inherently safe condition region in the shear stress—time relationship: The higher the shear stress is, the smaller number of times is needed for blister generation. For validating the verified safety data, we also built a testbed where a manipulator is used to optimally control the position and force of a cuff in contact with a piece of porcine skin for the purpose of reproducing the contact conditions as close as possible to those obtained when a human wears a robot.

Recently, various gait rehabilitation robots have been used as therapy in clinical fields for stroke, spinal cord injuries, and several neurological disorders. We investigated the kinematic differences with joint trajectories of two types of gait rehabilitation robots, i.e., exoskeleton and end-effector devices. Furthermore, we compared the end-effector device’s stair climbing and descending motions to actual motions. The exoskeleton device shows larger hip and knee angle than the end-effector device during gait. However, exoskeleton ankle joint was restricted in dorsiflexed position. The end-effector device’s stair climbing motion was similar to actual stair motion, although there was a delayed and lower maximum flexion. Compared with the actual motion, the stair descending motion had a lower maximum flexion angle for both hip and knee joints in the end-effector device. In addition, the end-effector device’s ankle trajectory was aligned with the dorsiflexion angle, while descending to the bottom stair.

Outcome assessments are vital in facilitating periodic, episodic and ongoing evaluation of persons with limb loss. There are many outcome measures used to quantify prosthetic fit, alignment, comfort, functionality and usability of lower limb prostheses. However, many measures are subjective, difficult to implement in a clinical setting and lack psychometric evaluation. This study used an immersive Computer Assisted Rehabilitation Environment (CAREN) virtual reality system with an instrumented spilt-belt treadmill and real time motion capture system as a research tool to evaluate and compare the gait of lower limb prosthesis users and non-amputees as a preliminary study to determine the effectiveness and appropriate use of outcome measures. The use of the CAREN system providing more real world scenarios such as ramps, inclines and unexpected inclines helped evaluate the hill assessment index (HAI) and the combined gait asymmetry metric (CGAM).

Wearable Robots for Mobility (WR-Mob), i.e. exoskeletons, are currently entering the market. This makes the topic of how to define and measure their performance more relevant and urgent. This abstract provides some considerations that could be taken into account when designing quantitative benchmark metrics that aim to quantify the performance of WR-Mob, focusing on measurement of reduction of metabolic cost and of improvement of balance. The considerations on metrics and their normalization are first steps to well-defined benchmark tests. Proper benchmarks contribute to solid comparison among devices that can be performed in different labs, and thus support a faster progress beyond the state of the art.

Benchmarking bipedal locomotion is an important topic not only for wearable robotics, but also for human movement analysis and humanoid robotics. In this paper we discuss how data from the KoroiBot walking database can be used to establish benchmarking data for human walking in different situations based on direct extraction and on model-based analysis. The goal is to identify human reference values for important quantities defining walking motions which can then serve for evaluating walking motions in the different areas of applications.

The objective of this paper is to identify clinical assessments that are potentially useful for benchmarking of bipedal locomotion. Results: Several clinical measures for static conditions and clinical measures for motion are suggested. Conclusions: Potentially useful clinical measures are identified. New metrics coming from new ambulant measurement techniques may be a good alternative for the qualitative clinical measures.

Brain-Machine Interfaces based on wearable robots’ control have been proposed in the research field for rehabilitation purposes. The combination of both systems allow the performance of more natural movements and a higher level of involvement of patients on their therapy. Studies focused on this topic should face several issues related to the integration of these systems. The current work is meant to test the accuracy of a real time Brain-Machine Interface based on the detection of gait attention during lower limb exoskeletal rehabilitation. Four users performed the experiment wearing an ankle exoskeleton. The system provides a coefficient between 0 and 1 depending on the level of attention experienced by the subject. These results show good similitude between real and decoded attention level.

In this study we explore a way of controlling a lower limb exoskeleton based on the detection of the user intention by recording and classifying information from force sensors placed on both knees and hips. The classifier is based on Linear Discriminant Analysis and has been tested offline in 5 healthy subjects, obtaining an average accuracy of 91.11 % for the sit-to-stand transition, 72.5 % for the stand-to-walk transition and 70 % for the stand-to-sit transition.

In this paper, a conceptual design of the two iterations of compliant actuators used within BioMot project, as well as the control strategy used to operate these actuators, is presented. The result of the presented approach are 2 exoskeleton gait prototypes that will be used for incomplete spinal cord injury (iSCI) patients’ gait rehabilitation.

A main concern that rises when developing active orthoses is how to actively engage the users and monitor how they are affected by the devices. Through EMG-informed neuromusculoskeletal modeling, it is possible to estimate users’ muscle contributions to joint moments generation. We present preliminary results about the application of such models to a subject wearing the BioMot ankle actuator.

Current powered exoskeleton (exo) control algorithms for locomotion assistance and rehabilitation are based on assistive, resistive and error augmentation paradigms. Within the assistive controller’s family, assist-as-needed consists in applying a corrective force proportional to the error (actual limb position versus reference pattern). Our final goal is to implement a fully adaptable control mechanism to allow a full lower limb exo to dynamically adapt the gait pattern to each patient. We propose to use a modified version of tacit learning algorithm in combination with a variable stiffness actuator to explore the improvement of the adaptability in comparison to stiff actuators. The preliminary results show that using this concept on a compliant actuator it is possible to modulate a fixed trajectory to adapt to the position limits that are induced by user’s movement capabilities.

The Hanyang Exoskeleton Assistive Robot (HEXAR)-CR50, which performs human synchronized gait motions to augment human power during load carrying, was designed and analyzed in this study. The HEXAR-CR50 was developed for industrial and military purposes to carry a payload of 20–30 kg while walking on level ground and climbing stairs. For the design of the exoskeleton robot with considered the joint functions and motion, we conducted a gait analysis that was based on biomechanics. The parameters for the design were based on the results of the gait analysis. The designed exoskeleton consisted of one leg with seven degrees of freedom (DOF). The simulations were conducted to verify the kinematic synchronized motion using LifeMOD$$^\mathrm{TM}$$.

Robotic gait training after spinal cord injury (SCI) is of high priority to maximize independence and improve the living conditions of these patients. Current rehabilitation robots are expensive and heavy, and are generally found only in the clinic. To overcome these issues, we present the design of a low-cost, low-weight robotic orthosis for subjects with SCI. The paper also presents a preliminary experimental evaluation of the assistive device on a subject with SCI. Results show that gait velocity, stride length and cadence of walking increased (24.11, 7.41 and 15.56 %, respectively) when wearing active orthoses compared to the case with standard passive orthoses.

A powered low-back exoskeleton is being developed to support manual material handling in industry. Controlling this device poses several challenges. At the low-level, the actuation units need to be capable of large torque outputs as well as transparent interaction. At the high-level, the exoskeleton needs to modulate its assistance based on information acquired from the environment and the wearer, so as to maximise its beneficial effect. These challenges have great relevance to industrial applications, where complexity, cost and invasiveness are key to successful deployment. In describing the current progress in the development of the exoskeleton, an attempt is made to highlight and discuss these challenges and possible technical solutions.

The main advantage of the fully autonomous system is its ability to decouple the weight/mass carrying function of the system from its forward motion function. It makes exoskeleton more efficient by reducing its power consumption, weight and size of the propulsion motors as well as by extending the run time of the batteries. A human machine interface has been achieved by means of flexible sensors to monitor subject shank and ankle movements and subject’s foot pressure. When subject stands on one leg and swings the other one the body weight is fully supported by standing leg exoskeleton structure where the knee joint motion is fully blocked. The join motors are small in size and consume less electrical energy from batteries because they do not support subject’s weight during the walk.

Exoskeleton technology can assist human effort performing manual handling tasks in industrial environments. Safety is vital both for the commercial and legal acceptance of industrial exoskeletons. We consider such systems as safety critical as they directly involve humans. Active safety functions can enhance the safety of exoskeletons and prevent accidents and injuries. We present the conceptual and evaluation criteria for Fall Detection, Active Balancing, Active Detachment, Collision Detection and Automatic Motor Braking. This evaluation highlights the safety critical scenarios which can be helpful in future hazard and risk assessment of exoskeletons, and also for safety certification evaluation.

This paper presents the SOLEUS project, which aims to design innovative countermeasure for astronauts in space, based on the development of an active foot orthosis and immersive virtual reality technologies. The paper introduces the problematics of space countermeasure and describes the expected benefits of the proposed technology. It provides details on the system architecture, components and the simulation tool that has been used to support the design process. Finally, the scientific evaluation strategy for the validation of the system is introduced.

Most assistive robotic devices are exoskeletons which assist or augment the motion of the limbs and neglect the role of the spinal column in transferring load from the upper body and arms to the legs. In the SPEXOR project we will fill this gap and design a novel spinal exoskeleton to prevent low-back pain in able bodied workers and to support workers with low-back pain in vocational rehabilitation.

A hip exoskeleton was designed that can assist hip flexion and extension. The device incorporates a motor, ball-screw, and spring in a lightweight package. The total weight including the battery is 2.95 kg. The system uses 20 W of power per leg. The system is controlled based on the phase angle of each leg and the torque is applied in synchrony with the user’s steps. The device assists walking, running, and does not interfere when going up and down stairs.

Most assistive robotic devices are exoskeletons which assist or augment the motion of the limbs and neglect the role of the spinal column in transferring load from the upper body and arms to the legs. In this part of the SPEXOR project we will fill this gap and design a novel, passive spinal exoskeleton to prevent low-back pain in able bodied workers and to support workers with low-back pain in vocational rehabilitation.

In this abstract, we describe a mono-articular soft exosuit to assist with hip extension during overground walking. The system is comprised of a mobile Bowden-cable-driven electrical actuation unit, soft textiles, and a load cell and an inertia measurement unit per leg. The exosuit applies forces with a peak of 300 N enabled by an IMU-based iterative control algorithm. This iterative controller detects the onset timing of the hip extension assistance based on an estimation of the maximum hip flexion angle. The timing and magnitude of the applied peak force is modulated by generating step-by-step actuator position profiles based on the previously measured assistive force. Results from a human subject during overground walking at self-selected speed indicate the robustness of the system to apply effectively forces with a high consistency in terms of magnitude and timing of the peak force profile.

Mobility can be limited due to age or impairments. Wearable robotics provide the chance to increase mobility and thus independence. A powered soft exosuit was designed that assist with both ankle plantarflexion and hip flexion through a multi-articular suit architecture. So far, the best method to reduce metabolic cost of human walking with external forces is unknown. Two basic control strategies are compared in this study: an ankle moment inspired controller (AMIC) and an ankle positive power inspired controller (APIC). Both controllers provided a similar amount of average positive exosuit power and reduced the net metabolic cost of walking by 15 %. These results suggest that average positive power could be more important than assistive moment during single stance for reducing metabolic cost. Further analysis must show if one of the approaches has advantages for wearers comfort, changes in walking kinetics and kinematics, balance related biomechanics, or electrical energy consumption.

Walking on uneven terrain with a wearable assistive robot requires the controller to adapt to rapid changes in human’s biomechanics. To do so, the changes due to terrain should be measured using wearable sensors. We investigated human ankle joint mechanics when stepping on different small, unanticipated bumps with either the forefoot or the rearfoot. It was shown that kinematics and kinetics change significantly depending on how humans step on a bump, and that changes in kinematics could be measured by IMUs. This result could be used to inform the design of adaptive controllers for wearable robots that provide optimal assistance to the ankle joint when walking on uneven terrain.

Progressive muscle weakness characteristic of Duchenne muscular dystrophy (DMD) results in loss of upper extremity active range of motion (AROM) despite residual muscle strength that is insufficient to overcome gravity. Admittance control is well suited for use by individuals with DMD as it allows for the utilization of residual muscle strength to intuitively control the motion of a powerful robot without requiring strength to overcome gravity and the friction and inertia of the robot. This study examined the feasibility of using the HapticMASTER, an admittance control motorized arm support, to increase the upper extremity AROM of individuals with DMD to a greater degree than that provided by the Armon Edero, a commercially available passive arm support. The results demonstrate that the HapticMASTER robot significantly increased the reachable surface area scores compared to the Armon Edero passive arm support (paired-samples t-test, t(5) = 3.984, p = 0.010, Cohen’s d = 1.6).

Spasticity and dystonia are challenging motor impairments that may interfere with the use of exoskeleton-based therapy. We suggest that two mechanical stimulation techniques that target and remediate these manifestations in cerebral palsy, will allow exoskeletons become a much more feasible rehabilitation technique. This will improve function and, importantly, safety. Our previous studies have shown the positive outcomes of vestibular stimulation, and published literature proposes the possible advantages of whole body vibration. Our current approach utilizes both techniques to define a rehabilitation method specific to the subject’s diagnosis of spasticity and dystonia. Our recent pilot data shows great potential in temporarily eliminating/reducing both spasticity and dystonia in a subject with CP. Results suggest a reduction in tone and possible improvements to mobility after a single session of stimulation. Therefore, this paper serves to propose the use of this approach to enable and enhance the benefits of robotic therapy.

The goal of this research is to assess how powered exoskeleton-training for 5 h per week over 20 weeks can change gait parameters to increase walking speed for chronic SCI. Gait parameters include Center of Mass (CoM) excursions, walking velocity, initial double stance time (IDS), single stance time (SS), terminal double stance time (TDS), swing time (SW), and spatial parameters such as step length, step width and stride length. Exoskeleton training had a significant effect on walking velocity due to specific temporal spatial gait parameters (IDS, TDS, Step and Stride Length) and increased stability (CoM).

An adaptive method for classification of arbitrary activities is presented that assesses continuously the activity in which a subject is engaged, thus providing contextual information facilitating the interpretation of any continuous data gathered from an (unsupervised) applied wearable robotics device and its bearer. Specifically the effect of a novel adaptive and fully automated initialization method using Potts energy functionals is discussed. Exemplary data suggests that this method very likely improves overall performance equally or better than more traditional methods. This includes state of the art methods based on segmental k-means initialization that do require substantial recurrent manual intervention.

The design of an aid for the hand function based on exoskeleton technologies for patients who have lost or injured hand skills, e.g. because of neuromuscular or aging diseases, is one of the most influential challenge in modern robotics to assure them an independent and healthy life. This research activity is focused on the design and development of a low-cost Hand Exoskeleton System (HES) for supporting patients affected by hand opening disabilities during the Activities of Daily Living (ADLs). In addition, the device, able to exert suitable forces on the hand, can be used during the rehabilitative sessions to implement specific tasks useful to restore the dexterity of the user’s hand. The validating and testing phase, conducted in collaboration with the Don Carlo Gnocchi Foundation, showed satisfying results both in terms of portability and wearability which are fundamental requirements for assistance during the Activities of Daily Living (ADLs) and for rehabilitation of people with hand impairments.

Physical therapy is an important resource for the recovery process of several medical conditions. Hand mobility impairment, for example, affects a patients life quality, making it a need to develop aid devices that improve the results of hand therapy, quickening the recovery process. The engineering design process of a wearable and portable rehabilitation glove, was based on the use of muscle wires or Nitinol and specially designed flex sensors. The automated control of this device is performed based on Pulse Width Modulation (PWM), its working cycle, and the feedback provided by the flex sensors, which allow the controlled movement of the different joints in each finger through the use of an interactive graphical user interface, simplifying the phases of measuring the bending angles of each joint before and after each session.

This paper presents the recent activities of Space Applications Services for the development of force feedback arm exoskeletons based on 3D printing technologies. The paper describes the design of the exoskeleton system and illustrates its application through two different projects of slave robotic arm teleoperation, ICARUS and DEXROV, where the concept is used.

SOLEUS system aims at providing efficient countermeasure exercises focused on the lower legs. The final product is foreseen to be based on an orthotics structure allowing exercising the ankle joint and muscle groups, even in microgravity. In order to test the pedal-pulling scenario for the operation of the system, multibody dynamics based musculoskeletal simulation has been performed. The result of the simulation shows the profile of the ankle plantar flexion torque by muscles with the given exoskeleton’s actuator force and motion conditions. Also the muscle activation patterns could be retrieved.

The human gait analysis by using wavelets transform of signal obtained from six inertial ProMove mini sensors is proposed in this work. The angular velocity data measured by the gyro sensors is used to estimate the translational acceleration in the gait analysis. As a result, the flexion–extension, the adduction–abduction joint angles of the hips, flexion–extension of the knees and dorsi and plantar flexion of the ankle are calculated. After measurements we propose to use one of wavelet transform (wavelet type) in order to analyze the signals, indicate a characteristic feature and compare them.

Several examples in literature demonstrate the potential impact of motor inertia on the electrical energy consumption of actuators. Nevertheless, optimizations of actuated prosthetics are often based on the mechanical energy consumption, disregarding the potential effects of motor inertia. In this short abstract, we simulate the electrical energy consumption of a powered prosthetic ankle actuated by a Series Elastic Actuator. Its compliant element is optimized for mechanical energy consumption, a typical strategy in state-of-the-art prosthetics. Our results confirm the importance of motor inertia. Due to the resulting changes in the operating points of the motor, the average motor efficiency is lowered by 17 %.

Stroke is the leading cause of disability in the United States with approximately 800,000 cases per year. This cerebral vascular accident results in neurological impairments that reduce limb function and limit the daily independence of the individual. Robotic rehabilitation may present an exercise intervention that can improve training and induce motor plasticity in individuals with stroke. A motorized hand exoskeleton that operates under admittance control provides support for wrist flexion/extension, abduction/adduction, pronation/supination, and finger pinch has been integrated with a pre-existing 3-Degree of Freedom (DOF) haptic robot (Haptic Master, FCS Moog) to determine the efficacy of increased DOF during proximal and distal training for neurorehabilitation.

In this study, our goal was to improve the standing balance of people with a Spinal Cord Injury (SCI) by using a powered Ankle-Foot orthosis acting in the sagittal plane. We tested four different controllers on two SCI subjects that have a lesion at a low level. In the experiments the subjects repeatedly had to recover from pelvis perturbations, while receiving ankle assistive torques from the orthosis. We found that the controllers that use centroidal dynamics as input parameters were able to provide proper support to the subjects after a perturbation had been applied, even though they worked against the subjects after they had recovered from the perturbation. These preliminary results show the potential of balancing controllers that operate in Center of Mass-space.

In this paper a method to reduce the mechanical impedance of the joints of a lower limb exoskeleton is presented. When user is in charge of the motion the exoskeleton mimic its movements. Gravity, Friction and Interaction Force compensators are designed in order to reduce the force necessary to move the exoskeleton joints. Gravity compensation is used to mitigate the effect of the exoskeleton’s weight. This weight adds a force component when the orientation of the limb is different to the gravity vector. The added Friction compensation effect reduces the frictional phenomena of the joints gearboxes. The Interaction Force is calculated from the measured strain of the segments of the exoskeleton. User intention is also detected using the Interaction Force. The gain block adjusts the weights of the friction and interaction compensators depending on the joint velocity. First in this work the context is shown. Followed by the experimental set-up, several compensators and its effects and the control algorithm.

This research developed a bilateral human-robot mutual adaptive impedance control strategy. The developed interactive impedance coordination methods can let human and robot arbitrarily switch the role between leader and follower seamlessly. Also, through iteratively increasing the impedance in the intended moving direction, human and robot can mutually borrow the force from each other to facilitate the task execution.

Most of today’s assistive devices are controlled to provide uniform assistance irrespectively from the configuration of the human arm and the direction of the movement. We propose an innovative control method for arm exoskeletons that takes into account both of these parameters and compensates the anisotropic property of the force manipulability measure, intrinsic to the biomechanics of the human arm. To test our controller we designed a set of reaching tasks where the subjects had to carry two different loads to targets at five different locations and of two different sizes. Reaching times and trajectories were analysed for the evaluation of the controller. Through the analysis of the average reaching times we found that our method successfully enhances the motion while the analysis of the average maximal deviation from the ideal trajectories showed that our method does not induce any additional dynamic behaviour to the user.

Lower limb wearable robotics also known as exoskeleton or power suit is a booming field of research. Potential medical applications cover a large range of gait disorders from rehabilitation to assistance in daily mobility. Surprisingly, or not, paraplegia seems to be the first target of all commercialized exoskeleton. In this paper we will try to understand this choice and look at other disorders leading to the inability to walk. Neuromuscular, autoimmune or neurological diseases such as muscular dystrophy, multiple sclerosis or stroke, can lead to similar gait disorders and are mostly incurable today. SCI (Spinal Cord Injury) symptoms are quite dissimilar from theirs and reveal specific design challenges. Existing devices’ architecture and human-robot interaction are presented and discussed in terms of adaptation toward non-SCI disorders.

Passive wearable lifting aids support workers by applying gravity force compensation at the arms. In this study we investigated the feasibility of a compensatory lower back moment, generated by a practically constant spring force (38.5 Nm), extending the lower back by pushing on the upper leg. This design is proposed as a light-weight solution to generate lower back moments. The method is compared to using counterweights at a different distances. We recorded EMG activity of the erector spinae longissimus (ES) muscle, the perceived workload (NASA TLX) and the preference of 12 subjects. Results showed no significant difference in ES peak EMG activity during the task, and no significant difference between perceived workload between conditions, as we expected. However, 10 out of 12 subjects indicated preferring the spring mechanism over both counterweights. The main reason of preference was the reduction of weight and inertia of the system. Therefore, the proposed constant spring force mechanism is a feasible alternative to counterweights.