Posts Tagged Rehabilitation robotics

[Abstract] Adaptive integral terminal sliding mode control for upper-limb rehabilitation exoskeleton


    Adaptive integral sliding mode control design for exoskeletons.

    Finite time convergence of the closed-loop system.

    Robustness of the control law with respect to parametric variations and disturbances.

    No requirement of the knowledge of the system bounds.

    Real experiments using an upper limb exoskeleton with and without human subjects.


A robust adaptive integral terminal sliding mode control strategy is proposed in this paper to deal with unknown but bounded dynamic uncertainties of a nonlinear system. This method is applied for the control of upper limb exoskeleton in order to achieve passive rehabilitation movements. Indeed, exoskeletons are in direct interaction with the human limb and even if it is possible to identify the nominal dynamics of the exoskeleton, the subject’s limb dynamics remain typically unknown and defer from a person to another. The proposed approach uses only the exoskeleton nominal model while the system upper bounds are adjusted adaptively. No prior knowledge of the exact dynamic model and upper bounds of uncertainties is required. Finite time stability and convergence are proven using Lyapunov theory. Experiments were performed with healthy subjects to evaluate the performance and the efficiency of the proposed controller in tracking trajectories that correspond to passive arm movements.


via Adaptive integral terminal sliding mode control for upper-limb rehabilitation exoskeleton – ScienceDirect


, , , , , , , ,

Leave a comment

[BOOK] Rehabilitation Robotics: Technology and Applications – Rehabilitation Robotics

Cover image The last decades have seen major advances in interventions for neuromotor rehabilitation. Forms of treatment based on repetitive exercise of coordinated motor activities have been proved effective in improving gait and arm functions and ultimately the patients’ quality of life. Exercise-based treatments constitute a significant burden for therapists and are heavy consumers of health-care resources. Technologies such as robotics and virtual reality can make them more affordable.

Rehabilitation robotics specifically focuses on systems—devices, exercise scenarios, and control strategies—aimed at facilitating the recovery of impaired sensory, motor, and cognitive skills. The field has a relatively long history, dating back to the early 1990s. Early attempts were part of the general trend toward automating heavy tasks by using “intelligent” machines, with minimal human intervention. The notion of “artificial therapist” was common in early scientific papers and patent applications. However, the most distinctive feature of these devices is not their ability to “automate” treatment but, rather, that of precisely quantifying sensorimotor performance during exercise, in terms of movement kinematics and exchanged forces. This resulted in a gradual shift toward more evidence-based and data-driven forms of treatment. Present-generation rehabilitation robots are designed as complements, rather than substitutes, of the therapist’s work. They support the recovery of functions by efficiently exploiting structure and adaptive properties of the human sensorimotor systems and provide rich information on sensorimotor performance and their evolution. Their design, implementation, and modalities of intervention incorporate findings from behavioral studies on sensorimotor adaptation and motor skill learning and their neural substrates.

Rehabilitation robotics is therefore characterized by highly specific design approaches and technical solutions, with roots in both engineering and neurophysiology.

This book addresses both technology and application aspects of Rehabilitation Robotics. Part I focuses on the state of the art and representative advancements in the design, control, analysis, and implementation of rehabilitation robots and the underlying neurophysiological principles. Part II addresses the existing applications and the clinical validation of these systems, with a special emphasis on therapy robots, which support exercise-based treatments aimed at recovering sensorimotor or cognitive functions.

PART I: Background and Technology


Planning and execution of movements results from the coordinated activity of multiple interconnected sensory and motor areas in the cerebral cortex. When an area in this specialized motor network is damaged—for example, through a traumatic brain injury or an ischemic event—the activity of the motor networks can be disrupted, thus leading to functional deficits. How the surviving motor networks reorganize to compensate for the injury depends on the location and extent of the lesion but may be affected by sensorimotor exercise.

Chapter 1 summarizes how neuroplasticity modifies motor networks in response to injury, by focusing on the changes after a cerebrovascular accident in the primary motor cortex. Neuroanatomical and neurophysiological evidence in animal models and human stroke survivors is reviewed to demonstrate how injuries functionally impair motor networks, how motor networks compensate for the lesion to improve motor function, and how selected therapies may facilitate recovery.

Chapter 2 focuses on the hierarchical architecture and synergistic functioning of the motor system. These aspects are crucial for the development of successful robotics applications with rehabilitation purposes. The same framework is used to discuss the mechanisms underlying rehabilitation interventions with a potential to facilitate the recovery process.

Technology and Design Concepts

Devices for rehabilitation benefit from advances in robot technologies, including sensors and actuators, mechanical architectures, and the corresponding control architectures. These devices are characterized by a continuous interaction with the human body, which poses specific design constraints.

Chapter 3 summarizes the notion of “biomechatronic” design for systems for robot-mediated rehabilitation, encompassing robot structure, musculoskeletal biomechanics, and neural control. Robots for rehabilitation are typically conceived to constantly work in constrained motion with the human body, which represents a challenge for designers. This requires a top-down design approach, in which a model of the human agent guides a concurrent, iterative design cycle of the robot’s mechanical, electronic, and multilayered control subsystems. Criteria for the identification of functional and technical specifications and the selection of key components of the robotic system are also derived. Two design case studies demonstrate how these design principles are translated into practice.

Chapter 4 addresses how actuators play a critical role in defining the characteristics of the robot-patient interaction. The different options for actuating and controlling a rehabilitation device are discussed, considering the complex flow of information between the user and the robot during a rehabilitation task. Strategies for both high- and low-level control are presented. Impedance and admittance control modalities are discussed as means of decoding human intention and/or modulating the assistive forces delivered by the robot. Mathematical tools for model-based compensation of nonlinear phenomena (backlash and friction) are also presented.

The way robots are used to facilitate training is crucial for their application to therapy and has important implications for their mechanical and control design. Intensity and frequency of practice are major determinants of the recovery process, but different exercise modalities are possible. Robots may be used for haptic rendering in virtual environments, to provide forces that facilitate task performance or task completion, and/or to make a task more difficult and challenging.

Chapter 5 reviews the control strategies for robotic therapy devices and summarizes the techniques for implementing assistive strategies, including counterbalance techniques and adaptive controllers that modify control parameters based on the patient’s ongoing performance.

Personalized treatment is becoming increasingly popular in neurorehabilitation. Two chapters discuss how new design techniques such as exoskeletons or wearable robots are applied to the design of modern therapy robots, for either upper or lower limb rehabilitation.

Chapter 6 specifically addresses the design of exoskeletons for upper-limb rehabilitation. After an introduction of the rationale behind the selection of this robot architecture and a review of the available solutions for actuation, the chapter discusses the state of the art and the most commonly adopted solutions. An overview of clinical evidences of upper-limb rehabilitation with exoskeletons is then provided, discussing evidences in favor of training with exoskeleton devices.

Chapter 7 reviews the current state and clinical effectiveness, safety, and usability of exoskeletons for gait rehabilitation. It provides an overview of the actuation technologies, including compliant and lightweight solutions. Control strategies aimed at guiding the patient according to his/her needs and encouraging his/her active participation are also discussed. Novel perspectives for “symbiotic” human-exoskeleton interaction based on interfaces with neural structures are also introduced.

Computational Neurorehabilitation

One important feature of therapy robots is that they integrate both therapeutic and measuring functionalities. Therapy robots have built-in technology and sensors that measure movement kinematics and kinetics, thus providing an accurate assessment of motor function by which it is possible to diagnose the patient state and to evaluate patient performance and their progress during treatment. The availability of quantitative information has triggered an entirely new paradigm for neurorehabilitation, unifying clinical assessment, and exercise. Computational neurorehabilitation is a new and emerging field, which uses modern data analysis and modeling techniques to understand the mechanisms of neural plasticity and motor learning, and incorporates this knowledge into personalized, data- and model-driven forms of treatment.

Chapter 8 reviews the quantitative measures—encompassing kinematic, kinetic, timing, sensory, and neuromechanical aspects of performance—which are most frequently used to describe motor behavior during robot-assisted rehabilitation of the upper limb. The chapter also analyzes how these indicators are used to monitor motor recovery during exercise, to understand the evolution of performance, and to precisely plan and, if necessary, modify the rehabilitation strategies. The relationship between robot-derived measures and their clinical counterparts is also discussed.

Chapter 9 addresses computational models for neuromotor recovery, with a focus on state-space models that describe the development of functional behaviors through exercise and the relation between neuromotor recovery and motor learning. The chapter first reviews models of the dynamics of sensorimotor adaptation and motor skill learning and then elaborates on similarities and differences with neuromotor recovery. Finally, it discusses how these models can be used to achieve a better understanding of the role of robots to promote recovery and to develop personalized forms of treatment.

Chapter 10 proposes a general framework to model the interaction between robot and patient during robot-assisted training. Human and robot are modeled as two agents, whose respective tasks are described by two cost functions. Optimal interaction strategies are then derived in terms of differential game theory. This approach allows to describe different forms of human-robot interaction. A specific prediction is that optimal interaction requires that the robot maintains a model of the behavior of its human partner. In this case, simulations and empirical studies exhibit more stable, reactive, and adaptive interaction. This form of “symbiotic” interaction is a step toward defining what it takes for robots to behave as “optimal” trainers.

Chapter 11 addresses the strategies implemented in rehabilitation robots to promote patient motivation, which is a major determinant of recovery through exercise. Motivation may be measured with self-report questionnaires or with indirect, more objective measures, such as exercise duration. Motivation may be promoted through interaction with virtual environments, which may consist of activities of daily living, which emphasize relatability, or games, which emphasize enjoyment. The design of these environments must take the hardware, the patients’ characteristics, and goal-related feedback into account. Motivation during exercise must be maintained by regulating task difficulty, thus ensuring an appropriate “challenge level.”

Software Environments for Rehabilitation Robotics

As a natural conclusion of this methodological section,  Chapter 12 reviews the software development environments that can be used to implement the different levels of control of a modern rehabilitation robot. The robotic field suffers from a lack of standardization in programming environments. Hence, it is not surprising that even in the specific context of rehabilitation robotics, there is currently no consensus on specific software and hardware platforms. The chapter surveys different solutions used for combining robots (and, more in general, haptic interfaces) and virtual environments. Advantages and disadvantages of each of these environments are discussed, together with typical applications, with a focus on upper-limb rehabilitation.

PART II: Applications

The second part of the book addresses the application of rehabilitation robots in different pathologies for training of diverse districts (upper and lower limb) and using different training strategies.

High Intensity, Assist-As-Needed Therapy to Improve Motor Functions

Chapter 13 provides an overview of 28 + years of efforts at MIT’s Newman Laboratory for Biomechanics and Human Rehabilitation for the developments of robotic tools to assist in the neurorecovery process. After a definition of the basic principles that are core for successful rehabilitation robotics technology, the chapter presents a snapshot of few of MIT’s rehabilitation robots, discusses the results of metaanalyses for upper extremity robotics, and finally presents two exciting examples for acute and chronic stroke. Overall, the above material points out that robotic therapy for the upper extremity that involves an interactive high-intensity, intention-driven therapy based on motor learning principles and assist-as-needed leads to better outcomes than usual care in both acute/subacute and chronic stroke.

The above principles have been extended to training in a three-dimensional workspace, using robots with an exoskeleton structure.  Chapter 14 describes the application of one of the first architectures developed with the purpose of mirroring the anatomical structure of the human arm and of enabling task-oriented training in the 3D space, mimicking activities of daily living.

Hand and finger functions are of critical importance for independence in everyday activities, but their recovery is often limited following neurological injury. This has motivated the development of novel therapeutic and assistive tools.  Chapter 15 provides a comprehensive overview of robotic approaches for the rehabilitation of hand function and underlines their potential to complement conventional rehabilitation. First, the design concepts of existing hand exoskeletons and end-effector devices are presented. Then, clinical evidence that underlines the feasibility of robot-assisted rehabilitation of hand function is presented. Finally, promising research directions are discussed to further exploit the potential of robot-assisted rehabilitation of hand function in neurological patients.

Robot-assisted gait training typically involves body-weight support and physical guidance to move the legs into the correct pattern. Gait rehabilitation robots allow greater exercise duration and movement repetitions; improve patient safety and motivation; reduce the therapists’ burden; and, eventually, improve the therapeutic outcome.  Chapter 16 introduces the rationale for robot-assisted gait training. In particular, existing gait-rehabilitation robots and their control strategies are presented. The available clinical trials are also summarized, showing that training with robotic rehabilitation devices is at least as effective as conventional physiotherapy. Further clinical studies are required in order to define the most appropriate robotic technical features based on the task, patients’ type, and degree of impairment.

Wearable systems open new perspectives for rehabilitation in individuals with disabilities, which can lead to difficulty in walking or making arm movements since they could be used to facilitate independent training in the clinic or at home. Wearable systems range from complex rigid exoskeleton structures for the assistance of joints or limbs to hybrid, soft, and interactive systems. The existing solutions are not yet widely used in clinical environments. The aim of  Chapter 17 is to review the scientific challenges and the current developments of wearable systems and to discuss their clinical potential.

Robots Not Only for Stroke Rehabilitation

Although most applications of robot rehabilitation focus on stroke and traumatic brain injury, these devices may find application in the treatment of other pathologies.

Chapter 18 addresses robot-assisted rehabilitation in multiple sclerosis (MS). Robot-assisted training leads to improved movement quality on reaching tasks, but clinical effects on standard assessment have not been always observed after multiple-session training. An increasing number of studies report effects of a multiple-week training program, but the magnitude of the effect was often similar to conventional training programs. Overall, there is evidence supporting the beneficial effect of robot- and technology-supported training, but its superiority compared with other or conventional treatment programs is still debatable. Research investigating the impact of different technological settings and the motor learning strategies implemented in technology must be encouraged for MS patients.

Persons with cognitive deficits are a completely different target population that can be addressed by therapy robots. Cognitive rehabilitation therapy (CRT) is a set of interventions designed to enhance cognitive performance. Ideally, CRT engages the participant in a learning activity to enhance neurocognitive skills relevant to the overall recovery goals. There is ongoing research to identify the determinants of a positive response to treatment.  Chapter 19 addresses the use of rehabilitation robots, socially interactive robots (SIR), and socially assistive robots (SAR), both virtual and embodied, to enhance, restore, or prevent early deterioration of cognitive abilities related to neurodegenerative disease or injury.

Integrating Robot Therapy With Neuro- and Psychophysiological Techniques

All the techniques and devices described until now use robot technology alone. Integration of different approaches and different technologies may improve the outcome, for instance, by training and restoring different functions within the same training session or by using physiological signals to monitor and/or control the recovery process. The following chapters focus on the use of neuro- and psychophysiological signals to enhance or complement robot-assisted therapy.

Chapter 20 presents hybrid FES-robot devices for training of activities of daily living aiming at the parallel restoration of functions by the external activation of paralyzed muscles and external mechanical support of postural functions. The combination of two modalities within the same treatment may multiply their individual effects as the external activation of muscles eliminates the need for large mechanical actuators and reduces the number of degrees of freedom to a controllable domain and, on the other hand, robot guidance removes the need for prolonged, fatiguing stimulation of muscles.

In spite of the acknowledged importance of proprioception for motor control and neuromotor rehabilitation, no effective method for assessment and rehabilitation of proprioceptive deficits has emerged in clinical practice. While there are many clinical scales for assessing proprioception, they all have insufficient psychometric properties and cannot be used in closed-loop treatment paradigms wherein treatment parameters are monitored and adjusted online or with a trial-by-trial frequency.  Chapter 21 discusses how robots can simultaneously address two interrelated needs: to provide sensitive and repeatable assessments of proprioceptive integrity and to automate repetitive training procedures designed to enhance proprioception and its contributions to functional movement.

The outcome of a training program can be conditioned not only by the patient’s physical conditions but also by his/her psychophysiological state during the whole course of the rehabilitation program.  Chapter 22 reviews psychophysiological response modalities that, together with task performance parameters and biomechanical measurements, may be used in a biocooperative approach to rehabilitation. The chapter focuses in particular on electrocardiogram, skin conductance, respiration signal, and peripheral skin temperature. Each signal is described in terms of acquisition modalities, signal processing, and features extraction. The psychophysiological responses in the case of multimodal challenge and physical activity are also examined, with reference to the differentiation of arousal and valence.

Understanding the mechanisms underlying muscle coordination during daily motor activities is a fascinating challenge in neuroscience and may provide important information pertaining to the recovery strategies of the neuromuscular system. Muscle synergies have been hypothesized as a neural strategy to simplify the control of the redundant motor actuators leading human movement and as a method to study motor coordination in healthy and neurological subjects.  Chapter 23 presents the theoretical framework for the extraction and the description of muscle synergies. Moreover, it summarizes how neuropathologies impact on muscle synergies and their potential for neurorehabilitation. Finally, it discusses how muscle synergies can be used to assess the effectiveness of robot-aided rehabilitation and the design of innovative control strategies.

Robots and Information Technologies Advances Toward Long-Term Intervention

As the world’s population ages, the management of chronic diseases will become more important. This shift will put pressure on health-care systems that often focus on providing effective care while reducing costs. The use of technological advancements to augment health-care services provides a method to meet these demands. Telerehabilitation robotics, addressed in  Chapter 24, combines established features of robot-assisted rehabilitation and tele-health care to provide distance rehabilitation services. While there is a growing market of robotic devices used in traditional rehabilitation settings, home-based implementations provide a unique set of challenges (e.g., remote monitoring, deployment constraints, and data management) that has limited the number of successful solutions. Clinical and kinematic outcomes show promising results and support further investigation. Cost analyses have demonstrated that telerehabilitation robotics is a cost-effective alternative compared with clinic-based therapy. While telerehabilitation robotics is a promising addition to conventional care, numerous barriers that limit practical integration will need to be addressed to allow a more widespread acceptance and use of this approach in rehabilitation.

As a final remark, robot rehabilitation involving an interactive high-intensity, intention-driven therapy based on motor learning principles and assist-as-needed leads to better outcomes than usual care in acute/subacute, chronic stroke and other pathologies. For this reason, clinical guidelines recommend the application of these technologies for the recovery of the lost functions.

This book highlights the most important technical aspects and strategies for the design, development, and application of robot technologies for rehabilitation purposes. With their ability to adapt exercise parameters based on physiological signals, objective and sensitive metrics reflecting the state and performance of a patient, the unique possibility to combine motor and somatosensory training, and the perspective of simple and wearable tools for home rehabilitation, robot devices promise further potential for the rehabilitation of neurological patients aiming at an improved motor function, a reduction of their disability, and overall an improved quality of life.


via Rehabilitation Robotics: Technology and Applications – Rehabilitation Robotics


Leave a comment

[ARTICLE] Hand exoskeleton for rehabilitation therapies with integrated optical force sensor – Full Text

This article presents the design of a hand exoskeleton that features its modularity and the possibility of integrating a force sensor in its frame. The modularity is achieved by dividing the exoskeleton in separate units, each one driving a finger or pair of them. These units or “finger modules” have a single degree of freedom and may be easily attached or removed from the robot frame and human fingers by snap-in fixations. As for the force sensing capability, the device relies on a novel force sensor that uses optical elements to amplify and measure small elastic deformations in the robot structure. This sensor can be fully integrated as a structural element of the finger module. The proposed technology has been validated in two experimental sessions. A first study was performed in a clinical environment in order to check whether the hand exoskeleton (without the integrated force sensor) can successfully move an impaired hand in a “Mirror Therapy” environment. A second study was carried with healthy subjects to check the technical feasibility of using the integrated force sensor as a human–machine interface.

A wide diversity of robotic devices, which can actuate/assist the movements of the human hand, can be found in the current scientific literature.1 Depending on the application, a hand exoskeleton may require uneven features. For example, a rehabilitation-aimed exoskeleton needs to be fairly backdrivable and allows a wide range of movement, so it is flexible enough to perform different rehabilitation exercises.2 In contrast, an assistance exoskeleton must be stiff enough to ensure a firm grasping of objects present during activities of daily living and can sacrifice flexibility of movement in favor of predefined grasping patterns.

These different requirements result on diverse force transmission architectures:

  • Some devices use linkages in order to transmit the force from the actuator to the human joints.35 This is a stiff architecture that requires a proper alignment between kinematic centers of the linkage and human joints, but allows a good control of the hand pose. Due to the flexibility of the design, with the correct sizing, these mechanisms can achieve complex movement patterns with simple actuators.
  • Another extended architecture is the cable-driven glove.68 These are more flexible and simpler alternatives that rely on the own human joints to direct the movement, so they are less prone to uncomfortable poses. In contrast, they require pulleys to achieve high forces and are harder to control in intermediate positions. Additionally, this kind of exoskeletons need a pair of cables in antagonist configuration in order to assist both extension and flexion movements.
  • Finally, some devices use deformable actuators, like pneumatic muscles or shape-memory alloys, attached directly to the hand by means of a glove.9,10 They result in very light and simple devices, but actuators are not placed in the most advantageous place to achieve great forces.

Regarding the exoskeletons based on linkages, especially those which rely on electric actuators, having a measurement of the interaction force between user and device may result an interesting feature in order to ease control tasks and improve safety. In certain devices, different sensor technologies have been implemented, such as torque sensors,11 strain gauges,12 flexion sensors,13 and miniature load cells.14 These sensors may be effective in their respective applications but present some shortcomings for their integration in exoskeletons. In particular, torque sensors measure loads in the motor shaft so, in over-constrained mechanisms, they might not measure all the interaction forces. Strain gauges are complex to fix in the proper place and shorter ones may not perform correctly, so for being usable they require geometries with size comparable to human phalanxes. Another miniature sensors, like load cells or force-sensitive resistors, normally can measure force in only one sense (compression or extension) and those that can measure both directions are too big for the scale of the human hand.

Research background and objectives

In our previous paper,15 we studied the feasibility of using multimodal systems in order to assist post-stroke patients during the execution of rehabilitation therapies with real objects. In this context, we evaluated the suitability of using a hand exoskeleton device,16 such as the aforementioned ones, for assisting an impaired person during the grasping of objects present in activities of daily living. This device has experienced substantial improvements with respect to the previous design in order to be able to interact safely with disabled users.

In that previous experimentation, the electromyographic (EMG) signal of the forearm muscles was proposed as a method to estimate user’s intention and consequently trigger the open/close movement of the hand exoskeleton. This method proved to be effective, but it can be used only for users with a coherent and relatively strong EMG signal, which might not be the case for most patients.17 From these results, there is a need for additional technologies that can detect the movement intention of the subject in order to cope with a wider range of user profiles.

Despite that the presented device will also be used in assistive context, the objective of the exposed research is to show whether the proposed improvements of the hand exoskeleton, including a miniature optical force sensor, allow its use in a real rehabilitation environment. Special attention will be given to the development of a force sensing method in order to measure the human–robot interaction forces and therefore to estimate user’s intention in rehabilitation scenarios.

Hand exoskeleton

Among the different existing architectures, we have decided to implement an exoskeleton based on the linkage approximation, since we consider that this is the most flexible solution in order to achieve a good compromise between the requirements of both rehabilitation and assistance scenarios. The motion transmission is based on a bar mechanism that allows the possibility of coupling the motion of phalanxes, so a natural hand movement is achievable using only one active degree of freedom per finger. Additionally, bars can transmit both tensile and compressive loads so the same mechanism is able to perform extension (most demanding movement in rehabilitation) and flexion (mandatory for assistance) movement of the fingers.

In detail, the designed exoskeleton is composed by three identical finger modules that drive index, middle and the pair formed by ring and little fingers. Each finger module has a single degree of freedom actively driven by a linear actuator. Unlike many of the referenced exoskeletons, due to the inherent uncertainty introduced by the human–exoskeleton interface (modeled as a slide along the phalanx longitudinal axis in Figure 1), we have decided not to rely on the human finger as the element that closes the kinematic chain. Conversely, we have adopted an approach similar to the one adopted by Ho et al.5 This way, adding a pair of circular guides whose centers are coincident with the joints of a reference finger, the mechanism is kinematically determinate without needing the human finger. Ho’s device uses slots with flange bearings to implement the guides; this may result effective but requires precision machining and miniature elements to achieve a compact solution. In contrast, we have designed a double-edged guide that slides between four V-shaped bearings (Figure 2). These elements allow the optimization of the required space and may be easily manufactured by prototyping technologies or plastic molding. To make up for the additional constraints, we have decided to actuate only medial and proximal phalanxes.



Figure 1. Kinematics scheme of the finger linkage attached to the human finger. Metacarpophalangeal (MCP), proximal interphalangeal (PIP), and distal interphalangeal (DIP) joints have been modeled as revolute joints. Additionally, the interface between the module and the phalanxes has been modeled by means of slide.



Figure 2. Left: Finger module represented in its extreme positions. Right: Detailed view of the designed circular guide to minimize mechanical clearances with minimum friction.


Continue —>  Hand exoskeleton for rehabilitation therapies with integrated optical force sensor – Jorge A Díez, Andrea Blanco, José María Catalán, Francisco J Badesa, Luis Daniel Lledó, Nicolas García-Aracil, 2018

, , , , , , , ,

Leave a comment

[Abstract+References] The Face Tracking System for Rehabilitation Robotics Applications – Conference paper


The paper presents the working model of the face tracking system. The proposed solution may be used as one of the parts of the rehabilitation or assistive robotic system and serve as the robotic vision subsystem or as the module controlling robotic arm. It is a low-cost design, it is based on open source hardware and software components. As a hardware base the Raspberry Pi computer was used. The machine vision software is based on Python programming language and OpenCV computer vision library.



  1. 1.
    H. Zhou, H. Hu: Human motion tracking for rehabilitation – A survey, Biomedical Signal Processing and Control, Volume 3, Issue 1, pp. 1âĂŞ18, January 2008.Google Scholar
  2. 2.
    C. Rougier, J. Meunier, A. St-Arnaud, J. Rousseau: Monocular 3D Head Tracking to Detect Falls of Elderly People. Proceedings of the 28th IEEE EMBS Annual International Conference New York City, USA, Aug 30-Sept 3, 2006.Google Scholar
  3. 3.
    L. Boccanfuso, J. M. O’Kane: Adaptive Robot Design with Hand and Face Tracking for Use in Autism Therapy. Social Robotics, Second International Conference on Social Robotics, ICSR 2010, Singapore, November 23-24, 2010, pp. 265–274, DOI: 10.1007/978-3-642-17248-9_28.
  4. 4.
    T. Fong, I. Nourbakhsh, K. Dautenhahn: A survey of socially interactive robots. Robotics and Autonomous Systems, Volume 42, Issues 3âĂŞ4, pp. 143âĂŞ166, March 2003.Google Scholar
  5. 5.
    P. Jia, H. H. Hu, T. Lu, K. Yuan: Head gesture recognition for hands free control of an intelligent wheelchair. Industrial Robot: An International Journal, Vol. 34 Iss: 1, pp.60 âĂŞ 68, 2007.Google Scholar
  6. 6.
    P. Raif, J.A. Starzyk: Motivated learning in autonomous systems. The 2011 International Joint Conference on Neural Networks (IJCNN), pp. 603–610, 2011.Google Scholar
  7. 7.
    J.A. Starzyk, J.T. Graham, P. Raif, A.H. Tan: Motivated Learning for Autonomous Robots Development. Cognitive Science Research, 14, 1, 2011.Google Scholar
  8. 8.
  9. 9.
  10. 10.
  11. 11.
    J.E. Solem: Programming Computer Vision with Python: Tools and algorithms for analyzing images. O’Reilly Media 2012.Google Scholar
  12. 12.
  13. 13.
  14. 14.
    K. Demaagd, A. Oliver, N. Oostendorp, K. Scott: Practical Computer Vision with SimpleCV: The Simple Way to Make Technology See. O’Reilly Media 2012.Google Scholar
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
    J. Howse: OpenCV Computer Vision with Python. CreateSpace Independent Publishing Platform 2015.Google Scholar
  20. 20.
    G. Bradski, A. Kaehler: Learning OpenCV: Computer Vision with the OpenCV Library. O’Reilly Media 2008.Google Scholar
  21. 21.
    P. Viola, M. Jones: Rapid object detection using a boosted cascade of simple features. In: Proc. IEEE Computer Society Conference on Computer Vision and Pattern Recognition. vol. 1, pp. 511–518, 2001.Google Scholar
  22. 22.
    P. Viola, M. Jones: Robust Real-time Object Detection. International Journal of Computer Vision 57(2), pp. 137–154, 2004.Google Scholar

via The Face Tracking System for Rehabilitation Robotics Applications | SpringerLink

, , , , , , ,

Leave a comment

[Editorial] Rehabilitation and assistive robotics – Advances in Mechanical Engineering

It is estimated that in the European Union (EU) the proportion of the population aged over 65 years will rise from 17.1% in 2008 to 30% in 2060 and that the proportion of persons aged over 80 years will rise from 4.4% to 12.1% over the same period (EUROSTAT population projections). Neurological conditions, especially stroke, are a major cause of disability among older people. Incidence of a first stroke in Europe is about 1.1 million and prevalence about 6 million. Currently, about 75% of stroke sufferers survive 1 year after. This proportion will increase in the coming years due to steadily increasing quality in hyper-acute lifesaving practice, follow-up acute and sub-acute care, and lifelong management of these conditions. Despite these positive developments in stroke care, approximately 80% of stroke patients experience long-term reduced manual dexterity, a 72% of those affected by stroke suffer leg weakness, affecting walking, and half of all patients with neurological conditions are unable to perform everyday tasks. Rehabilitation and assistive robotics have the potential to change older people lives improving their recovery and/or supporting them to perform everyday tasks.

The purpose of this special collection is to provide an opportunity for researchers working in academy or industry to show their latest theoretical, technological, and experimental aspects of rehabilitation and assistive robotics. A total of eight articles have been accepted after a strict peer review process.

In the topic of rehabilitation robotics, Fraile et al. present an end-effector rehabilitation robot, a 2-degree-of-freedom planar robotic platform for upper limb rehabilitation in post-stroke patients. In addition, they describe the ergonomic mechanical design, the system control architecture, and the rehabilitation therapies that can be performed by the aforementioned rehabilitation robot. There are other two more papers included in this topic. In the first one, Diez et al. propose a novel multimodal robotic system for upper-limb neurorehabilitation therapies in physical environments, interacting with real objects. This system consists of an end-effector upper-limb rehabilitation robot, a hand exoskeleton, a gaze tracking system, an object tracking system, and electromyography measuring units. Their experimental results show that the proposed system is feasible and safe enough. Wrong detections in electromyography (EMG) are the main cause of failure; however, in the 97% of the trials, it still resulted in successful grasping and releasing. In the second one, Simonetti et al. present the design and development of a modular architecture for delivering upper limb robotic telerehabilitation with the CBM-Motus, a planar unilateral robotic machine. Their architecture allows a therapist to set a therapy session on his or her side and send it to the patient’s side with a standardized communication protocol; the user interacts with the robot that provides an adaptive assistance during the rehabilitation tasks. Moreover, the experimental results with seven healthy subjects show the reliability of the novel architecture and the capability to be easily tailored to the user’s needs with the chosen robotic device.

In the topic of robotic prosthetics, Barone et al. propose a multilevel control of an anthropomorphic robotic hand with prosthetic features. The novel approach is based on two distinct levels consisting of (1) a policy search learning algorithm combined with central pattern generators in the higher level and (2) a parallel force/position control managing slippage events in the lower level. Their experimental results demonstrate that the proposed control has the potential to adapt to changes in the environment and guarantees grasp stability, by avoiding object fall thanks to prompt slippage event detection. Moreover, Sekine et al. present the development of a shoulder prosthesis based on a hybrid actuation system composed of pneumatic elastic actuators (PEAs) and servo motors. Their results show that the joints with PEAs could absorb more impact force, which is very important for safe use, than with motors.

In this special collection, there are two papers in the field of wearable exoskeletons. In the first one, Ning et al. present the design and development of a power-assisted gait orthosis. The paper analysed the gait characteristics with crutches, designed the mechanical architecture, and optimized it using genetic algorithms. Moreover, the performance of the final design is verified under many external conditions, such as no-load, gait movement, long-term continuous movement, and load tests. In the second one, Zhang et al. propose a human–machine force interaction designing architecture for a load-carrying exoskeleton. Their experimental results show that the human–machine interaction force detection at the back and feet and the identification of different body modalities and movement intention are feasible. Moreover, the actual load on the human back is far less than the payload, which shows that their exoskeleton has good power-assisted effect.

The last paper included in this special collection is about a novel algorithm to estimate the instantaneous tremor parameters such as the time-varying dominant frequency in the case of nonsynchronous sampling and to distinguish the tremulous movement from the raw data. The experimental results reported by Wang et al. demonstrate that the proposed solution could detect the unknown dominant frequency and distinguish the tremor components with higher accuracy than the existing procedures.

Source: Rehabilitation and assistive roboticsAdvances in Mechanical Engineering – Nicolas Garcia-Aracil, Alicia Casals, Elena Garcia, 2017

, , , ,

Leave a comment

[ARTICLE] Modifying upper-limb inter-joint coordination in healthy subjects by training with a robotic exoskeleton – Full Text



The possibility to modify the usually pathological patterns of coordination of the upper-limb in stroke survivors remains a central issue and an open question for neurorehabilitation. Despite robot-led physical training could potentially improve the motor recovery of hemiparetic patients, most of the state-of-the-art studies addressing motor control learning, with artificial virtual force fields, only focused on the end-effector kinematic adaptation, by using planar devices. Clearly, an interesting aspect of studying 3D movements with a robotic exoskeleton, is the possibility to investigate the way the human central nervous system deals with the natural upper-limb redundancy for common activities like pointing or tracking tasks.


We asked twenty healthy participants to perform 3D pointing or tracking tasks under the effect of inter-joint velocity dependant perturbing force fields, applied directly at the joint level by a 4-DOF robotic arm exoskeleton. These fields perturbed the human natural inter-joint coordination but did not constrain directly the end-effector movements and thus subjects capability to perform the tasks. As a consequence, while the participants focused on the achievement of the task, we unexplicitly modified their natural upper-limb coordination strategy. We studied the force fields direct effect on pointing movements towards 8 targets placed in the 3D peripersonal space, and we also considered potential generalizations on 4 distinct other targets. Post-effects were studied after the removal of the force fields (wash-out and follow up). These effects were quantified by a kinematic analysis of the pointing movements at both end-point and joint levels, and by a measure of the final postures. At the same time, we analysed the natural inter-joint coordination through PCA.


During the exposition to the perturbative fields, we observed modifications of the subjects movement kinematics at every level (joints, end-effector, and inter-joint coordination). Adaptation was evidenced by a partial decrease of the movement deviations due to the fields, during the repetitions, but it occurred only on 21% of the motions. Nonetheless post-effects were observed in 86% of cases during the wash-out and follow up periods (right after the removal of the perturbation by the fields and after 30 minutes of being detached from the exoskeleton). Important inter-individual differences were observed but with small variability within subjects. In particular, a group of subjects showed an over-shoot with respect to the original unexposed trajectories (in 30% of cases), but the most frequent consequence (in 55% of cases) was the partial persistence of the modified upper-limb coordination, adopted at the time of the perturbation. Temporal and spatial generalizations were also evidenced by the deviation of the movement trajectories, both at the end-effector and at the intermediate joints and the modification of the final pointing postures towards targets which were never exposed to any field.


Such results are the first quantified characterization of the effects of modification of the upper-limb coordination in healthy subjects, by imposing modification through viscous force fields distributed at the joint level, and could pave the way towards opportunities to rehabilitate pathological arm synergies with robots.[…]

Continue —> Modifying upper-limb inter-joint coordination in healthy subjects by training with a robotic exoskeleton | Journal of NeuroEngineering and Rehabilitation | Full Text

Fig. 1 Example of goal-directed pointing task (GDM). The four pictures show the motion from the starting position to the WAM button, while performing GDM task. In this case the subjects were not asked to follow any specific endpoint trajectory

, , , , , , , ,

Leave a comment

[ARTICLE] User-centered design of a patient’s work station for haptic robot-based telerehabilitation after stroke – Full Text


Robotic therapy devices have been an important part of clinical neurological rehabilitation for several years. Until now such devices are only available for patients receiving therapy inside rehabilitation hospitals. Since patients should continue rehabilitation training after hospital discharge at home, intelligent robotic rehab devices could help to achieve this goal. This paper presents therapeutic requirements and early phases of the user-centered design process of the patient’s work station as part of a novel robot-based system for motor telerehabilitation.

1 Introduction

Stroke is one of the dominant causes of acquired disability [1] and it is the second leading cause of death worldwide [2]. The high incidence of the disease and the current demographic developments are likely to increase the number of stroke patients in the future. Most of the survivors have physical, cognitive and functional limitations and require intensive rehabilitation in order to resume independent everyday life [3]. Therefore, the main goal of motor rehabilitation is relearning of voluntary movement capability, a process which takes at least several months, some improvement can occur even after years. In the rehabilitation clinic, patients usually receive a daily intensive therapy program. However, for further improvement of motor abilities, severely affected patients are required to continue their rehabilitation training outside the rehabilitation settings, after being discharged from the rehabilitation clinic. Langhammer and Stanghelle [4] found that a lack of follow-up rehabilitation treatment at home leads to deterioration of activities of daily living (ADL) and to motor functions in general. A possible solution is an individualized and motivating telerehabilitation system in the patient’s domestic environment. Some studies [5], [6] have confirmed the advantage of home rehabilitation after stroke and showed that telerehabilitation received high acceptance and satisfaction, both from patients, as well as from health professionals [7]. Most of the existing telesystems [7], [8] are based on audio-visual conferencing or on virtual environments and contain rather simple software for monitoring patients’ condition. However, in neurological rehabilitation the sensorimotor loop needs to be activated by provision of physiological haptic feedback (touch and proprioception) [3].

Robot-based rehabilitation is currently one of the most prevalent therapeutic approaches. It is often applied in hospitals alongside conventional therapy and is beneficial for motor recovery [9]. Rehabilitation training including a haptic-therapy device may therefore be even more promising for home environments than non-haptic telerehabilitation. Several telerehabilitation systems, which include not only audio and visual, but also haptic modality, already exist [10], [11] . Most of these solutions use low-cost commercial haptic devices (e.g. joysticks) for therapy training, with the goal of cost minimization and providing procurable technology. Nonetheless, devices specifically developed for stroke rehabilitation, which are already established in clinical settings, may have greater impact on motor relearning and could therefore also be more effective at home, compared with existing home rehabilitation devices.

In a previous paper [12], we presented a concept and design overview of a haptic robot-based telerehabilitation system for upper extremities which is currently under development. In the present work, we describe therapeutic requirements, user-centred development [13] and implementation of the patient’s station of the telesystem.

Continue —> User-centered design of a patient’s work station for haptic robot-based telerehabilitation after stroke : Current Directions in Biomedical Engineering

Figure 3 Implementation of the patient’s work station based on Reha-Slide (left) and Bi-Manu-Track (right).

, , , , , , , ,

Leave a comment

[ARTICLE] Compensating the effects of FES-induced muscle fatigue by rehabilitation robotics during arm weight support – Full Text


Motor functions can be hindered in consequence to a stroke or a spinal cord injury. This often results in partial paralyses of the upper limb. The effectiveness of rehabilitation therapy can be improved by the use of rehabilitation robotics and Functional Electrical Stimulation (FES). We consider a hybrid arm weight support combining both.

In order to compensate the effect of FES-induced muscle fatigue, we introduce a method to substitute the decreasing level of FES support by cable-driven robotics. We evaluated the approach in a trial with one healthy subject performing repetitive arm lifting. The controller automatically adapted the support and thus no increase in user generated volitional effort was observed when FES induced muscle fatigue occured.

Continue —> Compensating the effects of FES-induced muscle fatigue by rehabilitation robotics during arm weight support : Current Directions in Biomedical Engineering

, , , , , , , , ,

1 Comment

[Abstract] A Fully Fabric-Based Bidirectional Soft Robotic Glove for Assistance and Rehabilitation of Hand Impaired Patients


This paper presents a fully fabric-based bidirectional soft robotic glove designed to assist hand impaired patients in rehabilitation exercises and performing activities of daily living. The glove provides both active finger flexion and extension for hand assistance and rehabilitative training, through its embedded fabric-based actuators that are fabricated by heat press and ultrasonic welding of flexible thermoplastic polyurethane-coated fabrics. Compared to previous developed elastomeric-based actuators, the actuators are able to achieve smaller bend radius and generate sufficient force and torque to assist in both finger flexion and extension at lower air pressure. In this work, experiments were conducted to characterize the performances of the glove in terms of its kinematic and grip strength assistances on five healthy participants. Additionally, we present a graphical user interface that allows user to choose the desired rehabilitation exercises and control modes, which include button-controlled assistive mode, cyclic movement training, intention-driven task-specific training, and bilateral rehabilitation training.

Source: A Fully Fabric-Based Bidirectional Soft Robotic Glove for Assistance and Rehabilitation of Hand Impaired Patients – IEEE Xplore Document

, , , ,

Leave a comment

[ARTICLE] The Efficacy of State of the Art Overground Gait Rehabilitation Robotics: A Bird’s Eye View – Full Text


To date, rehabilitation robotics has come a long way effectively aiding the rehabilitation process of the patients suffering from paraplegia or hemiplegia due to spinal cord injury (SCI) or stroke respectively, through partial or even full functional recovery of the affected limb. The increased therapeutic outcome primarily results from a combination of increased patient independence and as well as reduced physical burden on the therapist. Especially for the case of gait rehabilitation following SCI or stroke, the rehab robots have the potential to significantly increase the independence of the patient during the rehabilitation process without the patient’s safety being compromised. An intensive gait-oriented rehabilitation therapy is often effective irrespective of the type of rehabilitation paradigm. However, eventually overground gait training, in comparison with body-weight supported treadmill training (BWSTT), has the potential of higher therapeutic outcome due its associated biomechanics being very close to that of the natural gait. Recognizing the apparent superiority of the overground gait training paradigms, a through literature survey on all the major overground robotic gait rehabilitation approaches was carried out and is presented in this paper. The survey includes an in-depth comparative study amongst these robotic approaches in terms of gait rehabilitation efficacy.

Download full text in PDF

Source: The Efficacy of State of the Art Overground Gait Rehabilitation Robotics: A Bird’s Eye View

, , , , , , , , , ,

Leave a comment

%d bloggers like this: