717

Intelligent Haptic Robotic Glove for Patients

Diagnosed with Cerebrovascular Accidents

Nirvana Popescu

Computer Science Department

University Politehnica of Bucharest

Bucharest, Romania,

nirvana.popescu@cs.pub.ro

Dorin Popescu Faculty of Automation, Computers and Electronics

University of Craiova

Craiova, Romania

dorinp@robotics.ucv.ro

Marian Poboroniuc Faculty of Electrical Engineering,

GH. ASACHI Technical University of Iasi,

Iaşi, Romania

mpobor@ee.tuiasi.ro

Cristian Dinu Popescu Rehabilitation Clinic Hospital of Iasi

Iaşi, Romania

cdpopescu@ssnn.ro

Abstract— This paper is based on a research project

regarding the design and development of an Intelligent Haptic

Robot-Glove (IHRG) for the rehabilitation of the patients that

have a diagnosis of a cerebrovascular accident. The IHRG is an

exoskeleton that supports the human hand and hand activities by

using a control architecture for dexterous grasping and

manipulation. This IHRG is a medical device that acts in parallel

to a hand in order to compensate some lost function. A novel

approach for the development of an exoskeleton assistive hand is

considered, a „mechatronic approach” for the integration of the

mechanical structure, the sensory, the actuation system and the

virtual and control system.

Keywords— intelligent robotic glove; exoskeleton; control

architecture; actuation system

I. INTRODUCTION

Rehabilitation Robotics is a branch of the areas of Robotics and Mechatronics that addresses to the study of complex robotic systems aiming to restore the human functions for those people who suffer major trauma as a result of strokes and cerebrovascular diseases (CVA). A CVA occurs when a blood vessel (an artery) that supplies blood to an area of the brain bursts or is clogged by a blood clot. Within minutes, the nerve cells in that area are affected and they can die in a few hours. As a result, the part of the body that is controlled by the affected area of the brain cannot function properly. According to the World Health Organization, each year, 15 million people suffer a cerebrovascular stroke, only 10 million survived and approximately 5 million remain permanently disabled and recuperate after a correct treatment. In accordance with the prestigious Japan Physical Therapy White Paper, over 800,000 people in Japan are using rehabilitation techniques with varying degrees of difficulty.

Total loss or loss of range of motion, decreased reaction times and disordered movement organization create deficits in motor control, which affect the patient’s independent living. Recent studies have shown that intensive and repetitive training may be necessary to modify neural organization and

recover functional motor skills. Several researchers have reported significant improvement in patients’ daily activities due to higher training intensities, even in the chronic phase of the disease. The limited amount of therapy offered by the current system makes it difficult to provide the training intensity needed for neural reorganization and functional changes. Efforts toward developing robotic treatments are motivated by the increasing public health burden associated with stroke-related disability and the current emphasis on cost reduction in health care that has resulted in shorter inpatient rehabilitation length of stay. These factors emphasize the need for optimal interventions for motor rehabilitation after strokes. Integration of robotic therapy into current practice could increase the efficiency and effectiveness of therapists by alleviating the labor-intensive aspects of physical rehabilitation and by enabling novel modes of exercise not currently available.

The potential benefits of robotic training have encouraged several research groups to develop robotic upper extremity rehabilitation systems [1, 2-4]. The team of the Northeastern University has developed various mechatronic devices in an effort to facilitate the rehabilitation of patients recovering from neurological ailments. A non-invasive interface for controlling mechanical hands by using forearm surface electromyography (EMG) was used. By this technique, the muscle activation potentials are gathered by electrodes placed on the patient’s forearm skin [5, 6].

Other authors proposed some exoskeleton powered hand that provides power augmentation for hand and wrist activities of human. They have developed a high-pressure hydraulic artificial muscle that is suitable for use in McKibben artificial muscles and has particularly large force-to-volume and force-to-mass ratios [7, 8]. Takaiwa studied a rehabilitation equipment using a pneumatic manipulator [9]. Some rehabilitation equipments have been developed [10-12] for an upper limb training with a surface force control of mechanical hands. Mori, Suzumori, have developed a high-pressure hydraulic artificial muscle that is suitable for use in McKibben

978-1-4799-2228-4/13/$31.00 ©2013 IEEE

718

artificial muscles for the development of power robot hand with shape adaptability [13, 14].

A key feature of our project presented in this paper is that subjects at any impairment level can repetitively practice and complete stereotyped movement patterns. There is increasing evidence that active repetitive practice of movements can have a profound effect on recovery from brain injury. Active retraining can also positively shape the cortical reorganization associated with motor recovery following brain injury. In chronic stroke subjects, constraint-induced therapy can lead to substantial increases in use of the more affected limb in activities of daily living. This therapy involves intensive repetitive exercise of the more affected limb coupled with constraint of the opposite limb and results in positive cortical reorganization in the motor cortex. Last research results of our consortium members lead to the conclusion that there is an increase in the cortical projection of a paralyzed limb due to a stroke if the rehabilitation exercises are performed by means of Functional Electrical Stimulation (FES).One novelty of our project presented in this paper would be to propose a hybrid system robot glove – FES (Functional Electrical Stimulation) to sustain the upper limb rehabilitation in stroke patients.

The paper is structured as follows: section II analyses the exoskeleton architecture, section III presents the actuation system, section IV analyses the virtual and control architecture, section V describes the hybrid IHRG-FES system for the upper limb rehabilitation, section VI discusses the clinical experiments that will be realized and section VII is concerned with conclusions.

II. THE ARCHITECTURE OF THE EXOSKELETON’S

MECHANICAL SYSTEM

For a correct analysis of hand function, a researcher must thoroughly analyze three main components: the upper hand and wrist, fingers (four) and thumb [15]. Different biomechanical movements of flexion, extension, abduction and adduction were analyzed, considering the amplitudes, movement directions, axes and planes. The model of hand articulations is presented in Fig. 1.

Fig. 1. Hand joints model

Carpal osteoarticular complex is so constituted as to permit carrying out movements of flexion-extension, abduction – adduction and circumduction. Because these joints are plane, each of them allows sliding movements of small amplitude. Wrist joint movements occur simultaneously: both in radiocarpal and mediocarpal joints. All these movements are accomplished by sequential displacement of region segments: the second carpal (distal) moves on the first carpal (proximal)

and then the latter slipped on the forearm. Functional position of the hand is the active position, ready to grab. This means that hand makes a dorsal flexion with 20°, the fingers are slightly flexed and the thumb is in opposition. The muscles are slightly tensed; the extensors of the hand and the flexors of fingers exert a dominant action on antagonists. This is an important aspect since, in the case of stroke occurrence and central level control interruption, this normal anatomic position is accentuated being accompanied by spasticity.

Using haptic interface technology allows for reaching and grasping movement executed by hand to be assisted by a robot that direct the movement to a specific target. Until now, the technologies developed for this purpose have the potential to revolutionize the way hospitals operate, reducing the recovery stroke cost while allowing therapists and clinicians to manage a large number of patients in the same period of time.

Many researchers believe that the best results of motor and functional recovery of patients with stroke can be achieved if patients receive repetitive and motivated robot-mediated therapy encouraging fine object gripping movements with the paretic hand. Another result of current technology is the development of modular systems for hand-based haptic interface which is a new control software that can be used at

home by patients.

The experimental method that allows determination of the variation laws of the kinematic parameters of the human hand movement is based on studying its movement using a data acquisition system based on hardware and software processing. Data acquisition system is called SimiMotion (www.simi.com) and is based on the analysis of motion video

capture.

An exoskeleton structure covering the required rehabilitation functions is shown in Fig. 2. The proposed solution is to develop a mechanical architecture consisting of a cascade of articulated elements, whose design to cover as much as possible the anatomic and functional finger phalanges, providing support for the actuation system.

Fig. 2. The exoskeleton architecture

Two architectures were analyzed: the 4-bar mechanism and the tendons mechanism. In tendons mechanism architecture, the three phalanges of the finger are realized as a serial structure of three rotation joints, each joint being controlled in cascade by the previous joint through a tendon-cable that is coupled on the pulleys system associated with each joint. In this case, kineto-static analysis principles were studied. Then it was considered a finger architecture based on

719

4-bar mechanism where the movement is achieved by successive deformations of the 4-bar mechanisms associated with each phalanx of the finger [16].

Fig. 3. The global architecture of one finger exoskeleton

The proposed architecture is a control architecture where the control is performed at the first active joint, the other joints being passive and realizing movements based on the associated bar system. The global architecture of one finger exoskeleton can be seen in Fig. 3. It is easy to observe the configuration of each finger phalanx and the control mechanism developed in the "4-bar" system. The proposed solution demonstrates flexibility and robustness which are required to meet the goals of rehabilitation operations.

Fig. 4. Movement of phalanges

Fig. 5. The global architecture of the whole hand exoskeleton

The mechanism allows the transmission of the control order, the angular variation of input to the output, with a transmission factor determined by the parameters of the mechanism. The movement of phalanges can be seen in Fig. 4. The global architecture of the whole hand exoskeleton is presented in Fig. 5. The virtual model was built using CAD SolidWorks.

III. THE ACTUATION SYSTEM

An assistive hand should assist a grasping force of human finger with a relative large grasping force. First, we analyze a new tendon driven system that can simulate variable compliance of a human fingertip using poly-articular tendon driven system. Then, a pneumatic solution is based on the Mc Kiben muscles and rotary-type soft actuators will be analyzed. The rotary-type soft actuator consists of a rubber tube and two silicone rubber tubes. In order to bend circumferentially, the polyester bellows is reinforced with fiber. In order to enhance prosthesis flexibility by keeping the intrinsic actuation solution and implementing simple control algorithm, we shall analyze an approach based on under-actuated mechanisms. When applied to artificial hands, the under-actuated mechanisms lead to an adaptive grasp, in a way that is closer to the human grasping than independent actuation. We will also investigate the feasibility of using Electro-Rheological Fluid (ERF) actuators in assistive hand actuation. We plan to develop compact actuators with an ability to supply high resistive torques in a controllable and tunable fashion, using the electrically controlled rheological properties of ERFs.

In this phase we will analyze, also, the possibility of the design, fabrication and testing of an ERF based assistive device and the innovative ERF actuators. Besides these actuation systems, we shall study the possibility to use the Shape Memory Alloys (SMA) in the category of technological actuators. For developing the gripping ability of the prototype robot- hand, we will use the Ni-Ti alloy in alternative forms. This solution will enable the system to operate more flexibly. In addition it saves energy consumption and increases the working speed.

IV. THE CONTROL ARCHITECTURE

The control architecture of the assistive hand is the center of cognitive and sensorimotor control system of the autonomous dexterous grasping and manipulation. In order to enhance the capability of the finger, the adoption of a suitable set of sensors is used. We will analyze a solution, currently tested, and based on flexsensors, usually used in data gloves in order to measure the bending of human fingers. These sensors are based on piezoresistive effect, and provide a variation of resistance proportional to the bending angle. The system will use highly sensitive capacitive-based pressure sensors to reliably quantify forces applied by the human hand. The sensor network will represent a comfortable sensor solution that connects wirelessly to the computer system and that also is wearable on the hand. Also, an observer will be used for the estimation of position and compliance force.

Our research will be focused on the distributed observer for the hyper-redundant models associated to the hand fingers.

720

The sensors are placed in several points or on the boundary of the arm segments. The conventional approach to sensor placement is to select the location based on considerations to ensure the necessary observability. Using quadratic performance measures, parameterized by the sensor locations, the optimal position will be found as the one yielding the maximum of the optimal control measures. The present research deals also with the optimal placement of sensors for the control of a continuum arm. The control algorithms are based by the state observers. The observability problems are solved by an approach derived from Luenberger observer type and the ―back-stepping‖ method.

The control architecture developed here and its algorithms shall constitute the bridge between the learning modules and the execution of the learnt skills in any unforeseen conditions to ensure the highest possible level of robustness. The algorithms involved in the proposed hierarchical multi-level multi-modal control architecture will make the system plan and execute fine in-hand dexterous manipulation. The developed architecture intends to be universal and valid for any intelligent glove.

At this point we also plan to create a virtual environment for rehabilitation therapy. Generally speaking virtual environments are suitable for rehabilitation therapy due to their inherent ability of simulating real-life tasks. Sensors are attached to the patient's hand to measure the motions, and the sensor readings can be transparently stored and evaluated by the system. Virtual environments will be used to develop a hand-reaching task for patients with chronic hemiplegia. Also, we will develop a library of VR exercises for hand rehabilitation in orthopedic patients. The system components are a computer, a Virtual Glove and the Intelligent Haptic Robot-Glove. The IHRG is integrated with VR exercises running on the computer.

V. THE HYBRID IHRG-FES CONTROL SYSTEM FOR

UPPER LIMB REHABILITATION

The electromyographic (EMG) signals of human muscles are important signals to understand the motion intention of human. The EMG signals can be used as input information for the control of the system. Functional electrical stimulation techniques mimic the voluntary initiation of the contraction while the IHRG may impose normal trajectories and perform sensorial feedback and measurements which can be used to estimate the rehabilitation process improvements. This work package will analyze the strategy that realizes the effective adaptation of the EMG-based assistive hand controller (i.e., robot-human interface) to any persons, IHRG-FES-based control strategies to perform daily life requested movements and quantification of the rehabilitation process improvements.

The model of the hand FES-based movement generation will be developed in Simulink&Matlab aiming to contribute to the control methods design and simulation. This model would be integrated within a hybrid one which adds the artificial actuation contribution of a mechanic glove. The model would be iteratively improved.

Based on the sensorial system different control methods will be developed and intensively tested in simulation before

to perform the in-vivo tests. The overall physical system will be realized based on the clinical and cosmetic specifications. The FES-IHRG system will be intensively test in laboratory and clinical environment. Based on the test observations its design and control methods would be improved.

VI. THE CLINICAL EXPERIMENTS

The rehabilitation exercises target four major hand impairments: finger range of motion, finger speed of motion, degree of independence of (finger fractionation), and finger strength. The speed of motion exercise is also monitored. If the speed performance does not exceed the target, a special signal is displayed on the screen. The range of motion exercise is executed separated, for the thumb and for the fingers. The exercise provides performance feedback for each finger. The design of the rehabilitation procedure relies on a moving target approach. The exercises are executed in blocks, each block containing a number of trials. Each block is assigned an average target. The individual targets for each trial are chosen from a normal distribution around the assigned average target. The average block target is computed by first averaging the performances of the patient in the previous block.

An analysis of currently used and experimental hand rehabilitation procedures will be performed. Hand movement retraining will be approached from a functional point of view and clinical tasks will be hierarchized according to their functional importance. A suitable clinical protocol will be developed, in order to ensure the maximum effect of the procedure on patient’s abilities

FES-based therapies enlarge the cortical projection of the exercised limb. Normal movements will be induced by a hybrid structure IHRG-FES-based control. Based on the IHRG sensorial system and a previous developed Grip Force Tracking System, the rehabilitation process will be quantified and the obtained data will be compared with those obtained during the current clinical practice.

A new database system stores the recorded data for a time period. Also, the therapist will have the possibility of monitoring the patient’s activity for a specified time period. A developed client-server architecture that brings the data from the rehabilitation site to the data storage site in real-time will be implemented. An analysis of literature will be performed so that the most appropriate parameters will be used to appreciate the clinical evolution of the patient, and the clinical evaluation protocol will reflect as correctly as possible the impact of the rehabilitation method on clinical parameters.

Transcranian magnetic stimulation (TMS) will be used to appreciate the cortical response to stimulation. Surface of the cortical projection of the hand, amplitude of motor answers, interhemispheric relations and influence are to be among the parameters of interest. A statistical analysis of the primary data may provide supplementary information regarding the effect of the current rehabilitation procedure, in comparison with previous experience and literature data. The visual and numerical parameters provided by the IHRG system will be validated as a useful approach of hand function evaluation in rehabilitation of stroke and multiple sclerosys patients.

721

Comparison with clinical tests (functional outcome) and the TMS findings will provide data to support that.

VII. CONCLUSIONS

This paper presented the design and development of an intelligent haptic robotic glove (IHRG) model for the rehabilitation of the patients that have been diagnosed with a cerebrovascular accident (CVA). Total loss or loss of range of motion, decreased reaction times and disordered movement organization create deficits in motor control, which affect the patient’s independent living. Recent studies have shown that intensive and repetitive training may be necessary to modify neural organization and recover functional motor skills.

The IHRG is an exoskeleton that supports the human hand and hand activities by using a control architecture for dexterous grasping and manipulation. The five-fingered assistive robotic glove was designed with mechanical compliance of human finger. The biomechanical elements of the exoskeleton assistive hand were designed considering: motion in different planes, adapted to patient’s hand, possible to train pinch and grasp, opening/closing game. A global architecture for the hand exoskeleton was presented. The actuation system was also studied and designed.

The paper is based on a project that is currently in work. That is why there were presented all the stages and experiments that will be done: the control architecture, the hybrid IHRG-FES control system for the upper limb rehabilitation, followed by clinical experiments.

ACKNOWLEDGMENT

This work is supported by PCCA 150/2012 grant of the Executive Agency for Higher Education, Research, Development and Innovation Funding (UEFISCDI).

REFERENCES

[1] P. S. Lum, C. G. Burgar, P. C. Shor, M. Majmundar, and M. V. D. Loos, "Robot-assisted movement training compared with conventional therapy techniques for the rehabilitation of upper-limb motor function after stroke", Arch Phys Med Rehabil, vol. 83, pp. 952-959, 2002.

[2] N. Hogan, H. I. Krebs, A. Sharon, and J. Charnnarong, "Interactive Robotic Therapist", MIT, Ed. United States, 1995.

[3] R. L. Farshid, F. Amirabdollahian, M. Topping, B. Driessen, and W. Harwin, "Upper Limb Robot Mediated Stroke Therapy—GENTLE/s Approach", Autonomous Robots, vol. 15, pp. 35-51, 2003.

[4] D. J. Reinkensmeyer, K. Leonard, E. M. Averbuch, A. McKenna-Cole, B. D. Schmit, and W. Z. Rymer, "Understending and treating arm movement impairment after chronic brain injury: Progress with the ARM guide", Journal of Rehabilitation Research & Development, vol. 37, pp. 653-662, 2000.

[5] A. Khanicheh, D. Mintzopoulos, B. Weinberg, A. A. Tzika, and C. Mavroidis, "MR_CHIROD v.2: magnetic resonance compatible smart hand rehabilitation device for brain imaging", IEEE Transactions on Neural Systems and Rehabilitation Engineering, vol. 16, pp. 91-98, 2008.

[6] A. Khanicheh, D. Mintzopoulos, B. Weinberg, A. A. Tzika, and C. Mavroidis, "Evaluation of electrorheological fluid dampers for applications at 3-Tesla MRI environment", IEEE/ASME Transactions on Mechatronics, vol. 13, pp. 286-294, 2008.

[7] L. Bobby Shields, A. John, ―An Antropomorphic Hand Exoskeleton to Prevent Astronaut Hand Fatigue During Extravehicular Activities‖, IEEE SYSTEMS and HUMANS, vol 27, No 5, 1997.

[8] M. DiCicco, L. Lucas, ―Comparison of Control Strategies for an EMG Controlled, Orthotic Exoskeleton for the Hand‖, IEEE ICRA 2004.

[9] Masahiro Takaiwa, Toshiro Noritsugu, Development of Wrist Rehabilitation Equipment Using Pneumatic Parallel Manipulator, Proc IEEE ICRA 2010, May 2010.

[10] J. C. Perry, J. Rosen, and S. Burns. ―Upper-limb powered exoskeleton design‖. 12(4):408–417, Aug. 2008.

[11] P. Schoonejans. Eurobot and exploration robotics. In Astra, ESA workshop on robotics and automation, 2009.

[12] P. Letier, M. Avraam, S. Veillerette, M. Horodinca, E. Motard, J.P. Verschuren, A. Schiele, A. Preumont, and M. Ilzkovitz. Exostation: 7-dof haptic exoskeleton and virtual slave robot simulator. In Astra, ESA workshop on robotics and automation, pages 3501 – 3506, September 2009.

[13] Mayuko Mori, Koichi Suzumori, Junichi Tanaka, Takefumi Kanda, ―Development of Rescue Robot using Ultra High Pressure Hydraulic Small Actuators and Field Tests for Verification of their Possibilities‖, in Journal of the Robotics Society of Japan, Vol.25, No.5.

[14] Mayuko Mori, Koichi Suzumori, Development of Power Robot Hand with Shape Adaptability Using Hydraulic McKibben Muscles, IEEE ICRA 2010.

[15] F. Lotti and G. Vassura,‖ A Novel Approach to Mechanical Design of Articulated Fingers for Robotic Hands‖ DIEM, Mech.Eng. Dept, University of Bologna, 2005.

[16] T. Lalibert´e and C. Gosselin, ―Simulation andDesign of Underactuated Mechanical Hands,‖ Mech.Mach. Th., 33(1), pp. 39-57, 1998.