Système robotique pour la mobilisation des membres inférieurs d'une personne paraplégique

This thesis is the first to be carried out within the framework of the project "Cyberthosis for Paraplegics" in progress at the "Laboratoire de Systèmes Robotiques" (LSRO) of the "Ecole Polytechnique Fédérale de Lausanne" (EPFL). This project, initiated in 1999 by the "Fondation Suisse pour les Cyberthèses" in Villeneuve (FSC), is based on the reflections on the current means of physical rehabilitation of spinal cord injuries people. Eager to bring new methods of rehabilitation, the FSC initiated a program focused on the development of rehabilitation devices and overground walking assistance devices, associating an orthotic system with the electrical muscle stimulation. The main objective is to create an active muscular participation while preserving the physiological dynamic effects of the movements and thus to accelerate the process of rehabilitation. Three developments corresponding to three successive phases of rehabilitation are envisaged. A. MotionMaker™: stationary system for mobilization and muscular reinforcement. B. WalkTrainer™: mobile equipment allowing overground locomotion to stimulate neural plasticity and relearn walking. C. WalkMaker™: Assistance device for autonomous walking. The first stage of this thesis was to validate, on the knee joint, the closed loop control of electrically stimulated muscles with the information provided by an orthotic system. To do this, a robotic knee was designed. An electric motor controls the knee bending/extension, with resistance or assistance to the movement. Integrated sensors transmit, in real time, the values torques and articular positions to the dataprocessing programs which control the orthosis and the associated stimulator. This allows the adjustment of the resistance during movement and the adjustment of the electrical stimulation of the quadriceps muscles according to the kinematics and dynamics chosen. With this prototype we demonstrated the feasibility of closed loop control of electrically-stimulated muscles on able-bodied people as well as on people with spinal cord injuries. In a second stage, a more complete device, the MotionMaker™, was developed. It comprises two orthoses for the lower limbs. In the current state, the control unit allows the management of the 3 articulations (hip, knee and ankle) and of 7 muscles on each side (rectus femoris, vastus medialis and lateralis, gluteus maximus, semimembranosus and semitendinosus, tibialis anterior and gastrocnemius). Several algorithms of electrical stimulation control were proposed and evaluated in this thesis. They are either based on a mathematical model of the 7 muscles according to the stimulation current and the joint positions or on an iterative learning control. The development of this cyberthosis is currently sufficiently advanced to carry out, on able-bodied subjects, leg-press movements with control of the strength provided by the electrically-stimulated muscles. Thus the outcome push forces on the feet follow the values defined for the exercise. This work is focused on the movement of both above mentioned devices and on the model and the control of the electrical muscle stimulation. The mechanical parts have not been designed by the author.

A Neurocomputational Model of Impaired Imitation

Aude Billard, Biljana Petreska von Ritter-Zahony

This abstract addresses the question of human imitation through convergent evidence from neuroscience, using tools from machine learning. In particular, we consider a deficit in imitation of meaningless gestures (i.e., hand postures relative to the head) following callosal brain lesion (i.e., disconnected hemispheres). We base our work on the rational that looking at how imitation in apraxic patients is impaired can unveil its underlying neural principles. We ground the functional architecture and information flow of our model in brain imaging studies. Finally findings from monkey brain neurophysiological studies drive the choice of implementation of our processing modules. Our neurocomputational model of visuo-motor imitation is based on selforganizing maps receiving sensory input (i.e., visual, tactile or proprioceptive) with associated activities [1]. We train the connections between the maps with anti-hebbian learning to account for the transformations required to translate the observation of the visual stimulus to imitate to the corresponding tactile and proprioceptive information that will guide the imitative gesture. Patterns of impairment of the model, realized by adding uncertainty in the transfer of information between the networks, reproduce the deficits found in a clinical examination of visuo-motor imitation of meaningless gestures [2]. The model makes hypotheses on the type of representation used and the neural mechanisms underlying human visuo-motor imitation. The model also helps to gain more understanding in the occurrence and nature of imitation errors in patients with brain lesions.

2006

Control strategies for a verticalized rehabilitation robot

Yves Stauffer

Caring and rehabilitation of spinal cord injuries have greatly evolved during the twentieth century. Unfortunately standard training still heavily relies on manual mobilization of the patient. Which not only results in short training periods, but also prevents the patient's muscles from contracting, if voluntary control is lost. In order to cope with these issues the Fondation Suisse pour les Cyberthèses (FSC) in collaboration with the Laboratoire de Systèmes Robotiques (LSRO) from EPFL initiated a three step reeducation project. The main idea is the combination of motorized orthoses and closed loop muscle stimulation in order to enable an active participation of the patient, with a smart muscle control. The three complementary projects are: MotionMaker: stationary device with hip, knee and ankle mobilization combined to muscle stimulation. WalkTrainer: mobile device for true overground deambulation, with leg and pelvis orthoses, active body weight support and muscle stimulation. WalkMaker: lightweight orthosis with muscle stimulation, for enhanced autonomy. This work focuses on the control of the WalkTrainer, both of the motors and the muscle stimulation. First optimal trajectory generation is discussed. For the pelvis a three step strategy is presented. That consists of measuring motions on valid subjects with different parameters, identifying the corresponding model that can then be incorporated in the trajectory generator of the WalkTrainer. Of course the same work could be undertaken for any motion, such as the ones applied to the legs or trunk. Second, a novel off line EMG based muscle model and closed loop stimulation algorithm were developed. The trajectory generation and the muscle stimulation were then combined to obtain a novel, third generation, rehabilitation strategy. That optimizes the afferent feedback of the subject, while allowing an active participation of his musculature. That strategy was successfully tested on 6 paraplegic subjects at the beginning of 2008.

EPFL2009

Control strategies for a verticalized rehabilitation robot

Yves Stauffer

EPFL2009