Force-Myography Based Estimation of Energy Absorption Capabilities of the Human Arm for Robotic Tele-Rehabilitation Therapy
Loading...
Date
Authors
Ramos, Andres
Keyword
Robotics , Teleoperation , Rehabilitation , Upper-arm , Force myography , Remote therapy , Excess of Passivity , Passivity , Stability , Telerehabilitation
Abstract
The need for rehabilitation therapy has been significantly increasing in the last few decades, due to neurological disorders caused by stroke. The cost of therapy for such a wide variety of motor impairments is high and the availability of therapists is insufficient for the current demand.
Therapists could perform remotely, using teleoperated robots that allow patients to receive help at home. This can reduce costs of transportation and increase availability of therapists.
Therapeutic forces delivered by a robotic system can help improve therapy by providing unlimited neurological stimulation to patients, taking maximum advantage of the plasticity of the brain for a timely recover from partial or total loss of kinesthesia, proprioception, and somatosensation.
Telerobotic therapy should be safe and easy to perform by therapists and patients, and this remote interaction must be realistic enough for therapy to work. Hence, a proper control system needs to be developed. Some controllers have been designed to achieve safe interactions, with a trade-off between stability and performance. Techniques such as damping and scaling degrade the interaction forces and the effectiveness of the therapy. In addition, instability can be induced from the active nature of the therapist and communication delays, injecting energy into the robotic rehabilitation system and generating unexpected harmful forces.
In this work, improved interpolations of energy absorption capabilities of the human arm are estimated through upper-arm Force-Myography. This approach by itself relaxes stability conditions that affect performance in robotic rehabilitation systems.
A new method for estimating energy absorption capabilities is implemented, using recursive least squares. This method allows adaptive behaviour of the estimation, providing valuable information according to the patient's real-time muscle contractions and energy absorption capabilities.
A fast orthogonal search is developed for estimating time-series EOP profiles that represent real-time energy absorption capabilities of the human arm.
A novel controller is also developed to allow safe delivery of therapeutic energy to the human arm, with real-time estimation of upper-arm energy absorption capabilities. In addition, a new kinematic scaling approach is implemented to reduce therapist's effort while delivering therapy. These methodologies are successfully tested in a teleoperation setup with two robotic devices.