Automated modified Ashworth scale adaptive impedance robotic assisted training platform for upper extremity muscle spasticity of neurologic disorder patients

Abstract

Robotic assisted training platforms have become a significant alternative to conventional training platforms as clinical therapeutic assistance to accommodate the increasing demand for neurological disorder physical treatments. Patients with neurological disorders usually experience conditions where their muscles are stiff, tight and prone to resist upon stretching, which in essence define muscle spasticity. The current method of muscle spasticity assessment is based on subjective assessment by therapists who heavily rely on their inner intuition, experience and skills. Based on the assessment, proper rehabilitation training tasks are prescribed as part of the training regimen. This, however, could be proven ineffective over the long run if the assessment is not done accurately. More so, in the case of robotic-assisted training systems used in training tasks, the deficiency in accurate information on muscle spasticity could largely affect any control strategy adopted to govern the robotic system. In order to address this problem, the research proposed to leverage on a synergetic combination of Modified Ashworth Scale (MAS) spasticity assessment tool and adaptive-impedance controller framed under a hybrid automata (HA) model applied on a patented upper limb rehabilitation platform, namely the Automated Muscle Spasticity Assessment System (A- MSAS). This required a dedicated spasticity characteristics model with control strategy during the assessment of muscle spasticity and an adaptive control based on impedance dynamics for the execution of the training tasks by A-MSAS. Spasticity characteristics model was developed using classification method and position-based impedance controller was adopted in strategizing the control of the A-MSAS. The latter was achieved through a dynamic mapping of the patient’s recovery parameters to the control parameters. The research involved clinical measurements of muscle spasticity from 39 subjects diagnosed with neurological disorders to classify the MAS scores quantitatively. From the research of assessment regimen it was found that by using spasticity characteristics model, the rate in predicting the MAS score of the subjects was 92.86% accurate. Meanwhile for training regimen, the adopted control strategy has resulted in an average angular velocity reduction, by 28.75% for pre-catch phase while average angular velocity increase which there were observable boosts by 46.46% for post-catch phases. The controller objective has been proven by allowing a degree of compliance even as A-MSAS platform dynamically deviated from the desired trajectory; proportional to the feedback received. Based on the findings, it was conclusively justified that an objective spasticity assessment prior to the training task would enhance the adapt- ability of the control strategy. This leads to a minimized muscle strain instigated from the feedback of spasticity characteristics pattern, hence warranting a more effective rehabilitation training.