Font Size: a A A

Real-time pathological tremor identification and suppression in human arm via active orthotic devices

Posted on:2014-10-07Degree:Ph.DType:Dissertation
University:Southern Methodist UniversityCandidate:Taheri, BehzadFull Text:PDF
GTID:1454390005495706Subject:Engineering
Abstract/Summary:
Tremor is a rhythmical and involuntary oscillatory movement of a body part. Along with the social embarrassment, tremor can be debilitating to the accomplishment of daily tasks. Mechanical loading via wearable exoskeletons, orthoses, has been under investigation as a non-invasive tremor suppression alternative to medication or surgery. In this method the challenge is identifying and attenuating the tremorous motion without adding resistance to the patient's intentional motion. In this research, three different control algorithms were designed to calculate the proper suppressive force to be applied by the orthosis to the patient's arm. The first control algorithm is based on a sinusoidal model of the tremor and the controller output, the required suppressive force to attenuate tremor, is sinusoidal. The second method is using an adaptive frequency estimator to find the tremor fundamental frequency in real time. The tremor amplitude is variable and the resistive force estimated by the controller has variable amplitude and the estimated frequency of the tremor. The third method was developed by using the so called backstepping method as a nonlinear control design tool. The main part of the controller is a high-pass filter and the backstepping controller finds the required suppressive force to make the output of the high-pass filter to converge to zero. Stability and robustness of the closed-loop system against the joint parametric uncertainties were analyzed for the proposed control algorithms. An experimental setup was designed and developed to emulate the dynamics of a human arm joint with tremorous motion. A pneumatic cylinder was used to apply orthotic suppressive force. The force produced by the pneumatic cylinder was controlled using a backstepping-sliding mode controller. The algorithm was implemented with a NI cRIO real-time controller for two types of tremorous motion: parkinsonian and essential. In addition to accurate tracking of the tremor frequency, the experimental results show significant tremor suppression in the range of 29.9--38.7 dB (96.8--98.8 %) at the fundamental frequency, and 6.5--14.9 dB (52.7--82.0 %) at the second harmonic.
Keywords/Search Tags:Tremor, Frequency, Suppressive force, Suppression
Related items