Pattern Recognition of Spastic Motion

Abstract

This article involves an application of an intelligent control system in biomedical engineering specifically dealing with patients who have difficulty in making smooth commanded arm motions. Individuals who experience random motor spasms during forearm motion were selected from diagnostic groups including head trauma, cerebrovascular accident, and cerebral palsy. An engineering analysis technique involving pursuit target tracking examination in the error phase plane provided distinct insight into the unique characteristics of uncommanded motion. Approximate reasoning, or fuzzy logic, is applied to the problem of recognition of an uncommanded motion of the forearm. Two fuzzy logic pattern recognition algorithms are presented as potential devices to distinguish between commanded and uncommanded motion. The fuzzy logic approach in the error phase plane is shown to have fairly good results, while the acceleration-velocity phase plane has reduced information and provides results that are less reliable. These patterns are then used as the basis for a proposed adoptive controller that would assist the individual in overcoming the motor spasm and returning to useful control.