Tuning Algorithms for Control Interfaces for Users with Upper-Limb Impairments

Abstract

Approximately 40% of Americans with disabilities cannot operate wheeled mobility devices and computers adequately because of diminished upper-limb motor control, sensory limitations, and cognitive impairments. We developed tuning software that can customize control interfaces for individuals with upper-limb impairments. This study compared the differences in each parameter among different diagnostic groups. The age of the subjects ranged from 18 to 80 yrs. The participants were classified into the following groups: athetoid cerebral palsy, spastic cerebral palsy, multiple sclerosis, upper-limb spasticity, and control. We used a validated tuning software protocol to customize an isometric joystick before a virtual tracing or driving task. Tuning parameters were then compared across groups. Seventy-five subjects were included. Gain, the parameter responsible for force-to-output ratios, in each directional axis (leftward gain: P = 0.018; rightward gain: P = 0.003; reverse gain: P = 0.007; forward gain: P = 0.014) was significantly different across the diagnostic groups. Post hoc analyses showed that the control group required smaller leftward gain than spastic cerebral palsy, multiple sclerosis and upper-limb spasticity groups and smaller gain in all other directions compared with spastic cerebral palsy. Gain may be a useful parameter in tuning by clinicians, and efforts aimed at gain customization may aid the development of commercially available tuning software packages.