Designing Android applications with both online and offline voice control of household devices

Utilizing assistive technology to aid persons with neuromuscular disabilities, this study compares two methods for voice control of household devices through an input-output board driven by an Android application on a smart phone. Two types of voice recognition software, online Google and offline Pocketsphinx, were used and tested for accuracy of several commands on a television remote. Each application was designed to recognize keywords and send signals to corresponding pins on the input-output board. The execution of a command was enabled through the interface of the smart phone with a television remote. Applications were tested for accuracy depending on distance between person and device as well as level of ambient noise.