Speech-controlled cloud-based wheelchair platform for disabled persons

Abstract

This paper describes the development of a prototype speech-controlled cloud-based wheelchair platform. The control of the platform is implemented using a low-cost WebKit Speech API in the cloud. The description of the cloud-based wheelchair control system is provided. In addition to the voice control, a GUI is implemented, which works in a web browser as well as on mobile devices providing live video streaming. Development was done in two phases: first, a small, initial prototype was developed and, second, a full size prototype was build. The accuracy of the speech recognition system was estimated as ranging from approximately 60% to up to 97%, dependent on the speaker. The speech-controlled system latency was measured as well as the latency when the control is provided via touch on a so-called smart device. Measured latencies ranged from 0.4 s to 1.3 s. The platform was also clinically tested, providing promising results of cloud-based speech recognition for further implementation. The developed platform is based on a Quad Core ARM Mini PC GK802 running Ubuntu Linux and an Arduino UNO Microcontroller. Software development was done in JavaScript/ECMA Script, applying node.js.