A Wheelchair Platform Controlled by a Multimodal Interface

In this paper, an embedded speaker-independent isolated word speech recognition module based on STM32F103C6T6 (STM32) is designed. The speech recognition algorithm includes: voice signal pre-processing, feature extraction and recognition. In addition, a communication module is designed based on the STM32 for receiving control commands from computer by universal serial bus. The result is that we realize a multimodal interface for wheelchair control in total three types of available inputs: speech, keyboard and traditional joystick. To evaluate the performance of wheelchair platform control by speech, four subjects participated in the experiment. The results demonstrated all of them can control the wheelchair by speech efficiently with an average success rate and response time are 92% and 2.2s respectively.