Developing a voice controlled wheelchair with enhanced safety through multimodal approach

Abstract

The rising population of disabled and elderly community and lack of caretakers to look after them have become a crisis in most countries. Voice controlled wheelchairs give its user a chance to interact with the wheelchair in a humane manne. Due to the uncertaities of the voice commands and unreliability of using one interaction modality, user safety is compromised. In natural wheelchair user and caretaker conversation a lot of distance related uncertain instructions like ‘little’ and ‘far’ will be used. With incorporating these uncertain terms, the interpretation of voice commands will be enhanced. In most existing systems, Joystick support is used as backup modality. Even though this is intended as a safety feature, accidental joystick operation can lead to unfortunate situations. Therfore a reliable safety system considering multimodal aspect of two modalities is needed. Therefore this paper proposes a voice controlled intelligent wheelchair system incorporating uncertain voice information with safe navigation system using multmodal approach. Uncertain Information Module(UIM) have been introduced in order to interpret uncertain voice information and Multimodal Safety System is introduced in order to ensure user safety. Usability experiments have been carried out to evaluate the reliability of voice control system and Multimodal safety system.