An embedded real-time hands free control of an electrical wheelchair
Agnès Ghorbel, Nader Ben Amor, Mohamed Jallouli · 2014
The main aim of this work is to propose a HMI alternative for hands-free control of an electrical wheelchair using both face and eyes information. An image processing based technique for the guidance of an electrical wheelchair is used. As processing unit, we choose the Panda board ES development kit based on the TI OMAP 4 SoC (System on Chip). This board is connected with the wheelchair for processing of the video streaming. The wheelchair command technique is based on image segmentation and facial detection. The adopted model relied on combining 2ARMs processing in a single chip for improving real-time performance, enhancing system flexibility and reducing system cost and power. The preliminary results are promising since a total processing time of about 0.9s is obtained.