Simulation-based usability evaluation of visually impaired outdoor navigation using sensor fusion

Chathurika Sewwandi Silva, Prasad Wimalaratne

Abstract: Conducting evaluation experiments with assistive navigation in real-life environments with visually impaired subjects is challenging for several reasons. Hence, simulation-based usability evaluation experiments are a pragmatic and cost-effective approach in such studies. This study presents sensor fusion in a virtual simulation of assistive navigation, including obstacle detection, localisation and motion planning. The obstacle detection involves a complementary sensor fusion approach to detect static and dynamic obstacles via vision and obstacle sensing. Localisation is carried out using Kalman filter-based sensor fusion of measurement data acquired by inertial measurements using an accelerometer and gyroscope as well as global positioning system (GPS) data of the visually impaired navigator. Precise motion planning for local and global navigation is based on the fused signals through obstacle detection and localisation modules. The evaluation showed that experiments effectively estimate the trajectory of a visually impaired navigator by fusing the heterogeneous sensor measurements in a virtual environment.

External IDs:dblp:journals/ijsnet/SilvaW22