Implementation of pedestrian navigation system for the visual impaired

Abstract

We implement a portable pedestrian navigation system for the visually impaired. The system consists of a smartphone with a voice recognition function, a budget direction guide and an osteoacusis headset, which can provide auditory guidance directions while recognizing the sound of the surrounding environment. Unlike previous pedestrian navigation smartphone apps, the developed system guides the user in walking direction by the scale of the left and right stereo sound of the headset worn by the user, voice guidance regarding forked or curved paths is given several meters in advance according to the speed of the user, and the user is immediately warned of walking in the opposite direction or proceeding off the path. To overcome GPS positioning errors, we proposed a robust trajectory planning algorithm for position errors. Experimental results show that the average directional angle error is 8.56 degrees (standard deviation: 14.78), average difference between GPS path and scheduled path is 5.55m (standard deviation: 3.79) and maximum difference is 13.22m in the experimental path, based on which can be stated that the user was navigated relatively stable by the developed system.