

This paper describes a series of trials based on mathematical information engineering that were designed to enable assistive technologies to help with the travel and mobility of the visually impaired. In order for technology to be of assistance to travel, two different concepts are necessary, i.e., to assist travel performance and to assist travel planning. In the former case, we developed a GPS-based “travel assist” system (TAS) so that an adaptive fuzzy inference neural network could be employed to stabilize positional guidance. In many cities and towns in Japan it is often the case that an area of buildings surrounding a particular street rather than the street itself assumes the identity of a recognized landmark. Therefore, GPS measurements acquired by a pedestrian walking on such a street are likely to randomly output landmarks on both sides of the street. The adaptive fuzzy inference neural network enabled the TAS to guide the user to just the landmark that was required by the pedestrian. In the latter case, we developed an AHP-based “plan assist” system (PAS), whereby the system assists in the self-determination of the route selection. These concepts are of special importance to visually impaired persons, even those with guide helpers, because such self-determination might enhance their efficacy when traveling, and, as a result, might add to the self-sufficiency of the visually impaired.