Automated Route Selection for Navigation

Automated vehicle navigation aids for drivers represent an area of considerable research interest. Whereas most previous work on the problem has assumed that such aids should be presented as maps (graphic output on a CRT), recent research suggests that drivers may perform better if given verbal (vocal) directions, rather than a graphic map. Furthermore, anecdotal evidence suggests that people often give directions to others by describing routes which are longer yet simpler than the ones they would normally use themselves. This paper presents an algorithm for selecting paths for travel by road. The algorithm is based on the principle that `ease of description is an important consideration in route selection. The route description is represented by a frame, and the route description cost is approximated by the number of slots in that frame. Then, total route cost is just the sum of the route length and the weighted route description cost.