Pedestrian Navigation Systems: a Case Study of Deep Personalization

Requirements Engineering (RE) focuses on obtaining the user goals and environmental constraints of a proposed system. In traditional RE, users are treated as a consumer-class: what holds for one member is assumed to hold for the rest. Personalization comes through providing options that a user can set at runtime to tailor features of the system to his or her personal preferences. In our work, we take up the personalization issue in more detail. In particular, we believe that some systems require a "deep personalization" that includes knowledge of an individual user's skills and limitations. In some cases, these skills and limitations might not be self-aware, i.e., a user cannot accurately self-reflect on his or her skills and weaknesses. In this paper, we will demonstrate the notion of deep personalization in the domain of personal navigation systems. We find this an interesting domain for several reasons: (1) There is a domain theory of navigation skills that draws from both Cartography and Psychology. (2) There are individual differences in navigation skills. (3) An individual user may not be self-aware of his or her skills. (4) If a system is delivered that does not match the skills of the user, it may be less than effective, and at worst, abandoned.