Context-Aware Mobile System for Public Bus Transportation

Introduction

Out of 440.5 million mobile devices sold in the third quarter of year 2011, 115 million were smartphones, meaning that one of four buyers bought a smartphone (Gartner, 2011). Modern smartphones boast with powerful embedded sensors and thousands of applications that suit needs of various people. With the help of sensors in mobile devices the environment in which mobile device is used can be determined, making it an ideal platform for creating context-aware solutions (Lane, et al., 2010). To understand what context-aware applications are, the term context should be explained first. There are several definitions that define this term, each of them being slightly different from another, but the one most relevant to our field of research is definition from Dey and Abowd (Abowd, et al., 1999). They have defined context as “any information that can be used to characterize the situation of an entity. An entity is a person, place, or object that is considered relevant to the interaction between a user and an application, including the user and applications themselves.” This means any piece of information that can be used to enlighten the situation of participant’s interactions is context.

Furthermore Dey and Abowd (Abowd, et al., 1999) elaborated that some context types are more important than others in practice. These so called primary context types are location, identity, activity and time. Based on these primary context types, secondary context types for same entity can be identified. For example, given person’s location (e.g., getting location from mobile device’s Global Positioning System (GPS)) other entities in vicinity of our person can be determined. Such categorization of context types is a two-tiered system, where primary context types are on the first tier and second tier consists of all other context types, which share same characteristic and are attributed to the entity with primary context. Such categorization can help developers to choose how they will incorporate context in their applications.

With knowing what exactly context is we can now think about how to incorporate context in context-aware mobile applications. Context awareness is not something new; it was already mentioned in 1994 by Schilt and Theimer (Schilit & Theimer, 1994). They described context-aware computing as software, which exhibits ability “to adapt according to its location of use, the collection of nearby people and objects, as well as the changes to those objects over time.” As mentioned before the knowledge about mobile device’s environment can be obtained through its sensors, thus making them the core of context-aware application.

Three of the four primary context types can be fairly easily determined with today’s smartphone features (Lane, et al., 2010). Location is determined with use of the location service, which uses embedded GPS receiver, Wi-Fi networks and cellular network to locate mobile device (Brownworth, 2012) Identity can be determined with use of devices International Mobile Equipment Identity (IMEI) or any other identification method such as username and password. Since every mobile device has internal clock, determining time is just getting accurate reading from that clock. Last primary context type, activity, is hardest to determine using smartphone features. Using combination of different sensors that individually cannot represent activity itself, but they may result in better characterization of activity context (Gellersen, Schmidt & Beigl, 2002). Example: location is somewhere outdoors, user is sitting but is doing specific motion pattern of legs and absolute position is changing. Sensor data of the given situation can be related to a situation when user is cycling. The other possibility of determining user’s activity is based on knowing what application the user is using.

Taking all this into consideration our goal was to develop a context-aware mobile application in the domain of public bus transportation which would use three of the primary context types (location, time and identity) to adapt information shown to user, and the fourth one (activity or intended activity) will be predicted based on the past actions of a specific user.