Improving Interactions with Spatial Context-aware Services

Smart device users can observe mobile services becoming better in quality and quantity on a monthly or even daily basis. In this thesis, we concentrate on the services that utilize the contextual information about the user, such as location, the user's preferences, local weather, time of the day and many other factors. These services, so-called context-aware services, are especially well-known to smartphone users by popular applications Apple Siri, Google Assistant and Microsoft Cortana. However, even though these applications are a large step towards better user experience with smart devices, these services are not yet ideal. This dissertation highlights and suggests how to cope with two major issues that users experience while interacting with current mobile context-aware services. The first highlighted issue is the lack of data on space usage rules in the user's location, such as legality of smoking, swimming, flying a drone and many other activities. In this thesis text, we investigate related work in the space usage rules area and suggest several methods to automatically and semi-automatically collect these rules into a single database and use it to inform users about space usage rules in their location and places that they are planning to visit in various ways. The second problem we cover is the lack of research in the scenic routing area, and specifically methods that would use raster information about the route's surroundings (e.g., images from Google Street View) in order to enhance user experience. We suggest two systems that are designed to improve the user interactions with scenic route generators. The first proposed system utilizes machine learning techniques in order to suggest user-specific scenic routes in the user's location. The second one analyzes different sides of the road to help to select a sit in a touristic bus that provides the most scenic views on the way. In this thesis, we make four main contributions to the HCI (Human-Computer Interaction) field, three for the first problem and one for the second one. These contributions present novel methods that target the same goals: to improve user interactions with context-aware services and suggest new ways how these applications could be further improved.

Publication year:2017

Keywords:space usage rules, computer vision, machine learning, scenic qualities, scenic routing, context-aware applications, location-based services

Accessibility:Open