Personalized Shared Public Displays: Supporting Multi-user Walk-by Content Personalization

| 0

Type: Master
Status: Completed September 2018
Student: Filip Zivkovic

Public display systems are increasingly becoming part of the urban landscape, with systems being deployed in venues such as railway stations, shopping malls, city squares, and universities. Most public displays today are simple slide-show systems that broadcast content based on a pre-selected schedule. However, public displays envisioned in the near future will provide a platform for running diverse interactive applications with highly personalized content. These applications will allow viewers to express their preferences both explicitly through interactive touch interfaces (active personalization) and implicitly using mobile handsets (walk-by personalization).

The aim of this project is to explore the development, deployment, and evaluation of web-based interactive display applications that can show personalized content influenced by the presence and preferences of multiple mobile phone users. The main focus of this project is on: 1) Designing and implementing two web-based applications, based on Public Transportation and USI Class Schedule information available through online API services, that can visualize and personalize such a content on public displays based on a set of personalization parameters. The apps should provide novel content visualization and scheduling algorithms that can support multiple, potentially conflicting, walk-by content personalization requests on shared public displays. 2) Integrating and deploying the new display applications within an existing touch-enabled display system at USI and supporting an existing Tacita personalization framework by implementing mobile interfaces that can be “loaded” into an existing iOS “controller app” and can be used to set personalization content parameters within the apps. 3) Evaluating the stability, latency, and overall performance of the apps within the existing display network. The evaluation will be performed based on the system log-files, time-stamps of events and personalization requests that occur within the apps, performance of the developed content visualization and scheduling algorithms, and the usage of system resources, e.g., CPU usage. The applications should be developed using HTML5 technologies within Play (Java) framework.

For more information contact: