Osprey Flight Path (OFP) is designed as an expandable mobile application that allows its users to glide quickly and easily across campus. Imagine the first day of a freshman who tries to find his or her way not only to campus but between classrooms as well. With the latest technologies such as Global Positioning System (GPS) and smart phones, it is easy to get to campus from his or her apartment by simply following the suggested routes. However, once on campus, the freshman has no automatic guiding system anymore. The existing guiding systems such as GPS devices and Google Maps do not suggest routes between buildings on campus. OFP is therefore created to fulfill such needs of students.
By allowing users to plot their starting building and the destination building with a user-friendly graphic interface, OFP uses the efficient A* search algorithm to find the shortest path from one building to another along a campus’s walkways such as side-walks, bridges, and other university/school specific modes of transportation. This path is then displayed point-by-point on the school’s map so that the user can easily follow the route as shown on a mobile device. It is currently being designed for implementation at University of North Florida, but can be modifiable to be used elsewhere and be expanded based on available support and interest by the community. This project has won the first place award in the 2011 School of Computing Student Symposium at University of North Florida. The project is also accepted and presented in the 2nd Annual Florida Undergraduate Research Conference on March 16-17 at Stetson University.
This project’s goal is to develop an android application that plays songs matching the step frequency of a runner using the phone’s accelerometer. In attempting this goal, some key information was learned. It was found that part of this goal is feasible though applying this concept to a user’s own database of songs proves difficult. One of the problems encountered is that there is no way to ensure that a user provides a wide enough range of song tempos in their personal libraries. Even assuming that a user has a wide enough range of music, how does an application accurately annotate this tempo data to the music? This can be done, however it is not 100% accurate. These issues suggest that this sort of concept is best applied as a supplement to streaming services such as Pandora. Additionally, combining the like/dislike aspect of Pandora with this service may also be beneficial to the user. In order to demonstrate this concept, we developed an application using SQLite as a database and running this application on an Android phone. Using an open-source pedometer application developed by Levente Bagi, we added a song selection process and music player. The algorithm we use for song selection is based on the algorithm from the paper titled “Development of An Automatic Music Selection System Based on Runner’s Step Frequency” by Niitsuma Masahiro et. Al, Keio University, Yokohama, Japan.
Felipe X. Aspillaga, a master student in School of Computing with a focus on software engineering, is developing an intelligent DJ software system for the iPad platform. The software system not only allows users to mix and play songs from their own music collection using a simulated DJ turntable, it also automatically suggests songs to be played next, helping amateurs and professionals to create pleasing compositions. Over his 15-year career as a professional DJ, Felipe has played in countless clubs, special events and festivals throughout Florida and Colorado and diligently documented his work with demos, live recordings, and several original and remix compositions of his own.
He is currently working with Dr. Ching-Hua Chuan to develop computational algorithms for automatically analyzing music features such as key and rhythmic patterns from audio signals. By applying these algorithms, they will be able to examine large number of compositions made by professional DJs to study their mixing strategies and personal styles, and eventually combine these strategies with established rules of music theory to make intelligent suggestions for mixing music in real time. Felipe will present the preliminary results of the system with Jonathan Cobb, a senior undergraduate in School of Computing, in the final project presentation of CIS4930/5930 Music Informatics and Computing on April 27. Felipe also gave a live demonstration of the system at the 12th International Conference on Music Information Retrieval, held in Miami on October 24 -27 in 2011.