Based on our insights and results from the first funding period, we aim in the second phase of the project at broadening the range of possible input data, enhancing the flexibility of our data structures and algorithms, and exploiting their applicability for trajectory analysis, data mining, and route planning.
In the first phase, we mainly focused on trajectory data where movement is restricted to or at least guided by an underlying graph (as, e.g., a road network). This allowed us to develop efficient indexing techniques based on preprocessing the graph. In the second phase, we plan to extend our research also to trajectories of objects that move freely in two-dimensional or even three-dimensional space.
Furthermore, our trajectory indexing techniques developed in the first phase are designed to work well on static trajectory sets and a static underlying graph, and are tailored to a specific set of queries. But for many applications, it is important to be able to cater for dynamic changes in the data in an efficient manner, to be able to add and delete trajectories, to incorporate new dimensions of the data on demand, and to allow for user-defined data analysis, query types, and visualization. Therefore, one important objective of the new project is to provide this level of flexibility to the user.
Finally, we want to demonstrate the usefulness and importance of the algorithms and data structures developed in both phases by investigating important applications from the realm of large-scale trajectory analysis. In particular, we will focus on trajectory clustering and trajectory-based route planning.