CDS Summer 2014 Research - GeoSpatial Data Management Group

Date of Presentation: 
Wednesday, August 27, 2014
Quarter: 
2014 Summer
Research Focus: 

This week's CDS Wednesday seminar (August 27th) will feature an overview of two projects in the GeoSpatial Data Management Group. The talk will run from 12:00pm - 1:00pm in the UWT Tioga Building 3rd floor atrium. All are welcome.

 

PreGo: Dynamic Multi-Preference Location Inference based Routing System

Presenter: Aqeel Bin Rustum

Existing routing services such as Google Maps, Bing Maps and MapQuest Maps can find a shortest route between a source and destination pair based on travel time or distance. However, in real life commuters have other preferences in addition to those i.e. a route with less highways or more facilities. Sometimes, the commuters are willing to enjoy their trip by taking a path that follows a scenic coastal road even though with extra time. All of this brings us to end that the shortest commute time is not always the only preferred attribute for commuters. Hence, extending routing services to handle multiple user preference can contribute greatly to the trip planning services field.

User’s preferences can be represented by additional attributes beside the distance and the commute time. This means that a road network can be promoted to include multiple costs on each of its edges. But then how the shortest path algorithms such as Dijkstra’s algorithm and A*-Search algorithm can consider multiple costs when searching for the shortest path? Well, a trivial solution is to run a shortest path algorithm for multiple iteration and at each iteration we consider one attribute and at the end return the most common sequence of edges. However, this solution is time consuming so we need to find out an efficient way to approach this issue. So, in our talk today, we will present our PreGo framework solution.

PreGo can best be described as a preference learning, flexible and dynamic routing system. It is also dynamic in the sense that it does not depend on static road network graph. Rather, we first introduce a new structure termed: Attribute Time Aggregated Graph, constructed from volunteered geographic information attributes (VGIs) such as GPS traces, crime, accident reports, open access maps and so on. With this, it can be promised that PreGo can efficiently handle the user multi-preference routing in just one pass.

---------------------------------------------------------------------------------------------

Efficient Computation for Shortest Path on 3D Land Surface

Presenter: Nitin Arya

Finding the shortest path on real land surface forms an essential component inside many applications such as tracking the flow of hazardous materials and objects, e.g., chemical spills and forest fires.

Existing work suffers from scalability issue as it can compute the surface shortest path on very small area of the underlying space. 

In this presentation, we will present an efficient algorithm to compute the shortest path on 3D land surfaces. In addition, we will demonstrate a practical implementation of this algorithm and show the output visualization. In the future, this algorithm will be adapted and employed to build the surface spatial predictive index inside the AMADEUS framework.

---------------------------------------------------------------------------------------------

About the Series: The Center for Data Science seminar series is centered around intellectual exchange and interaction, and the audience is encouraged to ask questions during presentations. The goal is a seminar that looks less like a lecture and more like a spirited discussion of issues raised in a relatively brief presentation of a paper or a research project.