Tracking Algorithms Based on Dynamics of Individuals and MultiDimensional Scaling

Jose Maria Cabero, Fernando De la Torre Frade, I. Arizaga, and A. Sanchez
IEEE international symposium on wireless pervasive computing 2008, May, 2008.


Abstract
Accurate indoor location of people is a key aspect of many applications such as resource management or security. In this paper, we explore the use of radio communication technologies to track people based on their dynamics. The network consists of two types of radio nodes: static nodes (anchors) and mobile nodes (individuals). From a set of sparse dissimilarity matrices with information about proximity or estimated distances between nodes and individuals' dynamics at each time instant, we infer individuals' trajectories. Depending on the information available, two algorithms are proposed: Dynamic Weighted Multidimensional Scaling with Binary Filter (DWMDS-BF) and Dynamic Weighted Multidimensional Scaling based on Distance Estimations (DWMDS-DE). DWMDS-BF is an algorithm that implements a Binary Filter function that obtains very good tracking results when only connectivity information is available and DWMDS-DE is designed for those networks where a goodestimation of distances between nearby nodes is available. Both algorithms implement a dynamic component that regularizes the obtained trajectories according to individuals' dynamics. Extensive simulations show the effectiveness and robustness of the proposed algorithms.

Keywords
Multidimensional scaling, indoor people tracking

Notes
Note: the associated projects are component analysis and indoor people localization

Text Reference
Jose Maria Cabero, Fernando De la Torre Frade, I. Arizaga, and A. Sanchez, "Tracking Algorithms Based on Dynamics of Individuals and MultiDimensional Scaling," IEEE international symposium on wireless pervasive computing 2008, May, 2008.

BibTeX Reference
@inproceedings{Maria_Cabero_2008_6041,
   author = "Jose {Maria Cabero} and Fernando {De la Torre Frade} and I. Arizaga and A. Sanchez",
   title = "Tracking Algorithms Based on Dynamics of Individuals and MultiDimensional Scaling",
   booktitle = "IEEE international symposium on wireless pervasive computing 2008",
   month = "May",
   year = "2008",
   Notes = "the associated projects are component analysis and indoor people localization"
}