Tracking Algorithms Represented as Classes
In this paper we consider some possibilities for applying object oriented approach to developing Target Tracking (TT) programs. We examine one of the main parts of TT algorithms — track evaluation — and propose a structure of classes that may simplify and alleviate creating and testing TT programs. These classes implement tracks, and consist of data and methods describing track kinematics - vectors, matrices and filtering algorithms. The track hierarchy contains classes for Linear Kalman Filter, Extended Kalman Filter and Probabilistic Data Association Filter. An example shows how Interacting Multiple Model (IMM) filter can be implemented using objects of these classes.
KeywordsTracking Algorithm Target Track Virtual Function Probabilistic Data Association Virtual Void
Unable to display preview. Download preview PDF.
- 1.Bar-Shalom Y., Blair W.: Multitarget-Multisensor Tracking: Applications and Advances, vol 3. Artech House (2000)Google Scholar
- 2.Bar-Shalom Y., Xiao-Rong Li: Estimation and Tracking: Principles, Techniques and Software. Artech House. (1993).Google Scholar
- 3.Bar-Shalom Y., iao-Rong Li: Multitraget-Multisensor Tracking: Principles and Techniques. YBS (1995)Google Scholar
- 4.Blackman S.S.: Multiple-Target Tracking with Radar Applications. Artech House (1986)Google Scholar
- 5.Georgiev Vl., Georgieva J. et all: Manual for using of computers and programming in C++. Sofia (2000) (in Bulgarian)Google Scholar
- 6.Karaivanova A., Djerassi E.: Track Formation Using MHT Approach. Proc. of the International Conference on Mathematical Modeling and Scientific Computations. DATECS Publishing. Sofia (1993) 31–33Google Scholar
- 7.Simov G.: Programming in C++. SIM. Sofia (1993) (in Bulgarian)Google Scholar