Object trajectory simulation : an evolutionary approach

Abstract

The ability to successfully predict the trajectory of an entity can have numerous interests. Using trajec- tory prediction we propose to enhance Radio-Frequency IDentification embedding intelligent behaviours that al- low these systems to improve their accuracy in the de- tection and guidance of personal in RFID enabled in- frastructures. This paper proposes a representative ap- proach to a simulated area filled with sensors and trav- elled by an ob ject. The ob ject has an initial and end points and does random movement between them. The remaining unknown path must be provided by the sen- sors and the prevision module while error metrics must be calculated dynamically to help the prediction of the followed path. In this scenario, a trajectory is a path that an entity follows through space between iterations, and it can be represented as a set of coordinates sorted over time. The level of accuracy needed for the predic- tion model and objectives of this environment required a grid like representation. The presented solution is a multiple dimension structure that covers the environ- ment variables and entities of a move within the envi- ronment. An application has been developed to simulate a censored place driven by a random trajectory object and to calculate as accurate as possible the path of the object.