2D motion prediction within wireless sensor network coverage

Taking advantage of collective information gathering and collaboration between nodes in a wireless sensor network we developed an efficient algorithm that can estimate future positions of a moving object inside the coverage area. The algorithm approximates the object's trajectory with line segments obtained using regression techniques. Distributed design, low computational power requirements and optimized communication describe the strengths of the proposed algorithm which can represent a standalone application for object tracking or might be part of complex applications helping in overall power consumption reduction by activating nodes as a moving object or event is approaching. A lab implementation of the algorithm has been developed and tested demonstrating the feasibility of such an approach.