Distributed Particle Filters for Object Tracking in Sensor Networks

摘要: A particle filter (PF) is a simulation-based algorithm used to solve estimation problems, such as object tracking. The PF works by maintaining set of “particles” candidate state descriptions an object’s position. determines how well the particles describe observations and fit dynamic model, in order form estimate. drawback basic that functions collecting all data at fusion centre. This leads high communication energy costs resourcelimited network sensor network. In this thesis, we analyze determine it can be modified for efficient use Our main priority keep low since increases lifetime. We propose two innovative filtering algorithms which minimizes associated costs.