Human Motion Tracking Based on Unscented Kalman Filter in Sports Domain

Noise‐aware manoeuvring target tracking algorithm in wireless sensor networks by a novel adaptive cubature Kalman filter

“…This is caused by the linearisation that makes the EKF only reach the estimation accuracy of the first‐order Taylor expansion. To avoid the adverse effect of linearisation on positioning accuracy, the unscented KF (UKF) and CKF without linearisation are created, whose estimation accuracy can reach the second‐order Taylor expansion and the third‐order Taylor expansion, respectively [32, 33].…”

Pedestrian Tracking Based on Improved Particle Filter under Complex Background

“…As a core problem in computer vision community,moving targets tracking has got extensive and in-depth research.There are three main algorithms for human body tracking algorithm, Calman filtering , Meanshift algorithm, and template matching method,etc.Template matching method needs complex calculation and cannot extracte target absolutely [1] ,Meanshift algorithm has fast tracking speed , but it is easy to fall into local optimum [2] ,and Calman filtering needs the target state satisfied with Gauss distribution,otherwise it will needs a large amount of calculation,even worse failure to tracking when facing the non Gauss and nonlinear problem caused by complex background or target occlusion [3] .In recent years,particle filter algorithm developed rapidly in target tracking field,and has become a hot topic in the research of tracking target field for it's ability to deal with nonlinear, non Gauss problem [4,5] .…”