|
|
自动化学报 2005
机器人定位中的自适应粒子滤波算法, PP. 833-838 Keywords: Robotlocalization,particlefilters,K-Ldistance,probabilityretrieval Abstract: ？Theresearchofrobotlocalizationaimsataccuracy,simplicityandrobustness.Thisarticleimprovestheperformanceofparticlefiltersinrobotlocalizationviatheutilizationofnoveladaptivetechnique.Theproposedalgorithmintroducesprobabilityretracingtoinitializeparticlesets,usesconsecutivewindowfilteringtoupdateparticlesets,andrefreshesthesizeofparticlesetaccordingtotheestimationstate.Extensivesimulationsshowthattheproposedalgorithmismuchmoreeffectivethanthetraditionalparticlefilters.Theproposedalgorithmsuccessfullysolvesthenonlinear,non-Gaussianstateestimationproblemofrobotlocalization.
|
