室内环境下基于IMM-EKF算法的移动目标定位.pdfVIP

计 算机 研 究与 发 展 7544“ssnlOOO一1239．2014．2Ol313(73 JournalofComputerResearchandDevelopment 51(11)：2408—2415，2O14 室内环境下基于 IMM-EKF算法的移动 目标定位 张云洲 付文艳 项 姝 魏东飞 杨 兵 (东北大学信息科学与 ¨』二程学院 沈阳 110819) (zhangyunzhou@ ise．neu．edu．cn) IM M -_EKF Algorithm’B·ased IndoorM ovingTargetLocalization ZhangYunzhou，FuW enyan，XiangShu，W eiDongfei，andYangBing (CollegeofInformationScienceandEngineering，NortheasternUniversity，Shenyang110819) Abstract Itisachallengingtask to improvetheaccuracyofthemobilelocalization in IOS (1ine—of sight)andNIOS (non—line—of—sight)mixed environment．W hen theM N (movingnode)movesin indoorenvironment，dueto the obstaclessuch aswalls，doors，and furniture，thecommunicating signalbetween M N and ANs(anchor nodes) change between IOS and NI()S frequently and randomly。which hasnegative effecton theaccuracy ofMN location estimation．To guarantee the accuracy，aKF (Kalman filter)based IM M (interactingmultiplemode1) isproposed to filterthe measureddistanceundertheIOS／NLOSmixedenvironment．Duetothedifferentcharacteristicof rangingerrorsbetweenIOS andNIOS，two parallelKFswithdifferentparametersareemployed in ordertosuitforIOSmodeandNIOSmode。bothofthemodeprobabilitiesarecalculatedbythemode likelihoodsandhistoryprobabilities．Themodestransitionbetween IOS／NIOS modesiSbasedon Markovchainandmodeprobabilities．Theweightedmeanofthetwomodesfilteringresultsistaken astheestimated distanceofIM M ．Oncetheestimated distancesareobtained，theEKF (extended Kalman filter) iS applied to Iocate the M N． The simulation results demonstrate the IM M can significantlym itigatethepositiverangeerrorandachievehighlocalizationaccuracy． Keywords wirelesssensornetwork (W SN)；mobilelocalization；indoorenvironment；non-line—of sight(NI()S)；interactingmultiplemodel(IMM ) 摘 要 如何在视距 (1ine—of—sight，IOS)与非视距 (nonline—of—sight，NI()S)混合的室内环境 中提 高移 动 目标定位的精度 ，这是一项富有挑战性的工作．移动 目标在