DataFusion ofMEMS IMU/GPSIntegrated System for Autonomous LandVehicle*
Meiling Wang · 2006
Toimprove theperformance ofMEMES IMU thecasethatthevehicle experiences dynamics rangenot basedintegrated system, thedatafusion methodintegrated Al included inthetraining set;thehardware costwillbe andKalmanfilter wasdiscussed. First, GPSsignal validity and increased foron-line training duetotheheavycomputational vehicle motion status wereidentified byusing fuzzy logics. And burden(2). whenGPSdoesn't outage, itsdataweresethigher weight tobe TheKalmanfilter depends onasetofmeasurements and usedtoguide ALV;otherwise, Al(Artificial Intelligence) based TheKamicsdel oaoseof estimes and integration algorithm withthehelp ofKalmanfilter wasadopted.aproper dynamics modeltoprovide optialma eiates ofthe Themethod canmakefull useoftheadvantages ofKF(Kalman states. The limiting factors oftheKalmanfilter based Filter) andAl,andreduce their limitations. INS/GPSintegration are:modeldependency, prior knowledge dependence, sensor dependency, linearization dependency and IndexTerms-Artificial intelligence, Kalmanfilter, MEMS soon(3). IMUIGPS, ALV,dead-reckoning Basedtheaboveanalysis, Albasedapproach integrated withKFwasdiscussed inthis paper. Itcanmakefull useof