Abstract ▼ AI-Helper

It is a very important problem that we recognize the accurate position at the vehicle navigation. Mainly used common methods of position calculating are GPS (Global Positioning System), DR (Dead-reckoning) and INS (Inertial Navigation System). GPS is useful in accumulating for few errors, but it has...

It is a very important problem that we recognize the accurate position at the vehicle navigation. Mainly used common methods of position calculating are GPS (Global Positioning System), DR (Dead-reckoning) and INS (Inertial Navigation System). GPS is useful in accumulating for few errors, but it has errors like Troposphere error, Multipath error and so on. Because Dead-reckoning of processing time takes so long, it can't measure the accurate position at the real-time environment. If a vehicle travels long distance, there is a defect that accumulates errors at INS. Car navigation system uses combinations of GPS, DR and INS to solve the problem, but these systems have considerably defects about economics and complication of system expulsion. In this thesis, I propose a position compensation algorithm which is only using GPS data to solve upper problems. It is deduced a rectilinear equation with coordinate values of digital map for the purpose of the GPS errors compensation. And it applies the first coordinate value to the Kalman filter for estimating the moving position point at a vehicle. Then, I compensate GPS errors compare estimated point with GPS point errors. The experimental results have shown with compensated position data that are applied data getting in real travel to algorithm. The result is that uncompensated GPS data show error of the mean 3 meter distance, and error of maximum 5 more meter range. In case of application with proposed algorithm, compensated GPS data show error of the mean 1 meter distance, and error of minimum about 0.2 meter.

주제어