REAL TIME KINEMATIC GLOBAL NAVIGATION SATELLITE SYSTEM – MULTIPATH DETECTION AND OPTIMIZATION FOR AUTONOMOUS CAR NAVIGATION USING ELECTRIC POWER STEERING

Abstract

The purpose of this thesis is to improve the existing RTK GNSS technology and electric power steering. The improvement of RTK GNSS is done by eliminating multipath signal in order to have a better the positioning accuracy. Two Emlid: Reach GNSS modules are used as base and rover station receiver. The multipath elimination is done by modifying the code in RTKLIB, so that there are two elevation masks for satellite filtering which are in the car’s heading direction and in the perpendicular heading direction. The elevation mask in heading direction is set to be 30° while the elevation mask in perpendicular heading direction is set to be higher than 30° to achieve more fixed solutions. The improvement of electric power steering is done by installing a rotary incremental encoder so that it can detect its current position. The rotary incremental encoder is equipped with safety validation measure to ensure that no miscount will happen as well as energy saving mode to reduce the power consumption by 36.08%. Various experiments are done to achieve more information that can support this thesis. When these improvements are implemented in an autonomous car, a selfdriving car which has great precision in any environment is possible to build.