Different Filter Approaches and Performance Analysis of Fundamental Sensors in Autonomous Ground Vehicles

Abstract

Autonomous ground vehicles are being designed and produced for industrial or military purposes. Since most of these vehicles are high cost the sensors used are high sensitive. However, sensors used in low cost robotic applications have low resolution. When non-precision sensor data added to measurement errors caused by different typical of geometry and drive methods, process functions of robots are hitching. To prevent this, five simple-structured filters are designed and compared in this study for low cost distance sensors and GPS sensors data. It has been observed that accuracy of hybrid-structured filters are more suitable than mono-structured filters in the results obtained.