State estimation for a golf buggy via Differential Global Positioning System

Addresses the multi-sensor data fusion problem in the navigation of a two-wheel steerable vehicle converted from a golf buggy. An estimator based on the extended Kalman filter approach is used to recursively provide an optimal estimate of the position and orientation of the vehicle. Sensory data from the Differential Global Positioning System (DGPS), a gyroscope and an odometer are fused in the algorithm. As measurement data from the DGPS occasionally contains spikes, an effective and practical method is described to remove such data, preventing it from corrupting the estimation. A technique based on adjustable process noise levels is also employed in our work to maintain the estimation consistency of the filter. Finally, the filter's performance is evaluated with different trajectories under different satellite conditions using true data obtained from field trials.