Optimization of ADCP and GPS position and velocity measurements using a Kalman filter

Abstract

This study optimized and evaluated a Kalman filter to improve measurements of boat velocity, boat position, apparent bedload velocity and water discharge using inputs of Stand Alone Global Positioning System (GPS) VTG (Doppler velocity) and GGA (triangulated position) and Acoustic Doppler Current Profiler (aDcp) bottom tracking (Doppler sonar). The filter significantly (alpha<0.05) improved mean velocity errors compared to Stand Alone GGA. Kalman filter velocity errors were lower than Stand Alone VTG velocity errors, but differences were rarely significant. Improvements were seen in mean position errors, although for low bedload transects, care was needed when using the Kalman velocity filter. The Kalman filter usually showed the lowest mean bedload biases, although biases were all very close ranging from -1.7 cm/s to 1.25 cm/s. The filter did not always show the lowest errors for discharge although it did show the least variability compared to each of the input signals.