ADDR-GPS data fusion using Kalman filter algorithm

Low speed reconnaissance unmanned air vehicle uses low cost air data dead reckoning (ADDR) as the primary navigation system. In order to keep the ADDR position error (due to heading and air speed sensor calibration/installation error and computation error) to a minimum value, Global Position System (GPS) is used in supervisory mode to update the ADDR position at regular intervals. This thesis deals with a new approach based on Kalman filtering for navigation sensor data fusion obtained from ADDR navigation system and GPS. The modeling of ADDR and GPS dynamics has been developed, simulated for its error sources and position accuracies to determine the covariance matrix. A feed forward Kalman filter based navigation sensor data fusion algorithm has been developed to fuse the information from ADDR and GPS to obtain the best position estimate. This algorithm provides a relatively accurate low cost navigation system for UAV applications.