Small UAV’s position and attitude estimation using tightly coupled multi baseline multi constellation GNSS and inertial sensor fusion

Abstract

The evolution of small UAV based photogrammetric and LiDAR survey systems significantly ease the geographical data acquisition process in mapping applications. However, the accuracy of the onboard position and attitude determination systems still limits the penetration of these low cost vehicles in real-time mapping tasks, such as the monitoring of floods or other environmental disasters. This paper focuses on the improvement of position and attitude determination of low cost UAVs by introducing a rigorous attitude and position computation algorithm using tightly coupled sensor fusion for multi antenna, multi GNSS and inertial sensor observations. The algorithm utilizes an Extended Kalman Filter (EKF) and in its current phase a post-processed kinematic (PPK) positioning solution. The developed algorithm is validated in a real case study with a UAV platform containing two low- cost GNSS receivers, a PIXHAWK onboard flight controller computer using several INS sensors and a Sony ILCE-6000 camera for photogrammetric data collection. The positioning solution is aided with a low-cost ground based GNSS base station. The developed algorithm estimates the position and the attitude of the platform by fusing accelerometer and gyroscope observations with GNSS code, carrier-phase and Doppler observations. The integer ambiguities are resolved by the LAMBDA method for the position and a newly developed quaternion constrained modified LAMBDA method for the UAV’s moving baseline. The attitude estimations are compared with the estimations of the onboard flight controller system and both of them are validated using post-processed attitude information obtained from photogrammetric data processing (PGP).The proposed method brought an improvement of 30% in terms of root mean square error for the yaw angle, while it provided 92.91% and 88.18% success rates in integer ambiguity resolution for the two baselines, respectively.