This is a repository of code and documentation from students projects in the STR lab at the Australian National University
2021 Angela Zhang, "Application of an Equivariant Filter on Ardupilot for State Estimation" Abstract: This thesis concerns the implementation of an Equivariant Filter (EqF) design on a real-world ap-plication, specifcally focusing on the ArduPilot platform. The state estimation problem considered isthe UAV system posed on homogeneous spaces with Lie group symmetry. Measurements from IMU,GPS, magnetometer and barometer are fused for estimating the UAV attitude, position, velocity,gyroscope and accelerometer bias. EqF performance is validated using the simulation logs generatedfrom ArduPilot’s simulation software, and also on a real quadrotor fight log. Compared to the currentExtended Kalman Filter (EKF) used in ArduPilot, EqF has shown a broadly similar performance inposition and velocity estimates with a weaker performance in attitude under the same simulation.Further investigation is required for EqF performance analysis, and additional features such as theincorporation of time-delays and estimating magnetic feld states should be considered in the nextproject iteration.