An IMU-based system identification technique for quadrotors

Abstract

Accurate system identification has an important role in designing high-performance controllers for robotic platforms. Most advanced controllers are designed based on the plants' dynamic models and their identified parameters. However, some models and their parameters change because of payload or condition variations. Online identification of these changes leads to develop high performance adaptive controllers. In this research, a rigorous amount of study is carried out to identify the parametric variations of a multi-rotor using a sensory system. The proposed method applies an inertial measurement unit (IMU) to identify the exact center of mass and its displacements. This method, which is based on the genetic algorithm (GA), is independent of the IMU location. The proposed algorithm is theoretically investigated while experimental results are presented to validate the performance of the algorithm.