Indoor Navigation System and Suspended Load Control for Multirotors

Abstract

This thesis considers a high precision indoor navigation system and a suspended load control for multirotor systems. A complete overview of a multirotor system with all the necessary hardware and software is presented. A navigation system that is low-cost, light weight, easy to deploy, and independent of satellite coverage is developed. The solution is based on Extended Kalman Filter (EKF) with sensor fusion approach to filtering using all available on-board sensory data, and uses the principle of trilateration utilizing range measurements from radio transceivers. The presented EKF algorithm is capable of outlier detection and rejection, and solves the divergence problem with adaptively weighted covariance based on Fuzzy Logic Adaptive System (FLAS). The suspended load problem is solved with two separate controllers that are independent from the multirotor autopilot. The so-called input shaping filter is used to avoid initial swing excitation in the transition between hover and forward flight, while a time-delayed feedback control damps remaining oscillations on the load based on load angle measurements. The results are verified by simulations and in-flight experiments.