Real-time indoor localization support for four-rotor flying robots using sensor nodes

Abstract

Flying four-rotor robots (quadrocopters) are on-board sensor controlled systems. In comparison to classical monorotor objects (helicopters), quadrocopters can be piloted with a much lower effort. However, lateral drifts can not be compensated only referring to the built-in sensors. The detection of such drifts is strongly necessary for indoor operation - without corrections a quadrocopter would quickly cause a collision. In order to compensate the dislocation, an additional indoor positioning system is needed. In our work, we provide a framework for time-of-flight based localization systems relying on ultrasonic sensors. It is optimized for use in sensor nodes with low computational power and limited memory. Nevertheless, it offers scalability and high accuracy even in case of single erroneous measurements. We implemented the system in our lab using ultrasound sensors that are light enough to be carried around by the flying object. Using this real-time localization system, a position controller can be implemented to maintain a given position or course.