Indoor localization and position monitoring are required to provide ambient-assisted life apps and
systems.
Six-degrees-of-freedom (6 DOF) estimation plays a key role in aerial robots’ navigation and control.
Different sensor data is usually fused in outdoors environments (e.g. GPS and INS); however, indoors or
in GPS-denied environments, such an approach is unfeasible.
Infinium Robotics is developing an inexpensive alternative solution based on computer vision. The in-
house developed algorithms indicate that accuracies better than 5 centimeters in the position
estimation (i.e. x, y, and z coordinates) and 2 degrees in the attitude estimation (i.e. roll, pitch, and yaw
angles) are feasible at update frequencies of 30 Hz with relatively low latency. Current research is
focused to improve the robustness of such algorithms.
