Hava-yer robot ekipleri için bağıl lokalizasyon ve koordinasyon

Recently, autonomous robot teams have been implemented broadly in many social and military applications such as firefighting, agriculture, search and rescue, mapping, target tracking, and docking. A mix of different types of ground robots and aerial vehicles can be employed in a robot team to accomplish tasks efficiently and robustly. Such heterogeneous systems show unparalleled benefits in complex tasks compared to teams composed of identical robot types. In a heterogeneous robot team, precise relative localization, i.e., estimating a robot's position with respect to its neighbor robots, plays a key role. We develop a relative localization system for air-ground robot teams where an aerial vehicle and multiple ground robots work in coordination to perform a reliable relative position estimation. The aerial vehicle is employed to detect special patterns on the ground robots by an onboard monocular camera, while the ground robots perform relative position estimation based on inter-robot distances acquired by ultrawideband sensors and the bearing and heading angles received from the aerial vehicle by communication. Thus, the aerial vehicle serves as an absolute frame provider for the entire team. Notably, each robot in the team uses onboard communication and computation capabilities solely without any need for an external localization infrastructure, making the team realizable in all conditions including GNSS-denied environments. We propose a multi-rate extended Kalman filter algorithm to handle different data rates of the sensor measurements. We carried out an extensive simulation study with a drone and five ground robots in a leader-first follower formation. Simulation results showed a successful estimation performance with an error rate of up to five centimeters in the relative position estimations in both axes. Keywords: Relative localization, Heterogeneous multi-robot systems, estimation algorithms, ultrawideband sensors