Multi-UAV wireless positioning using adaptive multidimensional scaling and extended Kalman filter

Abstract

Global Navigation Satellite System (GNSS) signal can be blocked when flight vehicles operate in challenging environments such as indoor or adversarial environments. While multi-UAVs are teamed during flight, cooperative localization becomes available to tackle this challenge. Multidimensional Scaling (MDS) method has been well studied for cooperative localization of Wireless Sensor Network (WSN) based on radio frequency (RF) measurement. When noise RF measurement model is lacking, conventional weighted MDS method represents confidence with the measurements by assigning weights relying on distance information between each pair of nodes. In order to process non-distance RF measurements, we present an improved weighted MDS method which applies a novel weighting scheme. In this article, the proposed method conducts velocity estimation for multi-UAV system based on odometry and Frequency Difference of Arrival (FDOA) measurements. Furthermore, an extended Kalman Filter (EKF) algorithm is applied to refine the initial estimation of the MDS method and derive position estimation. Finally, numerical experiments demonstrate the robustness and accuracy of the adaptive MDS-EKF refinement framework for multi-UAV system localization in an unknown dynamic environment lacking measurement noise information.