Multi-robot cooperative localization based on heuristically tuned extended H∞ filter

Aiming at the problems such as low accuracy of localization and weak consistency in the condition of unknown noise characteristics in multi-robot cooperative positioning system, a multi-robot consistent cooperative positioning approach based on the algorithm of Extended H-infinity Filter (EH∞F) is proposed. First, the kinematic model for multi-robot system is established and the adaptive EH∞F algorithm for cooperative localization of multi-robot system is formulated. Then, the consistency law of the EH∞F algorithm is analyzed and an improved scheme of consistency is proposed by expanding the covariance matrix. The proposed scheme of cooperative localization is evaluated by experiment of simulation. Simulation results show that the proposed method can effectively suppress the increasing of estimation error when the noise characteristics are unknown. Compared with the algorithm based on EKF, the positioning accuracy is improved by about 40% by the proposed EH∞F method. The proposed consistency improvement scheme can improve the consistency of multi-robot system positioning estimation by nearly 60% without reducing the positioning accuracy. © 2023 Editorial Department of Journal of Chinese Inertial Technology. All rights reserved.