A Novel ANFIS-AQPSO-GA-Based Online Correction Measurement Method for Cooperative Localization
Bo Xu, Shengxin Li, Asghar A. Razzaqi, Lianzhao Wang, Mingyu Jiao
Abstract
To solve the effect of unreliable acoustic range measurements in cooperative localization of autonomous underwater vehicles, a novel localization method is proposed by using a hybrid metaheuristic algorithm that consists of the adaptive quantum-behaved particle swarm optimization (AQPSO) algorithm combined with the genetic algorithm (GA) to train the adaptive neurofuzzy inference system (ANFIS) model. The method can predict the acoustic ranging errors online, then correct the measurements, and improve the accuracy and stability of the cooperative localization system. We first tested the optimization ability of AQPSO-GA with an open-source dataset, and then, a large number of simulation experiments are performed with actual data collected in lake-water trials to fully compare and analyze the performance of various cooperative localization methods. The experimental results show that, compared with the cooperative localization method based on the cubature Kalman filter, the proposed method can reduce the localization error by 90% and effectively guarantee the stability of the cooperative localization system.