Research On Multi-AUVs Formation Cooperative Location Method Under Abnormal Measurement

With the deepening of ocean exploration and development,single autonomous underwater vehicle(AUV)can not perform well in complex tasks.Cooperative operation has been widely recognized after considering the factors such as cost and quality.Multiple AUV cooperative localization is the basis and guarantee for high-quality cooperative tasks.However,the underwater environment is complex and changeable,and the occurrence of various abnormal situations is often unpredictable,which will inevitably have a negative impact on the positioning accuracy of the system and the execution of the task.This paper mainly focuses on how to weaken the influence of abnormal measurement on positioning accuracy.The main work of the paper is described as follows:1、The paper introduces the background of research and application of AUV and the research status of key issues in cooperative localization at home and abroad.Then,the related knowledge of sensors is introduced,including the error characteristics of sensors.Based on the theory of Extended Kalman Filter(EKF)and Cubature Kalman Filter(CKF),the positioning effects of single-pilot mode and multi-pilot mode are discussed,and the theoretical analysis is verified by simulation.2、The improved cooperative localization algorithm under the influence of measurement noise outliers is studied.The Maximum Correntropy(MC)theory was introduced into the filtering process to solve the outliers of the measured noise in the localization process,and the model of Maximum Correntropy EKF fusion algorithm(MCEKF)and Maximum Correntropy CKF fusion algorithm(MCCKF)were deduced and constructed respectively.Since the gaussian kernel width is introduced into the fusion algorithm,the influence of gaussian kernel width on the filtering results of the fusion algorithm is further explored.3、The improved cooperative localization algorithm for data loss and communication delay is studied.For the problem of measurement loss,the effects of improved EKF algorithm and Moving Horizon Estimation(MHE)method are simply verified.The estimation error is reduced by jumping the estimation process of the lost time or utilizing multiple measurements in the rolling window.Aiming at the problem of measurement information delay,an smoothing delay filtering algorithm is proposed based on EKF and DDF2 inspired by the moving horizon,which ensure positioning accuracy and improve system stability,thus weakening the influence of abnormal measurement environment on the positioning process.Moreover,the measurement information is fully utilized in the smooth process,which meets the requirements of high precision for post-measurement.4、A new method of formation topology optimization design is proposed.The evaluation function of the observable performance is constructed based on Cramér-Rao lower bound(CRLB)and Fisher information matrix(FIM)considering the factors such as the situation of multiple leaders and multiple followers(ML-MF)and the correlation between acoustic ranging error and distance.In view of the uncertainty of the location information,the Monte Carlo method was applied to complete the reconstruction of the evaluation function and the formation configuration in the possible distribution region is optimised.Finally,A step recursive strategy based on annealing idea is proposed to solve the constraint equation for ML-MF system.The feasibility and effectiveness of the algorithm was confirmed by the simulation results.