摘要
针对多自主水下航行器(AUVs)在恶劣水下环境通信数据包丢失的情况,提出了一种基于自适应神经模糊推理系统(ANFIS)的多AUV协同定位方法。将传统的扩展卡尔曼滤波(EKF)与ANFIS相结合。在协同定位系统正常通信时,跟随AUV的位置由EKF估计得到,并利用该阶段收集的数据对ANFIS模型进行训练;当通信包丢失时,ANFIS进入预测过程,ANFIS根据输入数据预测跟随AUV的位置。利用湖试数据验证了所提方法的有效性。实验结果表明,当通信包丢失时,ANFIS能够根据输入数据有效地预测AUV的位置。在通信包丢失时,所提方法平均定位误差与无ANFIS辅助的EKF滤波比较,平均定位误差减小78%,均方根误差减少77%,具有更好的准确性和稳定性。
A new cooperative localization method for multiple autonomous underwater vehicles(AUV)is proposed to deal with the problem of communication packet loss caused by harsh underwater environment.Traditional extended Kalman filter(EKF)is combined with adaptive neuro-fuzzy inference system(ANFIS),in which the position of follower AUV is estimated by EKF when the cooperative localization system can communicate normally,and the data collected in this stage is used to train the ANFIS model.When the communication packet is lost,ANFIS enters the prediction process,and the ANFIS predicts the position of follower AUV according to the input data.The efficacy of the proposed method in the cooperative localization of AUVs is verified by lake trial.The experimental results show that ANFIS can effectively predict the position of AUV according to the input data when the communication packet is lost.Compared with the EKF filter without ANFIS,the average error and root mean square error of the proposed method are reduced by 78%and 77%respectively when the communication packet is lost,which has better accuracy and stability.
作者
徐博
李盛新
王连钊
段腾辉
姚贺
XU Bo;LI Shengxin;WANG Lianzhao;DUAN Tenghui;YAO He(College of Automation,Harbin Engineering University,Harbin 150001,China;Hunan Aerospace Institute of Electrical and mechanical equipment and special materials,Changsha 410205,China;CSSC Marine Technology Co.,LTD,Beijing 100071,China)
出处
《中国惯性技术学报》
EI
CSCD
北大核心
2019年第4期440-447,共8页
Journal of Chinese Inertial Technology
基金
国家自然科学基金（61203225）
黑龙江省自然科学基金（F2018009）
省博士后科研启动金（LBH-Q15032）
装备预研重点实验室基金（614221801050717）
国家海洋工程重点实验室公开项目（1616）
国家装备预研基金（61403110306）
关键词
自主水下航行器
协同定位
自适应神经模糊推理系统
扩展卡尔曼滤波
通信包丢失
autonomous underwater vehicle
cooperative localization
adaptive neuro-fuzzy inference system
extended Kalman filter
communication packet loss
