多普勒计程仪（DVL）在水下导航系统应用越来越广泛。当海底环境发生变化时，DVL会发生数据刷新频率不稳定，数据无效等情况。为了提高导航的可靠性，本文提出了一种变训练集的SVR回归方法，对DVL的数据进行预测。根据水下机器人的速度变化率和加速度变化率调节训练集大小。把捷联惯导（SINS）的东向和北向速度作为输入，DVL东向和北向速度作为输出对模型进行训练。根据SINS的数据输出频率，选取合适的预测数据输出频率进行仿真。仿真发现算法有效地提高了SINS/DVL组合导航的精度，并在DVL数据无效时，有效地抑制误差，提高导航系统的稳定性。
DVL is more and more widely used in underwater navigation system. When the seabed environment changes, the data refresh frequency of DVL is unstable and the data is invalid. In order to improve the reliability of navigation, this paper proposes a SVR regression method of variable training set to predict DVL data. The training set size is adjusted according to the change of velocity rate and acceleration rate of underwater robot. Taking the east and north velocity of SINS as input, DVL's east and north velocity as output, the model is trained. We select the appropriate predictive data output frequency that relys on the data out frequency of SINS. It is found that SINS/DVL integrated navigation precision is effectively improved through simulation. When the DVL data is invalid, the error is effectively suppressed and the stability of the navigation system is improved.
2020,42(1): 161-167 收稿日期：2018-10-09
DOI：10.3404/j.issn.1672-7649.2020.01.033
分类号：U666.11
基金项目：中国科学院装备预研联合基金资助项目(6141A01061601)
作者简介：魏奥博(1994-),男,硕士研究生,研究领域为水下机器人导航控制
参考文献：
[1] 尹伟伟, 郭士荦. 非卫星水下导航定位技术综述[J]. 舰船电子工程, 2017, 37(3):8-11
[2] SHABANI, M; GHOLAMI, A. Improved underwater integrated navigation system using unscented filtering approach[J]. Navig., 2016, 69:561-581
[3] LARSEN M B. High performance doppler-inertial navigation-experimental result[C]//Ocean 2000 MITS/IEEE Conference and Exhibition. RI, USA:IEEE, 2000.
[4] 赵俊波, 葛锡云. 水下SINS/DVL组合导航技术综述[J]. 水下无人系统学报, 2018, 26(1):2-9
[5] SEMENIUK, L; NOURELDIN, A. Bridging GPS outages using neural network estimates of INS position and velocity errors[J]. Meas. Sci. Technol., 2006, 17:2783-2798
[6] HASAN, A. M.; SAMSUDIN, K.; RAMLI, A. R.; AZMIR, R.S Automatic estimation of inertial navigation system errors for global positioning system outage recovery[J]. Aerosp. Eng., 2011, 255:86-96
[7] ZHU Yixian, CHENG Xianghong, HU Jie, et al. A novel hybrid approach to deal with dvl malfunctions for underwater integrated navigation systems[J]. Applied Sciences., 2017, 7:759
[8] ZHAO L Y, LIU X J, WANG L, et al. A pretreatment method for the velocity of DVL based on the motion constraint for the integrated SINS/DVL[J]. Appl. Sci., 2016, 6:79
[9] 潘峰, 徐德民. 远程AUV组合导航滤波算法[D]. 西安:西北工业大学. 2006
[10] 周志华. 机器学习[M]. 北京:清华大学出版社. 2016.
