State estimation for autonomous surface vehicles based on Echo state networks

Zhouhua Peng, Jun Wang, Dan Wang

Результат исследований: Глава в книге, отчете, сборнике статейМатериалы для конференциирецензирование


This paper investigates the state estimation for autonomous surface vehicles in the presence of unknown dynamics and unmeasured states. The unknown dynamics comes from parametric model uncertainty, unmodelled hydrodynamics, and external disturbances caused by wind, waves and ocean currents. A nonlinear adaptive observer is proposed based on echo state networks, which are used to approximate the unknown dynamics using input-output data. By using the proposed observer, the unmeasured states and unknown dynamics can be simultaneously estimated in real time. The stability of the observer is analyzed via Lyapunov analysis. The proposed observer can be used in various motion control scenario, such as target tracking, trajectory tracking, path following, formation control, and even sideslip angle identification, not only for fully-actuated marine vehicles but also for under-actuated marine vehicles.

Язык оригиналаАнглийский
Название основной публикацииAdvances in Neural Networks - ISNN 2017 - 14th International Symposium, ISNN 2017, Proceedings
РедакторыAndrew Leung, Fengyu Cong, Qinglai Wei
ИздательSpringer Verlag
Число страниц8
ISBN (печатное издание)9783319590714
СостояниеОпубликовано - 2017
Опубликовано для внешнего пользованияДа
Событие14th International Symposium on Neural Networks, ISNN 2017 - Sapporo, Hakodate, and Muroran, Hokkaido, Япония
Продолжительность: 21 июн. 201726 июн. 2017

Серия публикаций

НазваниеLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Том10261 LNCS
ISSN (печатное издание)0302-9743
ISSN (электронное издание)1611-3349


Конференция14th International Symposium on Neural Networks, ISNN 2017
ГородSapporo, Hakodate, and Muroran, Hokkaido


Подробные сведения о темах исследования «State estimation for autonomous surface vehicles based on Echo state networks». Вместе они формируют уникальный семантический отпечаток (fingerprint).