Distributed Path Following of Multiple Under-Actuated Autonomous Surface Vehicles Based on Data-Driven Neural Predictors via Integral Concurrent Learning

This article addresses the problem of distributed path following of multiple under-actuated autonomous surface vehicles (ASVs) with completely unknown kinetic models. An integrated distributed guidance and learning control architecture is proposed for achieving a time-varying formation. Specifically...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transaction on neural networks and learning systems 2021-12, Vol.32 (12), p.5334-5344
Main Authors: Liu, Lu, Wang, Dan, Peng, Zhouhua, Han, Qing-Long
Format: Article
Language:English
Subjects:
Autonomous surface vehicles (ASVs)
Autonomous vehicles
Control theory
data-driven neural predictors
distributed path following
Guidance (motion)
integral concurrent learning
Kinematics
Kinetic theory
Learning
Observers
Path planning
Robustness
State observers
Surface vehicles
Task analysis
Trajectory planning
Trajectory tracking
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This article addresses the problem of distributed path following of multiple under-actuated autonomous surface vehicles (ASVs) with completely unknown kinetic models. An integrated distributed guidance and learning control architecture is proposed for achieving a time-varying formation. Specifically, a robust distributed guidance law at the kinematic level is developed based on a consensus approach, a path-following mechanism, and an extended state observer. At the kinetic level, a model-free kinetic control law based on data-driven neural predictors via integral concurrent learning is designed such that the kinetic model can be learned by using recorded data. The advantage of the proposed method is two-folds. First, the proposed formation controllers are able to achieve various time-varying formations without using the velocities of neighboring vehicles. Second, the proposed control law is model-free without any parameter information on kinetic models. Simulation results substantiate the effectiveness of the proposed robust distributed guidance and model-free control laws for multiple under-actuated ASVs with fully unknown kinetic models.
ISSN:2162-237X
2162-2388
DOI:10.1109/TNNLS.2021.3100147