Distributed model predictive control for vessel train formations of cooperative multi-vessel systems. (July 2018)

• Record Type: Journal Article
• Authors: Chen, Linying; Hopman, Hans; Negenborn, Rudy R.
• Is Part Of: Transportation research. Volume 92(2018)
• Page Start: 101
• Page End: 118
• Publication Date: 2018-07
• Subjects: Cooperative multi-vessel systems -- Vessel train formation -- Distributed model predictive control -- Autonomous vessels -- ADMM; Transportation -- Periodicals; Transportation -- Technological innovations -- Periodicals; 388.011
• DOI: 10.1016/j.trc.2018.04.013
• Languages: English
• ISSNs: 0968-090X

Abstract

Highlights: Proposal of the concept of a Cooperative Multi-Vessel System. Proposal of a serial iterative algorithm for the vessel train formation problem. Evaluation of the effects of varying updating sequence and responsibility parameter. Discussion of scalability of the proposed algorithm. Illustration of how the algorithm achieves a train-like formation of vessels. Abstract: Recently, the cooperative control of multiple vessels has been gaining increasing attention because of the potential robustness, reliability and efficiency of multi-agent systems. In this paper, we propose the concept of Cooperative Multi-Vessel Systems (CMVSs) consisting of multiple coordinated autonomous vessels. We in particular focus on the so-called Vessel Train Formation (VTF) problem. The VTF problem considers not only cooperative collision avoidance, but also grouping of vessels. An MPC-based approach is proposed for addressing the VTF problem. A centralized and a distributed formulation based on the Alternating Direction of Multipliers Method (ADMM) are investigated. The distributed formulation adopts a single-layer serial iterative architecture, which gains the benefits of reduced communication requirements and robustness against failures. The impacts of information updating sequences and responsibility parameters are discussed. We furthermore analyze the scalability of the proposed method. Simulation experiments of a CMVS navigating from different terminals in the Port of Rotterdam to inland waterways are carried out to illustrate the effectiveness of our method. The proposed method successfully steers the vessels from different origins to form a vessel train. Due to the effective communication, vessels can timely respond to the velocity changes that others make. After the formation is formed, the distances between vessels become constant. The results show the potential to use CMVSs for inland shipping with enhanced safety.