Trajectory Planning for Autonomous Underwater Vehicles in the Presence of Obstacles and a Nonlinear Flow Field using Mixed Integer Nonlinear Programming
KAUST DepartmentApplied Mathematics and Computational Science Program
Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division
Online Publication Date2018-08-24
Print Publication Date2019-01
Permanent link to this recordhttp://hdl.handle.net/10754/628492
MetadataShow full item record
AbstractThis paper addresses the time-optimal trajectory planning for autonomous underwater vehicles. A detailed mixed integer nonlinear programming (MINLP) model is presented, explicitly taking into account vehicle kinematic constraints, obstacle avoidance, and a nonlinear flow field to represent the ocean current. MINLP problems pose great challenges because of the combinatorial complexity and nonconvexities introduced by the nature of the flow field. A novel solution approach in an optimization framework is developed to address associated difficulties. The main benefit of the proposed methodology is the ability to find multiple local minima. The contribution of the paper is fourfold: 1) a novel approach to integrate the flow field into the MINLP model; 2) a diversified initialization strategy using multiple waypoints, different solvers and approximated models, namely, a mixed integer linear programming model and the MINLP model with and without the flow field; 3) an algorithm that forces the solver to seek improved solutions; and 4) a parallel computing approach capitalizing on diversified initialization. The performance of the resulting methodology is illustrated on idealized case studies, and the results are used to gain insight into trajectory planning in the presence of flow fields.
CitationWang T, Lima RM, Giraldi L, Knio OM (2018) Trajectory Planning for Autonomous Underwater Vehicles in the Presence of Obstacles and a Nonlinear Flow Field using Mixed Integer Nonlinear Programming. Computers & Operations Research. Available: http://dx.doi.org/10.1016/j.cor.2018.08.008.
SponsorsResearch reported in this publication was supported by research funding from King Abdullah University of Science and Technology (KAUST).
JournalComputers & Operations Research