Mobile sensor path planning for iceberg monitoring using a MILP framework

We look at the task of iceberg monitoring using a single mobile sensor, and we suggest a modular framework for this. The focus is on path planning for which we come up with a novel strategy, which includes solving a static optimization problem often to account for changes. We formulate the optimization problem in a MILP framework, and we illustrate how this yields acceptable computational time for problem size of about 15 icebergs. We also suggest a tuning rule for weighting between different objectives in the optimization formulation, which we demonstrate in simulations. Initializing the optimization with the previous solution can improve computational time dramatically. Finally, we discuss how we easily can add extra features to our framework.