Session: 01-08-01 Digital Twin Applications to Offshore Systems

Paper Number: 103740

103740 - Modular Collision Avoidance Using Predictive Safety Filters 

The number of maritime systems being launched in the ocean is increasing every year, including the development of offshore wind-farms, underwater robotics for ocean condition monitoring, and autonomous ship transport. Many of these activities are safety-critical, making it essential to have a closed-loop control system that satisfies constraints arising from underlying physical limitations and safety aspects in a robust manner. However, this is often challenging to achieve for real world systems. For example, autonomous ships at sea have non-linear and uncertain dynamics, and are subject to numerous time-varying environmental disturbances such as waves, currents, and wind. There is increasing interest in using machine learning based approaches to adapt these systems to more complex scenarios, but there is currently no standard framework to guarantee safety and stability for such systems. Recently, predictive safety filters have emerged as a useful method for ensuring constraint satisfaction, even when unsafe control inputs are used. The safety filter approach leads to a modular separation of the problem, allowing the usage of arbitrary control policies in a task-agnostic way. In this work, we develop a predictive safety filter to ensure anti-grounding and collision avoidance for a surface vessel. The filter takes in a nominal input sequence from a potentially unsafe controller and solves an optimization problem to compute a minimal perturbation of the nominal control inputs, which adheres to both physical and safety-related constraints. To validate the approach, we implement the safety filter on a small prototype ferry, and perform simulations for several realistic scenarios with map data from Trondheim, Norway. It is demonstrated that the predictive safety filter is able to avoid collisions with static obstacles by running the experiments for a large number of initial conditions. The predictive safety filter approach is very flexible, and can be used to improve the robustness of a variety of offshore applications, e.g. wind turbine stabilization, autonomous vessels, and marine robotics. This is essential for the safe application of machine learning in these contexts, which has outstanding potential for bringing more autonomy and adaptability to offshore technology. 

Presenting Author: Haakon Robinson Norwegian University of Science and Technology

Presenting Author Biography: Haakon Robinson is a PhD candidate at the Norwegian University of Science and Technology (NTNU). His work studies the intersection between modern machine learning and established methods within modelling and control, with a focus on methods to ensure or verify the safety of these hybrid models.

Authors:

Aksel Vaaler Norwegian University of Science and Technology
Haakon Robinson Norwegian University of Science and Technology
Trym Tengesdal Norwegian University of Science and Technology
Adil Rasheed Norwegian University of Science and Technology