Single and multiple glider path planning using an optimization-based approach

2021 IEEE International Conference on Robotics and Automation (ICRA)

Pizarro

et al. 2021

Systems of multiple low-cost, underactuated floats combined with fully actuated surface vessels can improve the scalability and cost-effectiveness of autonomous systems for marine science and environmental monitoring. Here, we consider a coordination problem where surface vessels must drop off floats at locations such that they are likely to drift to observe given points of interest, and later must pick up the floats for redeployment. We define the Multi-Vessel Multi-Float (MVMF) problem and present a hierarchical solution based on the Dec-MCTS algorithm. Our solution defines customised sampling, rollout, and action generation algorithms to accommodate the problem's large search space and provide computational performance sufficient for practical application. We report analytical and simulation results that demonstrate the computational efficiency of our method and validate its behaviour in practical problems. These results immediately enable field experiments to progress the development of this exciting concept in multi-robot marine systems.

Section: Problem Formulationmentioning

confidence: 99%

Hierarchical MCTS for Scalable Multi-Vessel Multi-Float Systems

D'Urso

2021 IEEE International Conference on Robotics and Automation (ICRA)

Pizarro

et al. 2021

“…There has been other work in finding discrete time solutions using a time-expanded graph. Dijkstra [7] and A* [8,9,13] approaches to this problem have appeared. A recent improvement to the A* method has also been presented using adaptive sampling [16].…”

Section: Related Workmentioning

confidence: 99%

Hierarchical Planning in Time-Dependent Flow Fields for Marine Robots

2020 IEEE International Conference on Robotics and Automation (ICRA)

Yoo

Anstee

2020

We present an efficient approach for finding shortest paths in flow fields that vary as a sequence of flow predictions over time. This approach is applicable to motion planning for slow marine robots that are subject to dynamic ocean currents. Although the problem is NP-hard in general form, we incorporate recent results from the theory of finding shortest paths in time-dependent graphs to construct a polynomial-time algorithm that finds continuous trajectories in time-dependent flow fields. The algorithm has a hierarchical structure where a graph is constructed with time-varying edge costs that are derived from sets of continuous trajectories in the underlying flow field. We show that the continuous algorithm retains the time complexity and path quality properties of the discrete graph solution, and demonstrate its application to surface and underwater vehicles including a traversal along the East Australian Current with an autonomous marine vehicle. Results show that the algorithm performs efficiently in practice and can find paths that adapt to changing ocean currents. These results are significant to marine robotics because they allow for efficient use of time-varying ocean predictions for motion planning.

“…Aside from our previous work, studies on underwater glider path planning in 3D flow fields have been surprisingly rare. A large portion of existing work considers a 2D oceanic flow either using surface currents [24,27] or implicitly using a depth-average current [21,28,29]. The few that do plan in 3D workspace either model the glider with simple kinematics and a directly controllable turning rate [7], or do not account for flow fields [30,31].…”

Section: Related Workmentioning

confidence: 99%

Streamline-Based Control of Underwater Gliders in 3D Environments

2019 IEEE 58th Conference on Decision and Control (CDC)

Yoo

et al. 2019

Autonomous underwater gliders use buoyancy control to achieve forward propulsion via a sawtooth-like, riseand-fall trajectory. Because gliders are slow-moving relative to ocean currents, glider control must consider the effect of oceanic flows. In previous work, we proposed a method to control underwater vehicles in the (horizontal) plane by describing such oceanic flows in terms of streamlines, which are the level sets of stream functions. However, the general analytical form of streamlines in 3D is unknown. In this paper, we show how streamline control can be used in 3D environments by assuming a 2.5D model of ocean currents. We provide an efficient algorithm that acts as a steering function for a single rise or dive component of the glider's sawtooth trajectory, integrate this algorithm within a sampling-based motion planning framework to support long-distance path planning, and provide several examples in simulation in comparison with a baseline method. The key to our method's computational efficiency is an elegant dimensionality reduction to a 1D control region. Streamlinebased control can be integrated within various sampling-based frameworks and allows for online planning for gliders in complicated oceanic flows.