Dataset Information

Learning efficient navigation in vortical flow fields.

ABSTRACT: Efficient point-to-point navigation in the presence of a background flow field is important for robotic applications such as ocean surveying. In such applications, robots may only have knowledge of their immediate surroundings or be faced with time-varying currents, which limits the use of optimal control techniques. Here, we apply a recently introduced Reinforcement Learning algorithm to discover time-efficient navigation policies to steer a fixed-speed swimmer through unsteady two-dimensional flow fields. The algorithm entails inputting environmental cues into a deep neural network that determines the swimmer's actions, and deploying Remember and Forget Experience Replay. We find that the resulting swimmers successfully exploit the background flow to reach the target, but that this success depends on the sensed environmental cue. Surprisingly, a velocity sensing approach significantly outperformed a bio-mimetic vorticity sensing approach, and achieved a near 100% success rate in reaching the target locations while approaching the time-efficiency of optimal navigation trajectories.

SUBMITTER: Gunnarson P

PROVIDER: S-EPMC8654940 | biostudies-literature | 2021 Dec

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Learning efficient navigation in vortical flow fields.

Gunnarson Peter P Mandralis Ioannis I Novati Guido G Koumoutsakos Petros P Dabiri John O JO

Nature communications 20211208 1

Efficient point-to-point navigation in the presence of a background flow field is important for robotic applications such as ocean surveying. In such applications, robots may only have knowledge of their immediate surroundings or be faced with time-varying currents, which limits the use of optimal control techniques. Here, we apply a recently introduced Reinforcement Learning algorithm to discover time-efficient navigation policies to steer a fixed-speed swimmer through unsteady two-dimensional ...[more]

PMID: 34880221

Similar Datasets

Project description:IntroductionPhase contrast MRI allows for the examination of complex hemodynamics in the heart and adjacent great vessels. Vortex flow patterns seem to play an important role in certain vascular pathologies. We propose two- and three-dimensional metrics for the objective quantification of aortic vortex blood flow in 4D phase contrast MRI.Materials and methodsFor two-dimensional vorticity assessment, a standardized set of 6 regions-of-interest (ROIs) was defined throughout the course of the aorta. For each ROI, a heatmap of time-resolved vorticity values [Formula: see text] was computed. Evolution of minimum, maximum, and average values as well as opposing rotational flow components were analyzed. For three-dimensional analysis, vortex core detection was implemented combining the predictor-corrector method with λ2 correction. Strength, elongation, and radial expansion of the detected vortex core were recorded over time. All methods were applied to 4D flow MRI datasets of 9 healthy subjects, 2 patients with mildly dilated aorta, and 1 patient with aortic aneurysm.ResultsVorticity quantification in the 6 standardized ROIs enabled the description of physiological vortex flow in the healthy aorta. Helical flow developed early in the ascending aorta (absolute vorticity = 166.4±86.4 s-1 at 12% of cardiac cycle) followed by maximum values in mid-systole in the aortic arch (240.1±45.2 s-1 at 16%). Strength, elongation, and radial expansion of 3D vortex cores escalated in early systole, reaching a peak in mid systole (strength = 241.2±30.7 s-1 at 17%, elongation = 65.1±34.6 mm at 18%, expansion = 80.1±48.8 mm2 at 20%), before all three parameters similarly decreased to overall low values in diastole. Flow patterns were considerably altered in patient data: Vortex flow developed late in mid/end-systole close to the aortic bulb and no physiological helix was found in the aortic arch.ConclusionsWe have introduced objective measures for quantification of vortical flow in 4D phase contrast MRI. Vortex blood flow in the thoracic aorta could be consistently described in all healthy volunteers. In patient data, pathologically altered vortex flow was observed.

Dataset Information

Learning efficient navigation in vortical flow fields.

Publications

Learning efficient navigation in vortical flow fields.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets