Smart Inertial Particles
arXiv:1711.05853 · doi:10.1103/PhysRevFluids.3.084301
Abstract
We performed a numerical study to train smart inertial particles to target specific flow regions with high vorticity through the use of reinforcement learning algorithms. The particles are able to actively change their size to modify their inertia and density. In short, using local measurements of the flow vorticity, the smart particle explores the interplay between its choices of size and its dynamical behaviour in the flow environment. This allows it to accumulate experience and learn approximately optimal strategies of how to modulate its size in order to reach the target high-vorticity regions. We consider flows with different complexities: a two-dimensional stationary Taylor-Green like configuration, a two-dimensional time-dependent flow, and finally a three-dimensional flow given by the stationary Arnold-Beltrami-Childress helical flow. We show that smart particles are able to learn how to reach extremely intense vortical structures in all the tackled cases.
arXiv commentPublished on Phys. Rev. Fluids (August 6, 2018)
References in corpus (12)
- Heavy particle concentration in turbulence at dissipative and inertial scales
- Flow Navigation by Smart Microswimmers via Reinforcement Learning
- Dynamics of a small neutrally buoyant sphere in a fluid and targeting in Hamiltonian systems
- Synchronised Swimming of Two Fish
- Learning to school in the presence of hydrodynamic interactions
- Turbulent pair dispersion of inertial particles
- Multifractal concentrations of inertial particles in smooth random flows
- Large scale inhomogeneity of inertial particles in turbulent flow
- Finding Efficient Swimming Strategies in a Three Dimensional Chaotic Flow by Reinforcement Learning
- An instrumented tracer for Lagrangian measurements in Rayleigh-Bénard convection
- Reinforcement Learning of Artificial Microswimmers
- Propelled micro-probes in turbulence
Cited by in corpus (19)
- Machine Learning for Fluid Mechanics
- Zermelo's problem: Optimal point-to-point navigation in 2D turbulent flows using Reinforcement Learning
- Controlling Rayleigh-Bénard convection via Reinforcement Learning
- Hydrodynamics can Determine the Optimal Route for Microswimmer Navigation
- Learning to Control Active Matter
- Navigation of micro-swimmers in steady flow: the importance of symmetries
- Optimal active particle navigation meets machine learning
- Swimming strategy of settling elongated microswimmers by reinforcement learning
- Long-distance migration with minimal energy consumption in a thermal turbulent environment
- Learning how to find targets in the micro-world: The case of intermittent active Brownian particles
- The statistical properties of turbulence in the presence of a smart small-scale control
- Learning to swim efficiently in a nonuniform flow field
- Migration of self-propelling agent in a turbulent environment with minimal energy consumption
- Stabilising viscous extensional flows using Reinforcement Learning
- A critical assessment of reinforcement learning methods for microswimmer navigation in complex flows
- Adaptive active Brownian particles searching for targets of unknown positions
- Reinforcement Learning for Active Matter
- Run-and-Tumble Particles Learning Chemotaxis
- A study of first-passage time minimization via Q-learning in heated gridworlds