Efficient collective swimming by harnessing vortices through deep reinforcement learning.

Fish in schooling formations navigate complex flow-fields replete withmechanical energy in the vortex wakes of their companions. Their schoolingbehaviour has been associated with evolutionary advantages including collectiveenergy savings. How fish harvest energy from their complex fluid environmentand the underlying physical mechanisms governing energy-extraction duringcollective swimming, is still unknown. Here we show that fish can improve theirsustained propulsive efficiency by actively following, and judiciouslyintercepting, vortices in the wake of other swimmers. This swimming strategyleads to collective energy-savings and is revealed through the first evercombination of deep reinforcement learning with high-fidelity flow simulations.We find that a `smart-swimmer' can adapt its position and body deformation tosynchronise with the momentum of the oncoming vortices, improving its averageswimming-efficiency at no cost to the leader. The results show that fish mayharvest energy deposited in vortices produced by their peers, and support theconjecture that swimming in formation is energetically advantageous. Moreover,this study demonstrates that deep reinforcement learning can produce navigationalgorithms for complex flow-fields, with promising implications for energysavings in autonomous robotic swarms.