Action-Perception Trade-Offs for Anguilliform Swimming Robotic Platforms with an Electric Sense

The work presented addresses the combination of anguilliform swimming-based propulsion with the use of an electric sensing modality for a class of unmanned underwater vehicles, and in particular investigates the relative influence of adjustments to the swimming gait on the platform's displacement speed and on sensing performance. This influence is quantified, for a relevant range of swimming gaits, using experimental data recordings of displacement speeds, and a boundary element method-based numerical simulation tool allowing to reconstruct electric measures. Results show that swimming gaits providing greater movement speeds tend to degrade sensing performance. Conversely, gaits yielding accurate sensing tend to prove slower. To reconcile opposing tendencies, a simple action-perception cost function is designed, with the purpose of adjusting an anguilliform swimmer's gait shape, in accordance with respective importance afforded to action (i.e. movement speed) and perception.