Robophysics-inspired experiments in canines to discover strategies for quadrupedal negotiation of regular obstacles

Just as four legged animals have provided inspiration for robotic technologies, robot experiments can inspire new questions in animal locomotion. This study continues research into how dogs handle a regular series of obstacles, inspired by past work that found that quadrupedal robots can have stable trajectories that depend systematically on the gait that they use when moving over obstacle arrays. Subsequently we examined dogs as they negotiated similar arrays of varied spacing. We found that the dogs did not display large changes in average gait, but did exhibit stride length that varied in proportion to obstacle spacing. Dogs appear to have further strategies to adapt to the obstacles, however, that we explore here: 1) they can accumulate error and then perform a "reset" step to correct themselves; and 2) the animals can zig-zag across the obstacle array, presumably to utilize a more "comfortable" stride length. We explore the parameters that may predict a "reset" step, including the canines' stride lengths and limb phases, and we further look for morphological or other properties of the animal that may predict zig-zag behavior. This study may shed light on behaviors for agile, stable, and/or economical quadrupedal locomotion over more varied terrain.