Meeting Abstract
Work on a variety of organisms suggests that at moderate to high speeds animals can select their gaits for many goals, optimizing the cost of transport, speed, or chances of avoiding injury. Here we examine the motion of the nematode Caenorhabditis elegans using a newly developed geometric gait optimization tool. For animals operating in friction dominated regimes, such as viscous swimming, the locomotion is governed by a “connection” as used in the theory of geometric mechanics. By combining tools of kinematic phase analysis with those of geometric mechanics, our gait optimization tool models the connection governing a gait, allowing the computer to climb the gradient of goal functions such as the cost of transport. Using swimming data kindly provided by the Penn Complex Fluids Lab, we validated a Purcell swimming model and discovered that its moment equation contributed little (< 1%) to the motion per cycle. Using a simplified model absent the moment equation, we optimized gaits for extremal motion (maximum displacement per cycle) and for cost of transport, using the animals’ observed motion as the initial condition. We classified the resulting gaits by the mean time derivative of absolute curvature, and by the mean power. This gave 4.2+/-.27 [rad/mms] and 74.+/-12. [fW] for the animals, falling close to the 5.6 [rad/mms] and 15. [fW] of the optimal cost of transport gait, and further away from the 11. [rad/mms] and 270. [fW] of the extremal gait. We conclude that these nematodes motions are consistent with cost of transport having a significant weight in their choice of gait. Although unused here, our tool is model free, and can model the connection from motion captured gait data. This work may further illuminate how animals select their locomotor patterns and allow hypotheses of optimality to be rapidly tested.
