Meeting Abstract

59.6  Saturday, Jan. 5  Proper orthogonal decomposition of bat flight kinematics RISKIN, D. K.*; WILLIS, D. J.; HEDRICK, T. L.; IRIARTE-DIAZ, J.; LAIDLAW, D. J.; BREUER, K. S.; SWARTZ, S. M.; Brown University; Massachusetts Institute of Technology; University of North Carolina; Brown University; Brown University; Brown University; Brown University dkr8@brown.edu

It is not known how many body markers are needed for a model to accurately reproduce the complex movement of bat wings during flight, or where those markers should optimally be placed. Too few markers will result in underestimating dimensional complexity, while too many will increase the amount of time required to digitize marker positions, without adding information about complexity of movement. Poor selection of marker locations can have either (or both) of these effects. Proper Orthogonal Decomposition (POD) is a computational tool mathematically equivalent to Principal Components Analysis, that can be applied to quantify the dimensional complexity of motion. We applied POD to a series of markers on a bat (Pteropodidae: Cynopterus brachyotis) flying in a wind tunnel in 9 trials, each at a different speed. In each trial we independently measured 3D motion of 15 wing and leg markers on the left side of the body, and body pitch (46 dimensions total). We found that regardless of speed, >30% of flight motion could be reconstructed using a single linear variable, and >95% constructed using 16 of the 46 variables. For each number of wing markers (1 to 15) we found the ‘optimal’ set, defined as that which resulted in the greatest dimensional complexity for a given number of wing markers. We found that the hindlimb moves independently from the rest of the wing. Also, the motions of different parts of digits III and IV are quite independent, so where capturing dimensional complexity is the goal, several markers should be placed on those digits.