Meeting Abstract

98-1  Saturday, Jan. 7 13:30 - 13:45  Dynamic traversal of large gaps and high bumps by cockroaches GART, SW*; LI, C; Johns Hopkins Univeristy swgart@jhu.edu http://li.me.jhu.edu

To survive, terrestrial insects must traverse obstacles like rocks, litter, vines, and exposed roots comparable in size to or even larger than themselves. Much has been known about how insects slowly negotiate obstacles and how lizards and birds run over small obstacles (up to half hip height). However, it is less clear how well insects can dynamically traverse large obstacles comparable to their size. Here, we challenged the discoid cockroach (Blaberus discoidalis) to run into two types of large obstacles, a gap (N = 7 animals, n = 350 trials) and a bump (N = 6, n = 270), and varied gap width from 0.2 to 1 body length (BL) and bump height from 1 to 4 hip height (H). The animal was able to dynamically traverse gaps as large as 1BL and bumps as high as 2H. For the gap, the probability of dynamic traversal (not falling into and then climbing out of the gap) decreased from 100 ± 0 % to 5 ± 1 % as gap width increased from 0.2BL to 1BL (P < 0.0001, repeated-measures ANOVA). Dynamic modeling well predicted observations of gap traversal, and demonstrated that the animal behaved like a forward-moving rigid body falling on one end and used its head to bridge large gaps. For the bump, probability of dynamic traversal (average forward speed does not fall below minimal running speed) decreased from 87 ± 2 % to 33 ± 4 % as bump height increased from H to 2H (P < 0.0001). The animal climbed over higher bumps (3H and 4H) at slower exploratory speeds. For both gaps and bumps, running at a higher speed (higher kinetic energy) and with a more erect posture (higher potential energy and larger body pitch) facilitated dynamic traversal. Our study provides inspirations for robots to use kinetic energy to dynamically traverse large gaps and high bumps common in terrains like landslides and building rubble, a scenario challenging for slow navigation using sensing and path-planning.