UCLA

Unlike two, four, six, and eight legged animals, Myriapoda--i.e., centipedes and millipedes--have been largely overlooked in the computer graphics literature. This thesis presents an artificial life (A-Life) framework for animating the locomotive behavior of Myriapoda with compelling physical and biological realism. Our system produces real-time animation and the creatures are autonomous, requiring minimal animator intervention. Creatures of various body morphologies can be animated simultaneously, and they are capable of complex multi-legged locomotion as well as anguilliform swimming.

Taking an Artificial Life approach, we develop a hybrid animation system that combines kinematic and dynamic simulation to animate a novel biomechanics model specifically tailored to the unique body structure of myriapoda. In our simulated myriapoda, the characteristically vivid leg wave patterns of their biological counterparts result as an emergent behavior of our distributed, decentralized leg control system for terrestrial locomotion. A compelling anguilliform swimming pattern emerges as a result of hydrodynamic simulation and the coordinated actuation of elastically deformable body segments. Locomotive transitions from land to water and vice versa in the creature's simulated physical environment are achieved smoothly by our locomotion controllers. The adoption of robust and efficient elasticity simulation techniques give rise to natural body deformations and support a novel approach to muscle actuation via rest-shape morphing. The virtual environment can be sensed by the simulated creature's antennae and the sensory information guides its adaptive behaviors, including obstacle avoidance and foraging.