Title :
Achieving periodic leg trajectories to evolve a quadruped gallop
Author :
Krasny, D.P. ; Orin, David E.
Author_Institution :
Dept. of Electr. Eng., Ohio State Univ., Columbus, OH, USA
Abstract :
For most large quadrupedal mammals, galloping is the preferred gait for high-speed locomotion. In this paper we evolve a gallop gait in a simulated quadruped robot at speeds from 3.0 to 10.0 m/s. To do so, we must generate periodic trajectories for the body and legs. An evolutionary algorithm known as set-based stochastic optimization (SBSO) is used to find the body trajectory while alternative methods are used to find periodic leg trajectories. The focus of this paper will be to evaluate three different methods for generating periodic leg trajectories. The combined solutions for the body and legs yield biological characteristics that are emergent properties of the underlying high-speed dynamic running gait.
Keywords :
gait analysis; genetic algorithms; legged locomotion; robot dynamics; stochastic programming; body trajectory; evolutionary algorithm; gallop gait; high speed dynamic running gait; high speed locomotion; periodic leg trajectories; quadruped gallop; quadruped robot; quadrupedal mammals; set-based stochastic optimization; Biomimetics; Evolutionary computation; Genetic algorithms; Hydraulic actuators; Leg; Legged locomotion; Optimization methods; Resonance light scattering; Robots; Stochastic processes;
Conference_Titel :
Robotics and Automation, 2003. Proceedings. ICRA '03. IEEE International Conference on
Print_ISBN :
0-7803-7736-2
DOI :
10.1109/ROBOT.2003.1242186