Abstract
Passive dynamic walker has excellent energy efficiency, but has problems keeping stable gait in real environments. Especially, it is sensitive to changes in slope angle. To overcome this problem, past researches altered foot shape with respect to slope angle to achieve gait stability. However, this poses a new problem of adaptivity. In this paper, we propose foot shape design which can robustly walk in multiple slope angles. We optimized foot shape by genetic algorithm and discussed how it improved its gait performance.
