出版社：The Japanese Society for Artificial Intelligence
摘要：``Embodied Evolution (EE)'' is a methodology in evolutionary robotics, in which, without simulations on a host computer, real robots evolve based on the interactions with actual environment. However, we had to accept robot behavior with low fitness especially in the early generations when adopting the EE framework. We introduced pre-evaluation into the EE framework so as to restrain robot behavior whose fitness is predicted to be low. This paper reports on the introduction of pre-evaluation into the Embodied-Evolution framework for a biped robot in order to reduce the risk of falling.
