Evaluating alternative gait strategies using evolutionary robotics

摘要: Evolutionary robotics is a branch of artificial intelligence concerned with the automatic generation autonomous robots. Usually form robot predefined and various computational techniques are used to control machine's behaviour. One aspect spontaneous walking in legged robots this can be investigate mechanical requirements for efficient bipeds. This paper demonstrates bipedal simulator that spontaneously generates running gaits. The model customized represent range hominoid morphologies predict performance parameters such as preferred speed metabolic energy cost. Because it does not require any motion capture data particularly suitable investigating locomotion fossil animals. predictions modern humans highly accurate terms cost given thus values predicted other bipeds likely good estimates. To illustrate transport calculated Australopithecus afarensis. allows degree maximum extension at knee varied causing adopt gaits varying from chimpanzee-like human-like. costs associated these gait choices information evaluate possible locomotor strategies early hominids.