Simulation model for locomotion in mobile legged robots that includes optimization and associative memory

It is usually accepted that complex locomotion behavior involves arrays of coupled neuronal oscillators. However, it seems reasonable to suggest that more neuronal machinery is needed to explain how lengthy sequences of movements are stored and selected to express a particular integrated motor behavior. Here simulation examples are presented where simple neural networks with the properties of optimization and associative memory are used as pre-processing elements to rhythmic oscillators forming then a more complete model of locomotion behavior machinery. Even if there is no experimental evidence that this occurs in biological movements the implementation of similar architectures in mobile legged robots could provide a more intelligent substrate for locomotion and other types of navigation