The Hi-NOON neural simulator and its applications to animal, animat and humanoid studies

摘要

This paper describes the Hi-NOON neural simulator, originally conceived as a general-purpose, object-oriented software system for the simulation of small systems of biological neurons, as an aid to the study of links between neurophysiology and behaviour in lower animals. As such, the artificial neurons employed are spiking in nature: to effect an appropriate compromise between computational complexity and biological realism, modeling was at the transmembrane potential level of abstraction. Further, since real neural systems incorporate different types of neurons specialized to some what different functions, the software was written to accommodate a non-homogeneous population of neurons. Hi-NOON has been used in animat (crick et phono-taxis) and biologically-based robot studies. In particular, it was employed to implement the nervous system of our ARBIB robot. A simple model of synaptogenesis has been added so improving the stability of its learning in the light of the stability-plasticity dilemma, and as a mechanism for long-term memory. The efficacy of the simulator is illustrated with respect to some recent applications to situated systems studies. Now that Hi-NOON has been expanded to simulate large nervous systems in a concurrent environment, it can be applied to humanoid robotics in the future.