The name is absent

•     Brooks' subsumption architecture (Brooks, 1986; 1989), which can be used as a mean

to implement robot control systems, which include tasks of perception and action, in
addition to the emergency of behaviours.

•     Rosenblatt and Payton's hierarchical network (Rosenblatt and Payton, 1989), a

mechanism similar in many aspects to the hierarchical models proposed by Tinbergen
and Baerends, but with nodes like formal neurons inspired in Brooks’ subsumption
architecture.

•     Maes' bottom-up mechanism (Maes, 1990a; 1990b), a distributed non-hierarchical

network of nodes, where each node represents an appetitive or consummatory
behaviour that the entity can execute.

•     Beer's neural model (Beer, 1990; Beer, Chiel, and Sterling, 1990), a semi-hierarchical

network of nodes, where each node represents a neuronal circuit. It is inspired in the
neuronal circuits of the American cockroach.

•     Halperin's neuroconnector network (Hallam, Halperin, and Hallam, 1994), a non

supervised neural network organized in layers.

•     Negrete's neuro-humoral model (Negrete and Martinez, 1996), a non-hierarchical

distributed network of nodes, where each node is a neuron with neuronal and humoral
capacities.

•     Goetz's recurrent behaviour network (Goetz and Walters, 1997), a network of nodes,

where a node can represent a behaviour, a sensor or a goal; the network converges to
a particular behaviour (attractor), in a similar way that a Hopfield's network (Hopfield,
1982) converges to a certain pattern.

Table 1 shows a comparison among the different ASMs before mentioned, taking into
account the most relevant aspects from these.

28

<<   32   33   34   35   36   37   38   >>

More intriguing information