The main components in the study of multi-agent systems in the field of IoT are intelligent agents and contexts. From the viewpoint of each agent, the primary objective is to choose actions that maximize the agent's future utility. The choice of a correct action depends on a process of complex data collection from context. Using a set of neuroscience model we consider the receptory field to slice the context area and to group the appropriate sensors. To permit the phenomenological analysis and consequent cognitive evaluation, we have applied Bellmund and Doeller’s model of hippocampus. This session describes only the phenomenological layer architecture. From a scientific point of view there is a novel fusion between fluent calculus and neural network ensemble. Such net are used to associate to situations’ fluent the appropriate schema of actions.
Keywords—IoT, Neural Network Ensemble, fluent Calculus, Place Map, Grid map, receptive field.