Juan Saez Pons

Country of Origin: Spain

Start: 01/10/04

Scientific Supervisor: Prof. H. A. Mallot

Job description:

Spatial memory in animals and humans can be modeled as a network (graph) of recognized places and actions used to navigate between adjacent nodes. In previous work, we have implemented a basic version of the graph model of spatial memory including visual sensors for place recognition and egomotion perception, and metric graph embedding based on multidimensional scaling. The model is used to simulated spatial behaviors such as way finding in insects or territory formation in lizards. Within the PerAct network, this work should be extended in one of the following directions: (i) Other sensory modalities (acoustics) for place recognition. (ii) Spacial memory in changing environments. (iii) Simulation of biological behaviors.