This project´s general objective is to make progress in the field of the Semantic Navigation of mobile robots. This level increases the level of autonomy of the robot, providing it with greater decision making ability and it makes easier the integration of the robot with humans and their environments.
The Navigation Systems of the robots have evolved from purely geometric concepts to semantic concepts, going through topological levels. Along literature, this trend is clearly reflected, being widely treated geometric and topological levels. However, the semantic level in most cases it stays in labelling of objects in the environment and it does not consider its full integration with other systems of the robot. The present proposal pretends to carry out research in this line, specifying the tasks of navigation of the robots and extending it to the case of autonomous vehicles driving through urban and interurban roads. The current proposal is focused on
- Modeling of the environment: study of an ontology applied to navigation in structured outdoor environment.
- Definition of an ontology that represents a structured environment for navigation in an outdoor environment, as in the case of urban roads.
- Development of a system of inference that handles the semantic information of the environment, generating the required commands on the driving system of the vehicle according to the detected semantic situations. (Affordances based on situations).
- Define and implement a learning mechanism that allows learning such actions upon detected situations.