Resolution of Conflict in Robotic Systems based on local and global Sensor Network


W. Jacak, K. Pröll - Resolution of Conflict in Robotic Systems based on local and global Sensor Network - Proceedings of International Conference Computer Aided Systems Theory EUROCAST 2009, Las Palmas, Spanien, 2009, pp. 1-9


In this paper we present the concept of a framework for an autonomous robotic agent that is capable of showing both neural network-based and sensor-based reactive behavior. The application of Intelligent Multi Agent Technology and Interactive Environment in a Complex Robotic System provides a solution of the task reachability problem in case of a space conflicts among active agents and known and unknown dynamic objects in the world model. Based on conflict resolution and updating of the world model on line mo-tion planning leads to resolution of task achievement. Usually an agent has only partial information about the world state obtained from own local perception system (sensors system). This information is usefull to solve only the local conflicts in the sense of reactive behaviour. To solve the possible conflict in artifi-cial intelligence maner is necessary to extend the information source .One way of the such extension is using additionally ther information from distributed sensors network. In our concept, the structure of a robotic agent consists of two independent subsystems: the action planning system and the action execution system with integrated conflict reso-lution method. • The action planning system uses the aggregated knowledge about all prepared ac-tions and conflicts which occurred in the past and knowledge about environment changes coming from global sensors network. • The action execution system is able to solve the local conflict situation (local space conflict by autonomous robotic agent) in a reactive manner. The so called conflict resolution part of agent makes use of knowledge not only from sensors mounted on active agent (local sensor network) but also from distributed global sensor network.