Self-organization of SN-MRS systems


The research objectives can be summarised by the following statement:
the study of bio-inspired task allocation algorithms for SN-MRS systems with high number of units, and with particular attention to scalability, real-time constraints, reliability of the transmitted data, success of robot operations and network task-aware self-organisation.
Recently, a few projects have studied the integration of SNs and MRSs. The approaches in the literature follow usually two ways. In the first one, sensors are treated as "helpers" for the robots. They are for instance landmarks that robots follow to navigate in dificult environments. In the second approach, the robots help the sensor network, for instance by replacing out-of-work pieces or, since they have room for bigger batteries and they can afford more powerful transmission, they work as gateways for long distance communication. In both cases, the interactions between robots and network are outbalanced either toward SN or toward MRS. The former approach ("sensors help robots") seems to be followed mainly by researchers with a strong background in robotics and Artificial Intelligence, the latter ("robots help sensors") by researchers coming from the telecommunication field.
Seen from a high abstraction level, the main objective of this proposal is to change the approach to SN-MRS systems. We do no want to treat them as the sum of two systems, but only as one heterogeneous system, where the difference between static and mobile units becomes fuzzier. To design and program such systems, it is required a good knowledge of both telecommunication networks and robotics. Both approaches have their advantages and disadvantages, but we believe that pushing the integration further will reveal new possible applications and interesting solutions for SN-MRS systems.
The following research aspects affect our developments:
- number of units - time constraints - task allocation - network-robot interactions

  • Projektdauer: 2005-11-01 - 2006-04-30

Projektmitglieder

  • PD Dr.-Ing. habil. Falko Dressler
  • M.Sc. Thomas Halva Labella

Förderer

  • DAAD
  1. Thomas Halva Labella, Isabel Dietrich und Falko Dressler, "BARAKA: A Hybrid Simulator of SANETs," Proc. 2nd IEEE/ACM Int. Conf. on Communiation Systems Software and Middleware, Bangalore, India, Januar 2007
  2. Thomas Halva Labella und Falko Dressler, "A Bio-Inspired Architecture for Division of Labour in SANETs," in "Advances in Biologically Inspired - Models, Methods, and Tools", Falko Dressler (Ed.) und Iacopo Carreras (Ed.) ser. Studies in Computational Intelligence (69), Springer, pp. 209-228, 2007  
  3. Thomas Halva Labella, Gerhard Fuchs und Falko Dressler, "A Simulation Model for Self-organised Management of Sensor/Actuator Networks," Proc. of GI/ITG KuVS Fachgespräch Selbstorganisierende, Adaptive, Kontextsensitive verteilte Systeme, Kassel, Germany, März 2006