Applied Multi-Agent Systems Principles to Cooperating Robots in Tomorrow’s Agricultural Environment Master Thesis in Computer Systems Engineering

Progress in software engineering over the past two or three decades has primarily been made through the development of increasingly powerful and natural abstractions, with which to model and develop complex systems. This progress has led to the notion of Multi-Agent Systems (MAS), which is seen as, yet, an increase of abstraction compared to the object-oriented approach. The MAS society has proposed many theories and concepts, but so far practical experiences with these theories and concepts are a limited resource. In this thesis, a number of MAS principles, including organisation, negotiation and agent capabilities, are applied to a real world context, in this case collaborating robots in an agricultural environment. A platform, called LEGOBot II platform is presented that utilises and combines the above mentioned MAS principles, which enable us to use MAS concepts on practical problems. The application of the MAS principles to the practical problems, showed that the investigated concepts were ready and mature, to undergo the transition from theoretical ideas to practical realisations. Furthermore, it was revealed that today’s MAS frameworks advantageously could combine concepts from organisation, agent reasoning and capabilities.