System Implementation 1/3- MRMA Development Environment
Our overall system is a Multiple Robot Mobile Agent (MRMA) environment shown in the slide above, which we have been using to identify roles for mobile agent mobility in multiple robot teams. A Mobile Agent (MA) is a software entity which exists in an execution environment within which it has a clearly defined boundary. It can control to a large extent its own execution and interaction with other agents or computing components i.e. it is not purely reactive and it may be autonomous, asynchronous, dynamic and intelligent. A MA differs from a static software agent in the sense that it is mobile within its execution environment. When, as in this case, an agent’s execution environment is composed of a number of distributed computers or robots, this allows it to move between computing nodes in order to perform tasks. MAs have software engineering advantages over an amalgamation of other distributed computing architectures, can make our system more adaptable and robust, and allow us to dynamically position control or learning at the most appropriate location within a distributed environment. By encapsulating learning in a MA we can apply these characteristics to this functionality, independent of the learning ability.