Strategies are the crucial part of the robot's behaviour. Without good strategies, the whole framework is more or less useless. Strategies have to be registered by the master resource control of the framework.
A strategy is in fact a collection of:
- server and client states
- the beginning states
- the data repositories (used for exchanging informations on the server / client from state to state and from finite state machine to finite state machine)
- the corresponding parameters in the configuration file
A detailed explanation, when and how the framework calls the methods of the strategie's states, is given in the
publications of the framework-developers.
Tasks in this field are:
- designing and implementing collaboration algorithms that can be put in states of a finite state machine
- collecting ideas of REALIZABLE tactics / brainstorming
- research on basic problems with solutions that can be used in different strategies. For example:
- "how my robot avoids to crash into a wall"
- "how my robot recognize what targets are friends"
- "how the robots can track an enemy"
- "how robots can control their movement"
- "what data structures are required to save information all other automates should have to"
- verification of the effectiveness of the strategies by competing other robots
- analysation of the tactics of the provided robots and freely available robots
- analysation of the log files of the world champion chips
- verification if algorithms are fast enough (if not, the synchronization of the framework will lead to a time shift of the reality of the RTB-Server and our clients)
- implementation of physical helper classes
Publications of the Strategy Team:
