Goal recognition and configuration recognition algorithms for modular robots

Modular robots are a class of robots which are composed of a set of similar and independent mechatronical parts. Researchers and scientists are interested in these robots mostly because of their massive adaptability. These robots are likely to be used in various fields of industry in near future. Given the fact that an important application of modular robots is accomplishing missions in remote and hard-to-reach areas, detailed knowledge of their configuration and relative arrangement of their modules are of great importance. In this paper, at first we introduce and simulate a distributed goal recognition algorithm and its flaws. In this algorithm, the robot´s modules analyze a shape matrix and recognize if their configuration matches the goal´s shape or not. Then, we present an improvement to the algorithm and discuss its advantages over the previous version. We also introduce an algorithm for a robot to recognize its own configuration when needed and send the proper message to the querying system. This algorithm is distributed and is based on the module´s communication to adjacent modules.