Streaming behavior generation and measurement error reduction in a super-mechano colony for effective object retrieval

Super-mechano colony is a parent-children type multirobot system proposed for an object retrieval mission over a large field. The present study addresses to gather objects efficiently when objects are concentrated in unknown several small areas. In order to realize an object gathering stream effectively, it is essential to reduce measurement errors included in position estimates. In this manuscript, two basic schemes are described for this purpose. One is to unify information reported by child machines on the parent machine. The other is to merge information among a number of child machines migrating as a party. The validity of this scheme is examined through simulations.