In recent, unmanned or human-robot collaborative systems are spreading in security applications in order to
protect humans against sudden attack. Especially, the automated security system has to accomplish its mission even
when there are limitations of its sensors in view of range or reliability. For this dependable operation, it is better to build
up a security system using collective multi-robots than a single robot. So, we suggest a method how to organize
collective robot behaviors for a self-localization algorithm that allows a recursive state estimation process to be
collective in a multi-robot coalition team that is guaranteed connected. A leader robot in our method obtains a
temporary estimate from at the current time step using information from other robots (follower-robots) and checks out
whether it’s followers needs to be localized. When the leader robot decides to localize a follower robot, it recovers the
centralized-equivalent estimate with the help of its followers near to the follower robot. A practical implementation is
finally provided for five robots to emphasize the effectiveness of the proposed collective approach to multi-robot
localization problem. Moreover, we implement our robot behavior controller in OPRoS in order to enhance robot
behavior SW’s portability.