The simultaneous start of actions in a distributed group of automatic devices: a decentralized control method with a signal repeater

This paper proposes a method for accelerating decentralized synchronization processes in the distributed control of a group of stationary or mobile automatic objects. With this method, the objects pass to specified states or affect the environment simultaneously or with required time delays. Some examples of such objects include actuators, computers in a computing cluster, distributed data processing facilities in supercomputers, and mobile robots. The object's action depends on the current state of all objects and the environment. The actions should start with minimum delay after detecting the possibility to perform them. Arbitrarily located sources of executive commands and their receivers are synchronized by exchanging signals and messages between objects through an intermediary (a signal repeater). Means are used to accurately measure the time intervals of signal transfer between each object and the repeater. Group operations are used to accelerate synchronization processes. These operations involve a large number of objects simultaneously. The object's data are used in operations simultaneously. Data are processed during their transmission without extra time. Operations are executed by network devices of the system objects and the common network device without any computing facilities (the repeater).