Dispatching of Robotic Control Programs

摘要

The problem of dispatching discipline choice when managing programs are linked into unified multi-loop computer control system is considered. It is shown that a problem of control of such a system may be reduced to the problem of evaluation of states both robot and controller. In multi-circuit computer control systems time intervals of residence of robot in any state depends on both time complexity of control algorithm and dispatching discipline. Two simplest disciplines of most common use are investigated: the cyclic dispatching and foreground (quasi-stochastic) one. With use the formalism of semi-Markov process models of functioning of control programs under investigated dispatching disciplines are worked out. Mathematical relationships for time of return to any state of semi-Markov process and time between switches are obtained. The parameters obtained are essential for choice the efficient regimes of data processing when control of robots.