Generalized fault-tolerant pipelined task scheduling for decentralized control of large segmented systems

The control of complex, flexible structures requires substantial amounts of computational power to achieve precision performance in both space and time. This is due to the fact that such structures are inherently multiple input, multiple output systems whose complexities increase significantly with each additional parameter. The application of decentralized techniques can reduce the computational demands of these systems because multiple lower-order controllers replace a monolithic controller that would otherwise need to account for multitudes of system states in their calculations. Additionally, a decentralized control model provides a framework for the development of parallel control algorithms for both the high performance and fault tolerance of a sophisticated control system. This paper introduces a novel approach to scheduling computational tasks on processors in a multiprocessor environment. The approach is described in detail and compared against a general straightforward scheduling mechanism. Pipelined task scheduling features increased throughput of control computations and fault tolerance, justifying its use over conventional methods. Both pipelined and straightforward task scheduling algorithms have been applied to a physical control-intensive system; the results indicate a sound design and encourage further work involving pipelined task scheduling.