Methodology For Parallelizing Networked Control Systems And Its Schedule

Advanced control, which provides better control performance, is highly preferable with the continuing requirements of modem industry on improving its productivity. But the implementation of advanced control is largely limited by its complicated and time-consuming computations. Recently, the development of NCSs (Networked Control Systems) provides a solution for applying advanced control freely by the way of parallel computation. And function block (FB) will be a fundamental notion to implement NCS-based parallelism. So the basic concepts of function block are introduced here and two types of FB-based parallelism, pipeline-parallelism over single loop NCSs and parallelizing of control algorithms within FB-based NCSs, are also forwarded. The two types of NCSs parallelizing method could improve resource utilization, so that sampling rates could be increased and control performance could also be improved.Though the paralleled NCSs sound better, the implementation of parallelism over NCSs is not so easy due to the speciality of NCSs. For instance, communication data must transmit via networks in NCSs, which leads to increment of transmission delays due to parallelism. And the increased delays may bring to degradation of system performance. Otherwise, re-designing of control algorithms and re-constructing of software and hardware architecture are also required for the implementation of parallelism over NCSs.For these problems, this dissertation details the analysis of network load in paralleled NCSs, with an example of a CAN bus networked control system. The analysis of data transmission delays between different network nodes is also detailed. Based on these analyses, a QoS management method for paralleled NCSs is put forward to figure out the increment of delays. And a QoS management algorithm is also designed to provide a feedback way for resource scheduling, which could make a tradeoff between the degradation of network QoS and the improvement of control performance.Finally, a graceful degradation method based on the FTTCAN protocol is put forward. On the basis of efficient sensor ability, a detailed implementation architecture of QoS graceful degradation algorithm is designed, with the master node as the QoSmanager in the FTTCAN protocol.