Recent advances in the field of hybrid communication systems for factory automation have shown that the processing delay introduced by some network components may severely affect the performances of the communication system, introducing a high degree of randomness in the delivery time of packets. Real–time behavior could result not effectively satisfied and a high number of deadlines might result lost. Measurements are hence needed in order to obtain a statistical description of such network devices: an information that is in turn useful at the design stage of an automation network. To this aim an original and effective method providing the required analysis is presented, exploiting a simple measurement setup and low instrumentation efforts. It is based on a substitution approach, in which a closed loop network is at first characterized. The actual Device Under Test (DUT) is subsequently included in the measurement loop. The DUT characterization is obtained applying a suitable deconvolution algorithm to the data collected in these two steps. An experimental session involving real devices is finally provided, showing the performances of the proposed measurement approach. Theory and comments to the obtained results may also provide specific guidance to a designer in order to optimize the performances of a control network.
Estimation of the delay of network devices in hybrid wired/wireless real-time industrial communication systems
BERTOCCO, MATTEO;NARDUZZI, CLAUDIO;TRAMARIN, FEDERICO
2012
Abstract
Recent advances in the field of hybrid communication systems for factory automation have shown that the processing delay introduced by some network components may severely affect the performances of the communication system, introducing a high degree of randomness in the delivery time of packets. Real–time behavior could result not effectively satisfied and a high number of deadlines might result lost. Measurements are hence needed in order to obtain a statistical description of such network devices: an information that is in turn useful at the design stage of an automation network. To this aim an original and effective method providing the required analysis is presented, exploiting a simple measurement setup and low instrumentation efforts. It is based on a substitution approach, in which a closed loop network is at first characterized. The actual Device Under Test (DUT) is subsequently included in the measurement loop. The DUT characterization is obtained applying a suitable deconvolution algorithm to the data collected in these two steps. An experimental session involving real devices is finally provided, showing the performances of the proposed measurement approach. Theory and comments to the obtained results may also provide specific guidance to a designer in order to optimize the performances of a control network.Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.