| Abstract
|
<p>In modern factory automation syst … <p>In modern factory automation systems, data communication plays a vital role. Different nodes like control systems, sensors and actuators can communicate over a wireless or wired industrial network. The data traffic generated is often scheduled for periodic transmission, where each single message or packet must arrive in time. For this real-time communication, methods have been developed to support communication services with a guaranteed throughput and delay bound for such periodic traffic, but merely under the assumption of error-free communication. However, the possibility for errors in the transmission still exists due to, e.g. noise or interference. A node receiving sensor values from a sensor in the system might then be forced to rely upon an older sensor value from the latest period, possibly leading to inaccuracies in control loops which can compromise the functioning of the system. In safety-critical systems, redundant networks or communication channels are frequently added to cope with errors, leading to more expensive systems. In this chapter, we will describe an alternative approach where erroneous data packets are retransmitted in a way that does not jeopardise any earlier stated real-time guarantees for ordinary transmissions. Using our framework, the reliability of real-time communication can be increased in a more cost-efficient way. We describe in this chapter an overview of our framework for reliable real-time communication, while details of our approach can be found in our earlier publications. In the light of the emerging use of wireless communication, the framework proves to be especially useful due to the high bit error rate inherent to the wireless medium. However, the framework is naturally also attractive for wired communication systems.</p>for wired communication systems.</p>
|
