R2D2 - Reliable Real-time communications for Dependable Distributed systems
|Duration:||March, 2009 - March, 2012|
|Senior researcher:||Elisabeth Uhlemann|
Wireless technologies offer significant benefits in many application areas; they provide a novel approach to existing applications, such as localization and tracking of goods, or enable new applications where wireless access is the only option, e.g., measure-ments and control of rotating or highly mobile devices. These application areas have diverse communication requirements, where real-time constraints are one important category. In a real-time system late delivery of information has the same effect as lost information and thus each real-time task has a deadline to meet. What happens if the deadline is missed varies with application. Many emerging applications using wireless communications also have concurrent requirements on reliable and real-time communications. However, existing communication protocols typically provide either reliable (e.g., emailing) or real-time communications (e.g., voice) — but not both. Research on real-time communication is traditionally focused on scheduling and priority in conjunction with different medium access control (MAC) methods, i.e. introducing real-time properties, but assuming a reliable (i.e., a wired) communication channel. When a wireless channel is considered and more reliable communications is called for, retransmissions are typically introduced and hence the real-time properties are lost.
In contrast to traditional research on real-time communications, the research in the R2D2 project is focused on increasing the data reliability by introducing changes to the lower layers in the communication stack such that real-time properties still hold (or are enforced). By using cross-layer design such that specific Quality-of-Service (QoS) requirements from the application layer are allowed to influence physical layer parameters (e.g., adaptive coding and modulation), the data reliability can be enhanced while respecting real-time deadlines.
Besides reliable wireless real-time communications, many applications employing dependable distributed systems also have additional requirements such as low latency (e.g., traffic safety systems) or energy efficiency (e.g., wireless sensor networks). Since the wireless channel is error-prone with limited resources that all users share, the communication protocols need to be made application specific such that the available resources can be fully exploited. A cross-layer design paradigm enables sufficient dynamics to meet the increased performance requirements of emerging applications such that the specific QoS constraints can be satisfied. Suitable trade-offs between information reliability, latency, energy efficiency, required bandwidth and complexity can then be obtained. R2D2 thus aims at developing adaptive, cooperative error control strategies tailored to the QoS parameters of each specific application or even situation. Information theoretical tools such as cooperative coding, cooperative diversity and network coding, will be applied and evaluated for practical systems with realistic assumptions.