No abstract

A lifetime of several years for wireless sensor nodes can be achieved if their activity period is minimized. This can be done by using low duty cycle protocols. One of the challenges of these low duty cycle protocols is the synchronization of wake-up times. This becomes even more demanding if a wired central sink, which takes care of the schedule, is missing and communication links are unreliable. In this paper, we present our own Distributed Low Duty Cycle MAC (DLDC-MAC) protocol, which achieves a lifetime of two years using off-the-shelf sensor nodes. Our protocol synchronizes wake-up times of nodes in a fully decentralized way, even when plenty of transmitted packets get lost. We evaluated this in a real world scenario with 10 sensor nodes working for 14 days in our office space. Interestingly, the experiment showed that direct communication over long distances led to very high packet error rate, sometimes higher than 30%. In that case, multi-hop transmission delivered data with much higher reliability, with packet error rate close to 0%. KEYWORDSSensor networks, communication protocols, decentralized systems, wireless communication, energy efficiency 2009 International Conference on Computational Science and Engineering 978-0-7695-3823-5/09 $26.00

No abstract

No abstract

Protocols for sensor networks are commonly coupled to the specific operating system (OS), for instance TinyOS, involving some drawbacks. First, programmers must learn OS architecture, programming guidelines, sometimes a new programming language, etc. Second, protocols run on the specific OS only, i.e. on the hardware supported by the OS. Third, only selected network simulators, mostly provided by OS, can execute the OS-coupled code.To tackle the problems with interoperability we examined the idea of cross-platform protocol development for sensor networks, i.e. software running on several OSs, and share our experience in this paper.Our primary cross-platform MAC protocol runs on two OSs and on a hardware platform supporting the C programming language. To achieve interoperability we decoupled MAC from OS calls and provided a hardware abstraction layer (HAL) for each OS. We discovered that such an approach does not result in a significant penalty in terms of occupied memory (e.g. smaller by 3 kB than the TinyOSdedicated version) and consumed energy (additional overhead smaller by three orders of magnitude from total tx energy).