Poster Abstract: Exploiting Multi-Channel Diversity to Speed Up Over-the-Air Programming of Wireless Sensor Networks

Wireless Sensor Networks consist of hundreds or thousands of ad-hoc tiny sensor nodes (motes) that are able to sense, compute, and communicate. Because of the large scale of sensor networks, programming them manually is a tedious and, sometimes, even impossible task. Thus, it is essential to have the capability to program them wirelessly.

Various schemes for over-the-air programming of sensor networks have been proposed. However, When node density increases, the performance of these schemes often degrades because of increasing traffic contention on the shared channel.

We propose to speed-up over-the-air programming by harnessing the multi-channel transceiving capability of motes. For instance, in the 902-928 MHz frequency region, there are as many as 25 non-overlapping channels (frequencies) over which motes can communicate. The idea, therefore, is to relieve network congestion by splitting the traffic among different channels. We developed a general, dynamic method for exploiting multi-channel resources with single radio nodes. This method, described in more detail the paper, divides nodes into different groups. Nodes from all the groups can always communicate over a default channel. However, whenever possible, nodes belonging to the same group temporarily switch and communicate between themselves over an auxiliary channel that is unique to each group. We have devised a new over-the-air programming scheme for TinyOS motes using this method, called Multi-Channel Deluge. We have implemented Multi-Channel Deluge on a testbed of MICA2 motes operating in the 902-928 MHz frequency region. Our initial experimental results show that Multi-Channel Deluge can reduce the programming time by as much as 60% compared to the standard implementation of Deluge*.

The work is still in progress, if you are interested or need source code, please contact weiyao@bu.edu.

*J. W. Hui and D. Culler. The Dynamic Behavior of a Data Dissemination Protocol for Network Programming at Scale. Proceedings of ACM SenSys '04, Baltimore, MD.