Wireless sensor networks (WSNs) are a collection of small, simple sensor devices. WSNs are used in applications such as fire detection, industrial process control, and environmental and health monitoring. They generally comprise a number of small sensor nodes that transmit data to a central point by multihop wireless transmission. These networks are typically expensive and difficult to maintain; developing and testing the various components is considerably simpler using simulation rather than the real networks.

Chatzigiannakis et al. assert that accurate simulation of these networks requires the inclusion of obstacles in the simulation process, and propose a model for these obstacles. They review earlier work, and detail a systematic and generic obstacle model for simulations. They define the various classes and shape of obstacles, and discuss, in some detail, the impact that these obstacles have on network performance (focusing on three representative network protocols). The authors also provide experimental results from the implementation of their model in the simDust network simulator.

As stated, the introduction of a common, generic obstacle model will improve the accuracy of WSN simulations. The paper is quite readable. The experimental results are well presented, with appropriate supporting graphs and diagrams. The paper and the obstacle model will be useful to anyone simulating general wireless networks--not just sensor networks.