Date Added: Jan 2011
Congestion control in sensor networks is important not only to improve the overall throughput but also to enlengthen the network lifetime by saving the scarce energy wasted during congestion. While throttling the incoming traffic during congestion can effectively alleviate congestion, it also lowers the throughput called accuracy level observed by the application deployed in the sensor network. In this paper, the authors investigate how congestion can be alleviated by increasing the available resource amount rather than suppressing the incoming traffic. To optimally increase the available channel capacity during congestion, the end-to-end channel capacity of a flow with multiple paths and their various configurations is analyzed, which is then verified by simulations.