Feedback aided distributed routing based throughput optimization

This research proposes a light weight distributed routing system using ar- rival rate feedback to optimize the throughput of a wireless sensor network (WSN). The nodes have close to no knowledge about the system, the only information they have is their own hop count to the gateway and the hop count of their neighbors. Using this information nodes are able to route data traffic to the gateway of the network, however the throughput of the system is rather low. Using periodic feedback about the arrival rate of packets at the gateway nodes try to improve the network efficiency by avoiding slow or congested nodes. The system shows a maximum throughput increase of 45%, showing that the small amount of information that the feedback provides allows nodes to significantly improve network throughput. The system is not able to compete with systems that utilize knowledge about network topology with respect to throughput optimization. However, in systems where knowledge about the network topology is not feasible, this system could provide a significant performance increase with a very limited amount of feedback information.