A Buffer Based Congestion Avoidance Algorithm For Wirless Sensor Networks

Wireless sensor networks (WSN), which have been gaining more and more attentions in academia and industry, have been used for more and more applications in recent years. The frequent changes of wireless channel, the interference of concurrent data on different wireless channel, the dynamic topology of WSN and the incident data flow caused by the sensed incidence can cause congestions in WSN. Congestion causes many problems. It decreases the quality of overall channel, increases the rate of packet loss and the delay of packet transmission, causes the packet loss due to the overflow of buffer, and costs extra energy of the sensors, which makes the lifetime of the networks shorter. Because of its own characteristic, the traditional congestion control strategies don't fit WSN. It is urgent to design congestion control algorithms that suit for WSN.This thesis introduced some basic concepts, principles, structures, applications and research development of WSN. Then some of the popular current congestion control algorithms were analyzed and compared. And the reason why the traditional congestion control strategies didn't fit WSN was analyzed. On the basis of these, this thesis proposed a buffer-based congestion avoidance algorithm. The main idea of the algorithm was that a sensor y could send a packet to another sensor x only when x had the buffer space to hold the packet. The algorithm was implemented with different MAC protocols. In particular, for CSMA with implicit ACK, the solution by changing contention window of binary exponential back off algorithm dynamically prevented hidden terminals from causing congestion. We described simple yet effective algorithm that prevented data packets from overflowing the buffer space of the intermediate sensors. The algorithm automatically adapted the sensors' forwarding rates to nearly optimal without causing congestion. The proposed algorithm, which didn't adopted the traditional rate-based congestion control strategy that was used in most of the congestion control algorithms of WSN, was evaluated by NS2 and had been proved efficiently. Finally, the paper summarized the work and made the next step tasks and research directions.