A Research Of Spatial Neighbor Based Routing Algorithms For Wireless Senser Networks

Wireless Sensor Networks (WSN) is made up of great amount of cheap and tiny sensors deployed in a monitoring region. Through wireless communication methods, the sensors form a multi-hop and self-organized network system with the purpose of collecting the information of object being monitored and sending the result back to observer. With the abilities of free-deploying, self-organized networking, surrounding-sensing, WSN has broad application prospects in digital battlefield, surrounding-monitoring, scientific research and so on.As an important basic technology of WSN, routing technology is crucial to the application and widespreads of WSN. Giving the features of WSN like energy crucial, large-scaled, low storage ability of sensors, easy to be destroyed, the routing algorithm is facing new challenge. The geography-based routing changed the blindness and heavy overheads of flood routing algorithm, but it lead to the problem of routing voids because sensors only have the information of their one-hop neighbors. Aiming at the routing problems caused by the features of WSN and the shortcoming of present routing algorithm, a Spatial Neighbor Based Routing algorithm (SNBR) is proposed in this paper. The contributions of this paper are:(1) SNDA (Spatial Neighbor Discover Algorithm), the basis of SNBR, is proposed. It is used to collecting spatial neighbor information for sensors;(2) We propose the formal definition of spatial neighbor, analyze the features of spatial neighbor and use them in SNBR and SNDA algorithm;(3) SNBR suggests sensors expand their local information from one-hop neighbor to spatial neighbor to counter the problem of voids and improve routing efficiency;(4) Through a further study towards GPSR, We get that its bottleneck lies in the high frequency of entering face mode. SNBR reduces this frequency by the information of spatial neighbor and improves the efficiency of routing.(5) The simulation results in this paper verify three points: first, the size of spatial neighbor information is small enough to be stored on sensor nodes; second, the expending of spatial neighbor collection is small, that is to say, SNDA costs little energy and SNBR is applicable; third, the routing efficiency of SNBR is beyond GPSR, especially when there exists some voids in the networks.The overhead of SNBR is low, while its routing efficiency is high in situations such as low network density, un-uniformly network distribution and more network voids. So, the work in this paper is a basis for routing technologies to avoid voids.