Study On Neighbor Discovery Algorithm With Low Latency In Wireless Sensor Networks

Wireless sensor networks, a blend of advanced technologies such as networks, information, communication and semiconductors, have been widely used in environmental monitoring, national defense and military affairs, industry, public health, as well as the protection of wild animals. For a long time, as hot research spots, MAC protocols and routing protocols of wireless sensor networks have drawn great attention, but the research on neighbor discovery that plays an essential and basic supporting role, hasn’t got enough attention.Neighbor discovery is of great significance for node identification, route construction, as well as collaborative work. However, the characteristics of constrained power of sensor nodes, intermittent of network connection, low duty cycle and low data rate, give rise to difficulties in waking up simultaneously and finding each other for two nodes. Traditional neighbor discovery algorithms are essentially pair-wise based, focusing on energy consumption; but some certain wireless sensor networks have higher requirement for discovery delay. The paper studies neighbor discovery algorithms with low latency in-depth from the aspect of discovery delay.First, the paper analyzes and extras the adaptive scheduling algorithm with lower latency of neighbor discovery. Discovery nodes get information of potential neighbor nodes from existing neighbor nodes, and wake up actively to discover potential neighbor nodes; meanwhile, we study the information recommendation mechanism among neighbor nodes, and discovery nodes selectively receive recommended information from neighbor nodes, through the comparision of the closeness among neighbor nodes(such as common neighbor ratio), in order to filter the redundant data information for reducing energy consumption, and then extend the life of the network.Second, we use the actual mobility model of nodes to predict the number of potential neighbor nodes within its communication range to further reduce latency, and then calculate the wake-up time of nodes according to the number of potential neighbor nodes and dynamically adjust the duty cycle; extend the wake-up time of nodes appropriately to monitor and discover more neighbor nodes quickly, so as to accelerate the discovery efficiency of neighbor nodes, as well as reduce discovery latency.Finally, the discovery latency and energy consumption of the two neighbor discovery algorithms are compared making use of simulations with different node density, duty cycle, communication irregularity and moving styles of nodes. Simulation results prove that both neighbor discovery algorithms can achieve good results in reducing discovery latency, improving the performance of networks. They have high practical value for wireless sensor networks with higher real-time demands.