Underwater sensor network（UWSN） adopts acoustic communication with more energy-consumption and longer propagation-delay, which bring great challenges to protocol design. In this paper, we proposed level-based adaptive geo-routing（LB-AGR） protocol. LB-AGR divides traffics into four categories, and routes different types of traffic in accordance with different decisions. Packets upstream to the sink are forwarded unicast to the best next-hop instead of broadcasting to all neighbor nodes as in present UWSN routing protocols. LB-AGR defines an integrated forwarding factor for each candidate node based on available energy, density, location, and level-difference between neighbor nodes, which is used to determine the best next-hop among multiple qualified candidates. Through simulation experiments, we show the promising performance of LB-AGR.
DU Xiu-juanHUANG Ke-junLAN Sheng-linFENG Zhen-xingLIU Fan
针对水下传感器网络的多用户干扰大和空间复用率低的问题进行了研究,提出采用AR模型预测信道未来状态,然后基于信道预测完成功率控制的算法,从而减小信道时空不确定性的影响。该算法利用随机几何理论,建立SINR(signal to interference-plus-noise ratio)模型,分析接收端的累积干扰状态,然后发送端在信道预测的基础上,以最小化网络中断概率为目标调整发送功率。实验仿真结果表明,基于信道预测的功率控制(power control based on predicted channel state,PCBPC)算法降低了网络能耗,提高了网络空间复用率。与NPC(nonpower control)算法相比,PCBPC算法在典型场景下将中断概率降低了14%,将网络功耗降低了33.3%,提高了空间复用率。
针对水声通信带宽低、时延长、误码率高、多普勒效应显著等特征,分析了传统可靠传输机制在水下传感器网络应用的局限性,提出了基于数字喷泉码-优化Raptor码的水下传感器网络可靠传输机制。采用反馈控制对Raptor码内码的鲁棒孤子分布和多项式描述的Shokrollahi度分布进行优化设计,降低了编码包平均度。进一步对Raptor码的内码—弱化的LT码(Luby transform codes)的编解码进行优化,实现了快速的编解码。基于优化Raptor码的可靠传输机制采用反馈控制,动态评估信道删除概率,从而提高编解码和通信效率。通过Aqua-Sim仿真工具对提出的可靠传输机制与基于编码的多跳协同可靠数据传输(coding based multi-hop coordinated reliable data transfer,CCRDT)机制进行仿真对比。结果表明,所提出的可靠传输机制明显降低了传输开销,提高了数据吞吐量。
In this paper, a strategy is developed for spectrum sharing among multiple cognitive users in underwater environment. This strategy requires all nodes to negotiate and reallocate the channels before sending data, and Hungarian method is used to maximize the sharing rewards. Simulation results show that the proposed strategy can avoid collisions between source-destination node pairs, and guarantee that the communication system gets maximum sharing rewards. Both the parameters of POMDP model and the number of available channels have influence on the system sharing rewards, and the rewards will increase when the channels have larger transition probabilities or more channels are available for communication. However, the channels with larger bandwidths can attract more nodes to access, and thus will lead to more collisions.
水下网络可用频谱范围比较窄,且部分被水下生物占用,导致了水下传感器网络可用的频谱资源更为稀缺.针对上述问题,提出一种基于累积干扰预测(Predicted Cumulative Noise,PCN)的水下认知网络动态频谱接入算法.该算法把水下生物作为认知网络的主节点,水下传感器节点作为次节点;通过建立水下生物业务行为的马尔科夫模型预测累积干扰,次节点根据预测结果,采用合作的方式动态地接入授权频谱.仿真结果表明,该算法能够保护水下生物正常通信的同时,实现最优化的频谱共享,网络容量增益达到6.3d B.
