用于智能通信的无线传感器网络

---

在标题为“无线传感器网络的智能通讯”特刊的第一版，共22个手稿共收到和这6个被接受。这个问题显示，网络拥塞，用户移动性，和相邻频谱的干扰是通信质量的无线传感器网络（WSN）的劣化的主要原因。

在无线传感器网络中，利用最优路由树算法可以延长网络的生命周期。将基于子梯度的拥塞最优Wi-Fi卸载算法与虚拟拥塞最优Wi-Fi卸载算法相结合，可以得到每个接入点(AP)的最优卸载率。更重要的是,通过使用设备间(D2D)资源分配和再(资讯)辅助机器学习算法,我们可以获得有效的频谱资源分配方案,同时利用移动状态检测算法和训练的人工神经网络(ANN)模型基于无线个域网节点,我们可以估算出标题的方向和位置的移动用户。

这种特殊的问题已经成功吸引了许多有趣的原创文章讨论了无线传感器网络的优化智能通信。例如，L. Wang等。研究不同造成社区之间频谱共享和干扰功率低效的频谱利用的问题，以及所使用的动态博弈论优化D2D通信蜂窝网络频谱资源分配方案。所提出的分配方案不仅量化D2D发射功率干扰用户的数据传输速率的影响，同时也对量化数据传输速率不同的移动用户之间的社会关系的影响。该方案全面测量两个以上因素上的数据传输速率的影响，同时依赖于纳什均衡基于效用函数来设计基于资源优先级搜索，然后将其用于优化频谱效率的资源分配的方法。在随后的研究中，Y. Sun等。提出了一种基于人工神经网络模型的自由设备的无线定位系统和所使用的ZigBee节点来构造用于无线传感器网络中不同的传感器节点之间的通信的硬件平台。通过设置RSS数据及其相应的索引作为输入和已知位置作为输出的神经网络模型训练的坐标的方差，能够获得在不使用特殊终端的令人满意的定位结果。随后，H. A. Shah等。报道使用频谱感知算法KNN认知无线电（CR）网络，以提高频谱利用率的策略。 In training phase, this strategy makes global decision based on the perceptual report generated by each CR user, i.e., sending information or keeping silent. At the same time, the majority decisions of different CR users are merged into global one, which is then returned to each CR user. In addition, at each CR user, according to the comparison between global decision and the actual primary user activity determined by confirmation signal, the sensing classes are formed. Then, in classification phase, by comparing each CR user’s current sensing report with the existing sensing class formed in training phase, the distance vector and posterior probability of each perceptual class are calculated to indicate the presence or absence of primary users. In all, this strategy uses a decision-making combination scheme to infer the reliability of each CR user, which is able to determine sending information or keeping silent based on global decision. In response to the Wi-Fi offloading problem, B. Liu et al. studied the problem of network congestion and user mobility management in smart communications. They proposed a Congestion-optimized Wi-Fi Offload (COWO) algorithm to obtain the optimal offload ratio of Wi-Fi networks, which is considered for the enhancement of network throughput as well as mitigation of network congestion. In addition, they improved the COWO algorithm through equivalent transformation and developed a simple Virtual Congestion-optimal Wi-Fi Offload (VCOWO) algorithm, which can well approximate the optimal result obtained by COWO. Finally, extensive simulation results show that the VCOWO is capable of achieving higher network throughput and lower network congestion compared with the existing state-of-the-art. In terms of user mobility, Z. Deng et al. pointed out three special states of human motion, i.e., random hand movement, change of heading direction, and terminal location variation. The performance of heading direction estimation depends on the discrimination of these three states, which can be well achieved according to the user movement states detection, namely, Rotation Matrix and Principal Component Analysis (RMPCA). Besides, the outlier elimination algorithm is also used to improve the accuracy of heading direction estimation of pedestrian. Finally, P. Cao et al. proposed to use multiple Central Processing Unit (CPU) cores to accelerate the process of constructing the optimal routing tree corresponding to the maximal lifetime of WSNs. The goal of this approach is to break down the lifetime maximization problem into several separate subproblems which can be easily solved on each CPU core at the same time. To achieve this goal, they propose three decomposition algorithms, in which two of them are based on the assumption that routing tree does not involve any cycle and the other one is based on the assumption that any node in routing tree has at most one parent node. According to the numerical testing carried out on an 8-core desktop platform, the proposed approach is verified with faster computation speed compared with the conventional ones using only one CPU core.

因此，从我们的角度来看，这期特刊带来了新的见解无线传感器网络的智能通信。我们希望这些信息将有助于铺平道路，在进一步研究无线传感器智能和认知通信的发展道路。

利益冲突

本人声明，本人及其他客座编辑均与公司无任何利益冲突或私人协议。

穆州
祁连梁
(吴
Weixiao孟
Kunjie徐

版权

版权所有©2018木洲等这是一篇开放获取下发布的文章知识共享署名许可，允许在任何媒体中不受限制地使用、发布和复制原创作品，只要原稿被正确引用。