在高速铁路建设中，采用的电力拖车、路堑、站台、桥梁、明洞等特殊、形状不规则的建筑物，会对电磁波传播机理和多通路产生一定的影响，使其呈现出明显的衰减特性。从空间上看，呈现出强烈的空间异同性和频度相关，从而导致经验模型无法准确表征铁路场景的电波传播特性，成为制约无线网络规划精确度与效率、限制铁路移动通信系统高质量发展的理论瓶颈之一。针对这一挑战性难题，本文开展了基于多源异构数据融合的铁路移动通信高精确度网络规划及优化系统研究。采用高性能射线跟踪(RT)技术针对铁路通信 930 MHz频段下的高铁场景进行了网络优化仿真。仿真结果表明，90%的接收信号强度相较于优化前提高了 13.9 dB，显著提升了铁路复杂场景中的无线网络质量。

随着移动通信技术的快速发展和更新换代, 各行各业的信息化、智能化日益依赖于无线通信的强有力保障。因此, 准确的信道信息、可靠的无线覆盖至关重要。射线跟踪技术面向不同行业的各类通信场景, 通过建立有效的三维场景模型, 以材料电磁参数、天线参数为基础, 依托高性能计算引擎, 将光学的射线技术引入电磁计算领域, 能够准确地表征反射、散射、绕射、透射等多种电波传播机理及其对无线系统的影响。本文介绍了作者团队自主研发的高性能射线跟踪仿真平台, 围绕高性能射线跟踪技术在智能交通领域的应用案例, 阐述了射线跟踪技术在轨道交通、水路交通、公路交通等智能交通领域无线网络规划与优化、无线信道建模及信道特性分析等方面的技术优势, 以期以自主可控的高性能射线跟踪仿真技术更好地助力我国交通强国战略的实施。

在大型风电叶片动态摄影测量中, 为了对相机的站位进行优化, 采用一种变异操作改进型遗传算法作为摄影测量网络优化方法, 通过光线束前方交会的误差传递建立测量误差模型, 以空间坐标测量误差的标准差为网络优化的目标, 同时根据被测风电叶片几何结构和实际环境确定了相应的约束条件进行仿真实验, 得到了最优的相机站位。结果表明, 在以叶片长度为40m的风机为被测物的仿真实验中, 最优站位的空间坐标测量误差标准差为2.7mm; 通过对叶片长度为3.5m的风机模型进行实测实验验证, 最优站位的相对测量误差为0.009%, 最大误差为0.617mm。该研究为风电叶片摄影测量的网络优化提供了参考。

本文针对传统车道识别方法在复杂路面中自适应能力差的特点，基于图像分割技术提出了一种基于全卷积神经网络与条件随机场的车道识别方法。该方法通过大量数据的训练，使神经网络模型可以识别出车道，并且再通过条件随机场使得分割出来的车道覆盖面积及车道边缘的处理更加完善。同时，本文为了解决高速公路中对检测实时性的高要求，设计了一个全卷积神经网络，该网络结构简单，只有 13万个参数，并且做出如下三点改进：采用 BN算法提高网络的泛化能力及收敛速度；采用了 LeakyReLU激活函数取代了一般使用的 relu或者 sigmoid激活函数，并且采用 Nadam作为网络的优化器使得该网络具有更好的鲁棒性；采用条件随机场作为后端处理解决车道边缘处分割不足并且加大了车道覆盖面积。最后本文为了解决城市道路检测中道路环境复杂的问题，利用 FCN-16s网络模型加条件随机场的后端处理实现了复杂城市道路的识别。实验证明，在面对高速公路的高速及车道简单环境下，本文设计的网络模型更具有实时性且足够胜任车道的识别。在面对城市道路的复杂环境下， FCN-16s模型加条件随机场更能精确地识别出车道，并在 KITTI道路检测基准上取得不错的结果。

The rapid growth of the Internet raises the importance of resource planning of Internet protocol (IP) over elastic optical networks (EONs), which is a challenging task due to more complex and obscure physical -constraints of it. Compared with network cost, the power consumption may eventually become the barrier to the expansion of the Internet. We present an energy-efficient virtual topology design (VTD) scheme for IP over EON. We explicitly explain and analyze the mixed integer linear programming model and the heuristic algorithm for this scheme. Numerical results show that the proposed VTD scheme can significantly save power consumption.

Compared to the traditional wavelength division multiplexing (WDM) optical networks with rigid and coarse granularities, flexible spectrum optical networks have high spectrum efficiency, which can support the service with various bandwidth requirements, such as sub and super channel. Among all network performance parameters, blocking probability is an important parameter for the performance evaluation and network planning in circuit-based optical networks including flexible spectrum optical networks. We propose an analytical method of blocking probability computation for flexible spectrum optical networks in this letter through mathematical analysis and theoretical derivation. Two blocking probability models are built respectively based on whether considering spectrum consecutiveness or not. Numerical results validate our proposed blocking probability models under different link capacity and traffic loads.

With the emergence of high-bitrate applications, cross stratum optimization (CSO) attracts the interest of network operators because of its application in the joint optimization of optical networks and application stratum resources. Given the large-scale growth and high complexity of optical networks, achieving a more effective, accurate, and practical CSO becomes an important research focus. In this letter, we present a CSO-oriented, unified control architecture for OpenFlow-enabled triple-M optical networks. A novel dynamic global load balancing (DGLB) strategy with dynamic resource rating for CSO is presented based on the proposed architecture. The DGLB strategy is then compared with four other strategies by conducting experiments on a SOFT-based testbed with 1000 virtual nodes.

Efforts in realizing all-optical packet switching are overwhelming in the past decade. While optical packet switching remains an attractive switching paradigm in the long run, technical challenges prohibit it from becoming a practical solution for the ever-growing bandwidth hunger during the next few years. Finding a technically viable way to meet the increasing capacity requirement with good scalability and flexibility becomes a clear pursue for the community. Hybrid packet and circuit switching is considered to be one promising technique in realizing high performance switching at low cost and less energy consumption, by taking the advantage of both packet switching and circuit switching. In this paper, we review existing work in hybrid optical packet and circuit switching. We discuss the key technical challenges in realizing hybrid optical packet and circuit switching. We further introduce our ongoing efforts in building a seamlessly transformable packet/circuit-switching node with hybrid optical and electronic components. We show that in a hybrid node, the scheduling complexity with typical scheduling algorithms may be reduced to half of a node running in full packet switching mode.

The fast parallel restoration (FPR) scheme is proposed to achieve the fast setup of restoration label switched path (LSP) in the distributed optical networks. The scheme is derived by dividing the whole restoration LSP into several segments of sub-LSP and triggering each sub-LSP along the new route to finish the signaling procedure concurrently, and subsequently merging all sub-LSPs into a whole LSP. The theoretical analysis and simulation results show that the FPR scheme outperforms the other two typical restoration schemes in terms of connection setup time.

A single channel with a 160-Gb/s optical time-division-multiplexing (OTDM) transmission over 100 km is fabricated. With the help of 500-GHz optical sampling oscilloscopes, the fiber length is adjusted to the order of 10 m, which corresponds to the accuracy of 0.4 ps for the dispersion compensation. The dispersion map is optimized for the 100-km transmission link. A completely error-free transmission with the power penalty of 3.6 dB is achieved for 2 h without using forward error correction.