线性分组码的神经网络译码

介绍了神经网络与线性分组码之间的关系，并在文献[1]的基础上证明了软判决译码与求解能量函数最大值之间的等价性，然后以(7，4，3)汉明码为例介绍了神经网络在循环码硬判决、软判决译码中的应用。     

Location and calculation-free node-scheduling schemes in large wireless sensor networks

In wireless sensor networks that consist of a large number of low-power, short-lived, unreliable sensors, one of the main design challenges is to obtain long system lifetime without sacrificing system original performance (sensing coverage and sensing reliability). To solve this problem, one of the potential approaches is to identify redundant nodes at the sensing interface and then assign them an off-duty operation mode that has lower energy consumption than the normal on-duty mode. In our previous work [J. Wireless Commun. Mobile Comput. 3 (2003) 271; Processing of ACM Wireless Sensor Network and Application Workshop 2002, September 2002], we proposed a node-scheduling scheme, which can provide a 100% sensing coverage preservation capability. This, however, requires each node to be aware of its own and its neighbors’ location information. Also, in that scheme, each node has to do accurate geometrical calculation to determine whether to take an off-duty status. In this paper, we propose and study several alternative node-scheduling schemes, which cannot completely preserve the original system coverage, but are nonetheless light-weighted and flexible compared with the previous one. Our simulation results compare these schemes with the previous one and demonstrate their effectiveness.     

无线网状网技术与应用

无线网状网(WMNs)由网状路由器节点和客户机节点组成,其中的网状路由器节点组成了无线网状网的网络骨干,其移动性很小。他们一起为无线网状网和其他常规无线网络的客户机节点提供网络的无线接入。WMNs技术结合了中心式控制的蜂窝网与分布式控制的无线自组织网的优点,可有效克服这两种技术的缺陷并显著提高无线网络的性能,已经成为下一代无线通信网络的研究热点之一。WMNs可为无线个域网、局域网、校园网、城域网的一系列应用提供高速无线宽带接入服务。虽然目前WMNs技术发展很快,但其协议栈各层仍存在许多有待研究的课题。首先简要介绍了无线网状网的结构与特点;随后重点分析了其主要的几个应用领域;最后探讨了WMNs各协议层的研究现状与关键技术,并分析了该技术存在的问题及未来的研究方向。     

The Role of Network Topologies in the Optical Core of IP-over-WDM Networks with Static Wavelength Routing

In this paper, we present a performance analysis of network topologies for the optical core of IP-over-WDM networks with static wavelength routing. The performance analysis is focused on regular degree four topologies, and, for comparison purposes, degree three topologies are also considered. It is shown that the increase of the nodal degree from three (degree three topology with smallest diameter) to four (degree four topology with smallest diameter) improves the network performance if a larger number of wavelengths per link is available. However, the influence of wavelength interchange on the nodal degree gain is small. The performance of regular degree four topologies with smallest diameter is also compared with the performance of mesh–torus topologies (which are also degree four topologies), and it is shown that the blocking probability of degree four topologies with smallest diameter is about two orders of magnitude lower than the blocking probability of mesh–torus topologies. It is also presented a performance comparison of WDM-based networks with nodal degrees ranging from two to five and it is shown that the increase of the nodal degree from two to three leads to high nodal degree gains, while de increase of the nodal degree from four to five leads to low nodal degree gains. These results show that degree three and degree four topologies are very attractive for use in the optical core of IP-over-WDM networks.     

弹性分组环技术

弹性分组环技术是一种新兴的城域网技术。作者主要讨论了弹性分组环技术的特点和优势,并且简要介绍了弹性分组环技术目前在国内城域网建设中的一些应用。弹性分组环技术集以太网的经济性、IP的智能化、光纤环网的丰富带宽和可靠性于一体,是目前宽带IP城域网运营商的最佳组网方案。     

Nanoweb Formation: 2D Self‐Assembly of Semiconductor Gallium Oxide Nanowires/Nanotubes

This paper reports a method to produce networks of crystalline gallium oxide comprised of one‐dimensional (1D) nanostructures. Because of the unique arrangement of wires, these crystalline networks are termed as ‘nanowebs’. Nanowebs are of great technological interest since they contain wire densities of the order of 10⁹ cm^–2. A possible mechanism for the fast self‐assembly of crystalline metal oxide nanowires involves multiple nucleation and coalescence via oxidation–reduction reactions at the molecular level. The preferential growth of nanowires parallel to the substrate enabled them to coalesce into regular polygonal networks. The individual segments of the polygonal network consist of both nanowires and nanotubules of β‐gallium oxide. Individual wire properties contribute to a nanoweb’s overall capacity and the implications for devices based on nanowebs are expected to be enormous.     

Positively Thermo‐Sensitive Monodisperse Core–Shell Microspheres

In this paper, we report on a novel family of monodisperse thermo‐sensitive core–shell hydrogel microspheres that is featured with high monodispersity and positively thermo‐responsive volume phase transition characteristics with tunable swelling kinetics, i.e., the particle swelling is induced by an increase rather than a decrease in temperature. The microspheres were fabricated in a three‐step process. In the first step, monodisperse poly(acrylamide‐co‐styrene) seeds were prepared by emulsifier‐free emulsion polymerization. In the second step, poly(acrylamide) or poly[acrylamide‐co‐(butyl methacrylate)] shells were fabricated on the microsphere seeds by free radical polymerization. In the third step, the core–shell microspheres with poly‐ (acrylamide)/poly(acrylic acid) based interpenetrating polymer network (IPN) shells were finished by a method of sequential IPN synthesis. The proposed monodisperse core–shell microspheres provide a new mode of the phase transition behavior for thermo‐sensitive “smart” or “intelligent” monodisperse micro‐actuators that is highly attractive for targeting drug delivery systems, chemical separations, sensors, and so on.     

自适应RBF神经网络在CDMA移动通信上行链路功率控制中的应用研究

提出了一种自适应RBF神经网络功率控制方案。详细研究了该网络在DS-CDMA通信中，进行上行链路闭环功率控制(基于信扰比(SIR))的应用理论，给出了该网络参数的计算方法。最后用计算机仿真法模拟出该控制器的运行性能。结果表明基于SIR的自适应RBF神经网络功率控制器能自适应地调整移动台的发射功率，使基站接收信号的信扰比始终非常接近于一个常数，且有比定步长功率控制更小的SIR跟踪误差，从而可以降低接收信号的中断概率、提高信道容量。     

Experiments of Fast Learning with High Order Boltzmann Machines

This work reports the results obtained with the application of High Order Boltzmann Machines without hidden units to construct classifiers for some problems that represent different learning paradigms. The Boltzmann Machine weight updating algorithm remains the same even when some of the units can take values in a discrete set or in a continuous interval. The absence of hidden units and the restriction to classification problems allows for the estimation of the connection statistics, without the computational cost involved in the application of simulated annealing. In this setting, the learning process can be sped up several orders of magnitude with no appreciable loss of quality of the results obtained.     

上海市电话局用户接入网发展状况

上海市电话局用户接入网建设经历了调研、网络规划、启动实施和进一步完善、规范四个怅阶段、现已基本形成了覆盖政府机关、大专院校、医院、商务大楼及高层建筑、新建住宅小区、区级以上图书馆等范围的用户光纤网络，为加快经济建设步伐、满足多层次用户需求、拓展电信新业务打下了坚实基础。