高性能Ed25519算法硬件架构设计与实现

针对签名验签速度难以满足特定应用领域需求的问题,该文设计了一种高性能Ed25519算法的硬件实现架构.采用宽度为2 bit的窗口法实现标量乘运算,减少了标量乘所需的总周期数;通过优化点加倍点操作步骤,提高了乘法器的硬件使用率;使用低计算复杂度的快速模约简实现模乘,提高了整体运算速度.为了使模L运算可复用标量乘中的快速模...     

基于混合遗传算法的SAR图像边缘检测

在研究SAR图像特性的基础上,提出了一种基于混合遗传算法的边缘检测方法.它将边缘检测归结为一个优化问题,利用遗传算法的全局寻优能力,能够搜索到定位准确,更加连续和精确的边缘,并且局部优化算子又能大大加快算法的收敛,因而算法兼有良好的检测效果和快速的处理速度.仿真结果也证明了其有效性.     

基于OpenCV的集成芯片基板定位技术

针对集成芯片制造中对定位、校准的高精度、实时性要求,提出了用机器视觉技术解决芯片基板定位的方法,通过对几种模板匹配算法的研究,采用基于OpenCV的图像分析技术实现了对集成芯片基板的准确定位,解决了传统机械定位精度低、速度慢的问题。     

Design of a low power wide-band high resolution programmable frequency divider

The design of a high-speed wide-band high resolution programmable frequency divider is investigated. A new reloadable D flip-flop for the high speed programmable frequency divider is proposed. It is optimized in terms of propagation delay and power consumption as compared with the existing designs. Measurement results show that an all-stage programmable counter implemented with this D flip-flop using the Chartered 0.18 /spl mu/m CMOS process is capable of operating up to 1.8 GHz for a 1.8 V supply voltage and a 5.8-mW power consumption. By using this counter, an ultra-wide range high resolution frequency divider is achieved with low power consumption for 5-6-GHz wireless LAN applications.     

多智能体强化学习在足球机器人中的研究与应用

 本文提出一种基于投票的多智能体强化学习方法,使球队在比赛中学会协作,自动适应环境,提高实时性和进球数.首先通过定义称为策略的联合行为,将协作问题转化为对策略的学习,简化问题的处理;然后对球场进行划分,以区域表示位置,有效减少了状态空间维数,加快了学习速度;接下来通过区分环境状态并只考虑协作状态,减小状态空间,进一步提高了学习速度;并使用投票的方式综合各个队员的决策,达到协作的目的.最后通过实验结果表明了该方法的正确性和有效性.     

一种无接缝纹理映射算法的并行实现

随着多核处理器的问世,并行编程变得简单和必要。通过对一种纹理映射算法的研究,对其耗时部分进行并行化,以加快算法运行速度。该纹理映射算法先建立三维网格顶点和视点之间的对应关系,然后利用重心坐标权重插值的颜色填充方法构造出一个无接缝的纹理图像。通过利用算法过程中存在的并行性,对算法进行并行实现。实验结果显示,能够得到一个无接缝的纹理图像,而且在双核处理器上的平均加速比达到了1.70。     

分层实体制造激光头切割路径的建模与优化

分层实体制造(LOM)技术中分层制造时间是由该层的切割速度与切割路径确定的，当切割工艺参数(如：切割功率、切割速度)确定之后，每一层制造的时间是由该层切割(扫描)路径确定的。因此优化切割(扫描)路径对提高成型效率有重要意义，而分层实体制造技术中激光切割路径优化的实质是空行程路径的优化。建立了切割路径空行程路径优化的数学模型。由于求解该模型的复杂性，采用了分级规划的两个分步算法：首先用改进的最近邻域算法求解轮廓边界线上的切割起点，然后当切割点确定后把路径优化问题归结为旅行售货员问题(TSP)，采用了高效的智能仿生算法一蚁群系统算法来求解。运行结果表明，该算法显著缩短了分层制造中的空行程，提高了快速原型制造的效率。     

Selective body biasing for post-silicon tuning of sub-threshold designs: A semi-infinite programming approach with Incremental Hypercubic Sampling

Sub-threshold designs have become a popular option in many energy constrained applications. However, a major bottleneck for these designs is the challenge in attaining timing closure. Most of the paths in sub-threshold designs can become critical paths due to the purely random process variation on threshold voltage, which exponentially impacts the gate delay. In order to address timing violations caused by process variation, post-silicon tuning is widely used through body biasing technology, which incurs heavy power and area overhead. Therefore, it is imperative to select only a small group of the gates with body biasing for post-silicon-tuning. In this paper, we first formulate this problem as a linear semi-infinite programming (LSIP). Then an efficient algorithm based on the novel concept of Incremental Hypercubic Sampling (IHCS), specially tailored to the problem structure, is proposed along with the convergence analysis. Compared with the state-of-the-art approach based on adaptive filtering, experimental results on industrial designs using 65 nm sub-threshold library demonstrate that our proposed IHCS approach can improve the pass rate by up to 7.3× with a speed up to 4.1×, using the same number of body biasing gates with about the same power consumption.     

基于Bloom filter的多模式匹配引擎

基于Bloom filter,结合位拆分状态机设计了一种适合硬件实现的多模式匹配引擎,由bloom filter过滤出可疑字符,位拆分状态机进行精确匹配.提出了过滤引擎和精确匹配引擎的流水线连接结构,通过增加分配器、缓存等硬件单元解决两引擎处理速度不匹配的问题,利用引擎的并行处理达到较高的吞吐性能.还通过设定规则长度等简化设计使引擎在保持高吞吐量的同时减小资源占用量,提高了可扩展性.     

A fast search algorithm for vector quantization usingL2-norm pyramid of codewords

Vector quantization for image compression requires expensive encoding time to find the closest codeword to the input vector. This paper presents a fast algorithm to speed up the closest codeword search process in vector quantization encoding. By using an appropriate topological structure of the codebook, we first derive a condition to eliminate unnecessary matching operations from the search procedure. Then, based on this elimination condition, a fast search algorithm is suggested. Simulation results show that with little preprocessing and memory cost, the proposed search algorithm significantly reduces the encoding complexity while maintaining the same encoding quality as that of the full search algorithm. It is also found that the proposed algorithm outperforms the existing search algorithms.     

Multiplaten Z-buffer algorithm and its application to multiplescattering

The authors present a new algorithm: the multiplaten Z-buffer algorithm. By employing a geometry based on this algorithm we can speed up the process of ray tracing, which is used to calculate the multiple scattering of a complex object. Comparison with the experimental results shows this algorithm to be effective     

An accelerated decomposition algorithm for robust support vector Machines

This paper proposes an accelerated decomposition algorithm for the robust support vector machine (SVM). Robust SVM aims at solving the overfitting problem when there is outlier in the training data set, which makes the decision surface less contoured and results in sparse support vectors. Training of the robust SVM leads to a quadratic optimization problem with bound and linear constraint. Osuna provides a theorem which proves that the Standard SVM's quadratic programming (QP) problem can be broken down into a series of smaller QP subproblems. This paper derives the Kuhn-Tucker condition and decomposition algorithm for the robust SVM. Furthermore, a pre-selection technique is incorporated into the algorithm to speed up the calculation. The experiment using standard data sets shows that the accelerated decomposition algorithm makes the training process more efficient.     

基于支持向量约简的快速目标检测

支持向量机(SVM)由于其出色的泛化能力,已成为目标检测领域应用最为广泛的分类器之一.然而在检测过程中,过多的支持向量会产生很大的时间开销,从而降低目标检测系统的实时性.针对此问题,提出一种约简支持向量的方法,以降低分类器的决策开销,加快检测速度.此方法采用迭代的方式来估计特征空间中向量的原像,通过构建精简原像集来简化支持向量机,从而达到了提升分类速度的效果.利用精简的SVM结合Selective Search+ BoW模型构建了一款快速检测器,测试结果表明:该检测器能够在保证检测率的前提下,通过约简支持向量,提高目标检测的实时性.     

卡尔曼滤波在跟踪运动目标上的应用

在多关节机器人跟踪运动目标过程中,图像处理的主要内容包括图像的识别和目标轨迹的判定,需要在整个视觉空间进行搜索和匹配。这是一个典型的全局搜索问题,耗时较长、实时性较差,难以实现运动目标的快速跟踪。针对多关节机器人跟踪运动目标过程中存在的彩色图像数据量大、处理时间长,从而影响实时性的问题,采用线性卡尔曼滤波将全局搜索转化为局部搜索,减少搜索空间、加快匹配速度。通过实验可知,匹配时间由原来的157 ms降低到20 ms。实验结果证明了该方法的有效性。     

Computer-aided design tool for electromagnetic sensors - [instrumentationnotes]

Numerical simulation is an invaluable tool for the designer of any device. By providing a model of the sensor within the design parameters, simulation can substantially speed up and simplify the design process and give a reliable estimation. The use of optimization algorithms can lead to much better sensor characteristics and reduce sensor costs and operation.     

一种基于集群计算的软件无线电接收系统实现

提出并实现了一种基于集群计算的实时软件无线电接收系统,对系统的硬件架构、软件组成和工作原理进行了详细描述.通过采用并行网络数据分发、高效负载均衡调度算法、两级并行计算等关键技术,系统数据流量最高可达200MB/s,最大处理信号带宽可达40MHz.实际运行表明,系统能够完成PSK、QPSK、8PSK、16PSK、mQAM等多种高速无线通信信号的实时接收和处理.     

Fast encoding techniques for Multiview Video Coding

Multiview Video Coding (MVC) is a technique that permits efficient compression of multiview video. MVC uses variable block size motion and disparity estimation for block matching. This requires an exhaustive search process that involves all possible macroblock partition sizes. We analyze the time complexity of MVC and the methods that have been proposed to speed up motion and disparity estimation. We then propose two new methods: Previous Disparity Vector Disparity Estimation (PDV-DE) and Stereo-Motion Consistency Constraint Motion and Disparity Estimation (SMCC-MDE). PDV-DE exploits the correlation between temporal levels and disparity vectors to speed up the disparity estimation process while SMCC-MDE exploits the geometrical relationship of consecutive frame pairs to speed up motion and disparity estimation. We build a complete low complexity MVC encoding solution that combines our two methods with complementary previous methods to speed up motion and disparity search. We evaluate the complexity of our solution in terms of encoding time and number of search points. Our experimental results show that our solution can reduce the encoding time and number of search points of the standard MVC implementation (JMVM 6.0) using the fast TZ search mode up to 93.7% and 96.9%, respectively, with negligible degradation in the rate-distortion performance. Compared to the best published results, this is an improvement of up to 11% and 7%, respectively.     

一种可配置的CNN协加速器的FPGA实现方法

针对卷积神经网络中卷积运算复杂度高而导致计算时间过长的问题,本文提出了一种八级流水线结构的可配置CNN协加速器FPGA实现方法.通过在卷积运算控制器中嵌入池化采样控制器的复用手段使计算模块获得更多资源,利用mirror-tree结构来提高并行度,并采用Map算法来提高计算密度,同时加快了计算速度.实验结果表明,当精度为32位定点数/浮点数时,该实现方法的计算性能达到22.74GOPS.对比MAPLE加速器,计算密度提高283.3%,计算速度提高了224.9%,对比MCA（Memory-Centric Accelerator）加速器,计算密度提高了14.47%,计算速度提高了33.76%,当精度为8-16位定点数时,计算性能达到58.3GOPS,对比LBA（Layer-Based Accelerator）计算密度提高了8.5%.     

A general framework to construct stationary mobility models for the simulation of mobile networks

Simulation has become an indispensable tool in the design and evaluation of mobile systems. By using mobility models that describe constituent movement, one can explore large systems, producing repeatable results for comparison between alternatives. In this paper, we show that a large class of mobility models - including all those in which nodal speed and distance or destination are chosen independently - have a transient period in which the average node speed decreases until converging to some long-term average. This speed decay provides an unsound basis for simulation studies that collect results averaged over time, complicating the experimental process. In this paper, we derive a general framework for describing this decay and apply it to a number of cases. Furthermore, this framework allows us to transform a given mobility model into a stationary one by initializing the simulation using the steady-state speed distribution and using the original speed distribution subsequently. This transformation completely eliminates the transient period and the decay in average node speed and, thus, provides sound models for the simulation of mobile systems.     

事件驱动的动态武器目标分配研究

动态武器目标分配（Weapon Target Assignment,WTA）问题是军事运筹学研究的重要理论问题,也是防空作战中迫切需要解决的问题。根据防空作战中事件对动态WTA问题的影响。提出了基于事件驱动的动态WTA的改进蚁群算法。通过处理事件,动态地调整蚁群算法的参数,使武器目标分配方案反映事件的影响。并根据当前最...