Design and Implementation of Initial Cell Search in 5G NR Systems

The initial cell search plays an important role during the process of downlink synchronization establishment between the User Equipment(UE) and the base station. In particular, the uncertainty of the synchronization signals on the frequency domain and the flexibility of frame structure configuration have brought great challenges to the initial cell search for the fifth-generation(5G) new radio(NR). To solve this problem, firstly, we analyze the physical layer frame structure of 5G NR systems. Then, by focusing on the knowledge of synchronization signals, the 5G NR cell search process is designed, and the primary synchronization signal(PSS) timing synchronization algorithm is proposed, including a 5G-based coarse synchronization algorithm and conjugate symmetry-based fine synchronization algorithm. Finally, the performance of the proposed cell search algorithm in 5G NR systems is verified through the combination of Digital Signal Processing(DSP) and personal computer(PC). And the MATLAB simulation proves that the proposed algorithm has better performance than the conventional cross-correlation algorithm when a certain frequency offset exists.

Cell Search Algorithm in Beyond 3G Systems

The downlink frame structure for beyond 3G mobile communication systems is presented. Beyond 3G systems utilize the OFDM technique. However, a problem encountered in OFDM is that channel dispersion destroys orthogonality between carriers, caushag inter-symbol interference. It is also sensitive to high peak to mean power ratio （PAPR）. Therefore it spends much time on obtaining frequency, time, and frame synchronization. This paper proposes to add a frame synchronization channel in the time domain to overcome the shortcoming of OFDM. As transmitter diversity improves the system performance, beyond 3G systems employ space-time block coded （STBC）. Fast cell search algorithm including slot synchronization, frame synchronization and cell ID identification is then discussed, which is based on the frame synchronization channel in transmitter diversity systems. Detection and false alarm probabilities in AWGN and Rayleigh channels are analyzed, and the mean acquisition time is obtained. Computer simulations are conducted to evaluate the performance of the cell search algorithm under different channel conditions.

A novel cell search scheme for OFDM cellular systems

A novel cell search scheme for OFDM cellular systems is proposed. It is based on one OFDM symbol with several identical slots as preamble, the time domain repetition structure of which can be utilized to accomplish OFDM timing/frequency synchronization. The cell ID is comprised of two parts: a sub-carrier mask index g and a sequence index x. Each sub-carrier mask activates or deactivates some of the sub-carriers, after which a differentially coded sequence is loaded on pairs of the adjacent active sub-carriers. The user equipment (UE) recognizes the mask with index g via power detection of the received frequency domain signal. Then it estimates the index x from differential demodulation followed by detection of the frequency domain se-quence. In order to improve the performance, a method of jointly estimating g and x is devised. Simulation results showed that the proposed scheme is able to support a very large number of cell IDs while maintaining a good performance even in bad multi-cell environment.

Cell search scheme for long-term evolution of TD-SCDMA system

Cell search is an important aspect for 3G long-term evolution （LTE）. This article deals with cell search in the time-division-synchronou code-division multiple access （TD-SCDMA） LTE system. On the basis of the synchronization channel （SCH） and cell specific reference symbols （CSRSs）, the proposed cell search procedure includes five stages： frame detection and coarse timing, coarse carrier frequency offset （CFO） estimation, fine timing, fine CFO estimation, and cell identification. The key features of the proposed method are as follows： first, the neighboring three cells＇ CSRSs are frequency division multiplexed （FD） to mitigate inter-cell interference. Second, the frequency domain differential cross-correlation, computed from CSRSs are maximally ratio combined for cell identification. Finally, the large set Kasami sequences are quadrature phase shift key （QPSK） modulated to be cell specific sequences （CSSs）, to support a large number of target cells. Simulations show that the FD method is better than the code division multiplexed （CD） method.

Cell search algorithms for the 3G long-term evolution

This article presents downlink initial synchronization and cell identification algorithms for long term evolution （LTE） of third-generation （3G） mobile communication systems, which are based on synchronization channel （SCH） and cell specific pilot symbols, respectively. The key features of the proposed scheme are： it can improve performance of the frequency synchronization through oversampling of the SCH, it can support a large number of target cells by modulating a cell-specific pilot sequence over two symbols within a subframe, and it can guarantee cell identification performance by maximally ratio combining the frequency domain differential cross-correlation. Simulations show that the proposed scheme has a potential use in 3G LTE.

Dual-CRC polar codes and dual-SCFlip decoding algorithm for cell search in 5G system

Polar codes become the coding scheme for control channels of enhanced mobile broadband(eMBB)scenarios in the fifth generation(5 G)communication system due to their excellent decoding performance.For the cell search procedure in 5 G system,some common information bits(CIBs)are transmitted in consecutive synchronization signal blocks(SSBs).In this paper,a dual-cyclic redundancy check(dual-CRC)aided encoding scheme is proposed,and the corresponding dual-successive cancellation flip(dual-SCFlip)algorithm is given to further improve the performance of polar codes in the low signal-to-noise ratio(SNR)environment.In dual-CRC aided encoding structure,the information bits of polar codes in different transmission blocks add cyclic redundancy check(CRC)sequences respectively according to CIBs and different information bits(DIBs).The structure enlarges the size of CIBs to improve the block error ratio(BLER)performance of the system.The dual-SCFlip decoder can perform bit flip immediately once CIBs is decoded completely,and then decode DIBs or terminate decoding in advance according to the CRC result,which reduces the delay of decoding and mitigates the error propagation effect.Simulation results show that the dual-CRC aided encoding scheme and dual-SCFlip decoder have significant performance improvement compared to other existing schemes with low SNR.

A robust symbol timing strategy for cellular systems

In cellular systems,establishing the initial symbol timing of potential preambles is the first step of a cell search.The envelope fluctuation of the downlink signal hinders the successful timing of conventional symbol timing methods.To solve this problem,a hybrid timing strategy is proposed with two novel detectors,namely the normalized replica-based detector and normalized differential detector.The strategy first detects all potential preambles via the normalized replica-based detector and then employs the normalized differential detector to verify the target preamble,which comes from the target cell and has the highest power.The strategy is unaffected by envelope fluctuation and has computational complexity comparable to that of conventional methods.Simu-lations and real-data tests show that the hybrid timing strategy is robust and practical for initial symbol timing.

Sequence Pair-Based Scatter Search Approach for Cell Formation and Layout Considering Unequal Machine Dimensions and Cell Input and Output Locations

In the purpose of dealing with the inconsistency of the machine dimensions in cell formation and layout problems,comprehensively taking into account such factors as alternative process routings, lot splitting, cell input and output( I / O)locations,an integrated cell formation and layout model considering unequal machine dimensions is established which aims at minimizing intra-cell material handling cost,inter-cell material handling cost and cell expansion cost. A sequence pair-based scatter search algorithm is proposed which simultaneously adopts the roulette wheel selection operator as a subsets combination method to improve the quality of new solutions and the dynamic reference sets updating method to accelerate the convergence rate. Through the comparison of the results achieved by the proposed algorithm, Lingo and simulated annealing under different problem scales,the correctness of the model and the effectiveness of the algorithm are verified.

基于SSA-ELM神经网络控制器的光伏MPPT方法

光伏电池板所处环境的非线性变化使得光伏电池的功率保持在最大功率点(maximum power point,MPP)非常困难。传统的最大功率点跟踪(maximum power point tracking,MPPT)方法普遍存在技术缺陷,无法满足当前需求。针对光伏发电MPPT问题,该文提出了一种基于麻雀搜索算法优化的极限学习机(sparrow search algorithm-extreme learning machine,SSA-ELM)神经网络控制器的MPPT方法。与传统技术相比,该MPPT方法在稳定性、速度、超调和MPP的振荡等方面的效果均较好。使用MATLAB/Simulink平台进行仿真实验,验证了所提控制策略及理论分析的正确性。

面向个性化需求的燃料电池测试台价值评估方法

针对不同个性化需求的燃料电池测试台(fuel cell test bench,FCTB)难以评价和量化评估的问题,提出一种基于改进和声搜索算法的FCTB价值评估方法.针对不同FCTB的个性化需求,建立了FCTB综合评估指标体系;结合用户的个性化需求,采用模糊层次分析法分配指标权重,构建价值定量评估模型,将权重求取问题转换为约束优化问题;提出一种改进和声搜索算法对问题进行求解,通过设计解向量生成机制和参数自适应调整策略,用于提高传统和声搜索算法的求解效率和搜索能力.仿真结果表明,本文方法在计算效率和精度方面具有优势,并能够根据不同的需求特性实现对FCTB方案做出定量的价值评估.

MC-NAS:一种可视化模块贡献神经架构搜索方法

现有的神经架构搜索方法无法直观地将网络模型与候选模块以及模型识别准确率之间的关系展示出来;同时很多NAS方法可扩展性差,无法将其搜索策略扩展至任意搜索空间。针对上述挑战,提出了一种可视化模块贡献神经架构搜索方法。提出了模块贡献这个概念,并通过分析贡献计算过程的窘境给出了任意搜索空间下的统一采样原则,利用统一的贡献度指导原则给出了不同搜索空间的贡献度计算策略。针对特定的约束条件通过动态网络规划算法生成神经网络体系结构。大量的实验结果表明该算法在任意搜索空间中的有效性。使用CIFAR-10、CIFAR-100和ImageNet16-120数据集在NAS-Bench-201基准测试上平均准确率达到了93.33%、71.07%、42.69%。

基于改进SSA-DBN的质子交换膜燃料电池水故障智能分类方法

为了实现质子交换膜燃料电池(PEMFC)系统水故障的高效快速分类,提出了基于改进麻雀搜索算法(SSA)优化深度置信网络(DBN)的PEMFC故障分类方法。采用归一化处理消除故障数据参数之间量纲不同的影响,使用核主成分分析对数据进行故障特征提取,有效地缩减了原始数据维度,降低了运算复杂度,并避免低贡献度数据对故障分类造成干扰。引入柯西-高斯变异策略改进SSA,并利用SSA对DBN进行参数寻优,确定网络结构,通过优化后的DBN实现对PEMFC水故障的快速分类。对3 000组PEMFC水故障数据进行测试,结果表明:所提方法可以快速准确地识别PEMFC的正常状态、膜干故障、水淹故障3种健康状态;总体的分类准确率为98.67%,运算时间为0.89 s,相比支持向量机、概率神经网络方法,所提方法的故障分类精度分别提升了4%、3.34%,运算时间分别减少了15.35、0.35 s。

储能用质子交换膜燃料电池长期老化预测

质子交换膜燃料电池(PEMFC)的长期老化预测有助于缩短耐久性测试时间,降低成本,为维护策略提供依据。针对超参数问题,提出一种将优化算法和储备池计算相结合的数据驱动预测方法。基于耐久性测试数据集,以电堆输出电压为老化指标,利用麻雀搜索算法(SSA)优化回声状态网络(ESN)的储备池尺寸、泄漏率和正则化系数,以构建预测模型。分别利用原始数据的前30%、40%、50%和60%作为训练集训练模型,验证模型在各训练集比例下的长期老化预测性能。训练集比例为30%时,所提方法在静态工况下的长期预测均方根误差(RMSE)达到0.0083,准动态工况下可达到0.0359。

基于Cell-DEVS的森林灭火资源调度狼群优化算法

针对森林消防调度兵力组织难、精细化程度不高的问题,提出了一种基于Cell-DEVS的森林灭火资源调度狼群优化算法。首先,利用Rothermel林火蔓延模型,对林火蔓延速度进行等级划分;其次,根据森林消防"控制要害"的原则,对森林灭火资源调度进行耦合驱动模块化建模,建立了基于Cell-DEVS模型的狼群强者生存更新机制(WSSUM);最后,针对森林灭火资源在单位时间步长局部调度搜索中不够精细的问题,提出了一种改进局部搜索策略狼群优化算法(WOA),对资源在局部游走阶段实现调度交互。与WSSUM算法相比,WOA在减少任务执行时间的同时,改善了局部搜索性能,实验结果表明:收敛速度较改进前提高了10.1%。该研究适应于配备了定位设备的单兵消防作战指挥系统,实现差异化精细调兵。

TRIZ方法在PECVD设备流热场改进中的应用

PECVD是太阳能电池加工过程中的重要设备,其内部热流场的均匀性对太阳能电池表面薄膜的均匀性影响很大。本文使用TRIZ(发明问题解决理论)对PECVD设备的热流场进行分析,提出TRIZ可作为流体仿真的重要补充。同时提出流热场问题不仅涉及流的传导、使用和分配问题,还涉及浓度不均、压力不均和工作效率低的问题,建议总结前人解决方法并将其纳入TRIZ流相关知识体系,以助于工程师参考解决类似问题。

INTEGRATION OF CELL FORMATION AND LAYOUT DESIGN IN THE UNIDIRECTIONAL LOOP MATERIAL HANDLING ENVIRONMENT

A two-phase approach is proposed to deal with the integration problem in theloop layout. Tabu search is applied to cell construction in phase 1 to minimize the inter-cell flow,and the heuristic for layout design is used as phase 2 to optimize the sum of intra-cell andinter-cell transportation cost. The final computational results demonstrate the validation of thetwo-phase approach.

TFT-LCD制造cell阶段多目标绿色调度问题研究

针对薄膜晶体管液晶显示器(TFT-LCD)制造cell阶段的多目标绿色调度问题,构建了最小化最大完工时间、总能耗和总生产成本为目标的优化模型。采用基于机器和工序的两段式编码,使用在步长因子前加入动态系数的改进布谷鸟搜索算法,结合双元锦标赛和动态淘汰制,根据聚集距离法筛选Pareto最优解来获得Pareto最优解集,对TFT-LCD制造cell阶段多目标绿色调度问题进行求解。结果表明,改进布谷鸟搜索算法优于标准布谷鸟搜索算法和带精英策略的快速非支配排序遗传算法,可以提高TFT-LCD制造cell阶段多目标绿色调度问题的求解效率和质量。

基于改进TSO优化Xception的PEMFC故障诊断

针对质子交换膜燃料电池(PEMFC)的故障诊断问题,提出了一种利用改进的瞬态搜索优化(TSO)算法优化Xception网络的故障通用诊断方法。首先,对故障数据进行线性判别分析降维和归一化处理,在保留主要特征的前提下降低计算复杂度;其次,引入Tent混沌映射和反向学习策略增强TSO算法的全局搜索能力,在训练阶段对Xception神经网络的超参数进行优化;最后,使用充分训练的Xception网络对PEMFC故障进行分类识别,并与经典的分类模型进行对比。在基于实验测量的水管理故障数据和仿真产生的多类故障数据上,Xception均取得了最高的分类准确率,分别为100%和98.08%,这表明Xception对数据特征的提取能力较强,且所提方法能作为一种PEMFC故障的通用诊断方法。

Cell—Talk手机语音搜索平台

本文介绍的是语音搜索平台Cell-Talk的主要功能以及相关软硬件架构,Cell-Talk除提供运营商多样化内容资源搜寻平台,还提供运营商内容服务管理功能以及语音辨识流程,提供给运营商一套强大完整的语音辨识搜索平台。

Development of a Web-Based Decision Support System for Cell Formation Problems Considering Alternative Process Routings and Machine Sequences

In this study, we use the respective advantages of the tabu search (TS) and the Web-based technologies to develop a Web-based decision support system (DSS) for cell formation (CF) problems considering alternative process routings and machine sequences simultaneously. With the assistance of our developed Web-based system, the CF practitioners in the production departments can interact with the systems without knowing the details of algorithms and can get the best machine cells and part families with minimize the total intercellular movement wherever and whenever they may need it. To further verify the feasibility and effectiveness of the system developed, an example taken from the literature is ado- pted for illustrational purpose. Moreover, a set of test problems with various sizes drawn from the literature is used to test the performance of the proposed system. Corresponding results are compared to several well-known algorithms previously published. The results indicate that the proposed system improves the best results found in the literature for 67% of the test problems. These show that the proposed system should thus be useful to both practitioners and researchers.