Design of integrated data/voice stations for a CSMA/CD LAN

This paper describes the design and implementation of integrated data/voice stations for a LAN environment. The stations are based on the DB16000 Development Board using a 16-bit CPU. Continuously variable slope delta modulation is selected as the voice encoding technique. Protocols were designed and implemented to handle data and voice separately. Prototypes were built to test the station's performance. The stations communicate with one another at a rate of 1.5 Mbit/s. The voice digitization rate is 16 kbit/s and the terminal speed is 9.6 kbit/s.     

Buffer analysis of an integrated voice and data terminal

Current research activities in computer networks tend to emphasize the integration of services such as voice and data by designing networks that take into account the performance requirements of these services. In some cases, however, design and implementation of a totally new network may not be feasible. An alternative approach in such cases would be to integrate the services at the terminal level. Such an integrated terminal could also be widely used in future office automation systems. The performance of such an integrated terminal is assessed with the assumption that the terminal is connected to a voiceband channel. The analysis indicates the performance limits on the application imposed by the buffer size at the terminal and the large fluctuation of delay performance for data traffic. A window-type flow control for data traffic is also considered. By appropriately choosing the buffers in the terminal and window size, both voice and data traffic service requirements can be met.     

An X.25-compatible protocol for packet voice communications

Developments in packet-switched networks have generated interest in how to integrate voice and data traffic in such networks. The paper deals with a proposed communications protocol for voice transmission within X.25 packet-switched networks. Basic protocol procedures for handling voice communications are examined, and possible formats for the identified types of voice frame and packet fitting for current CCITT standards governing packet data communications are presented.     

基于交通流特性的高速公路交通网络建模与应用

为了满足高速公路安全管理系统的功能需求,提出了一种适合高速公路交通网络数据表达、操作与分析的基于交通流特性的高速公路交通网络数据模型,弥补了现有交通网络GIS数据模型在复杂的网络建模和分析中的不足。在分析高速公路交通流特性的基础上提出了高速公路交通网络的建模要素,并设计了概念建模、逻辑建模和物理建模的方法。该模型有利于迅速响应影响网络交通流分布的突发事件并生成安全管制决策,文中以高速公路网交通封闭实施位置算法为例说明其在高速公路安全管理领域的应用。     

交通流冗余数据识别和约简方法

针对交通检测器检测到的数据存在冗余现象、影响后续决策并需要进行约简的问题,提出了一种冗余数据的识别和约简方法。采用等级分组法实现对冗余数据的识别,先通过等级法计算每个交通参数的权值并按照分组思想,将大数据集分割成许多不相交的小数据集,在各个小数据集中识别冗余数据。为避免漏查,选择其他关键参数多次重复识别。识别出的冗余数据采用平均法约简。实例验证表明,等级分组法识别冗余数据具有较好的精度,随着阈值的增加,查准率和查全率减小,但仍在93%以上;同时采用平均法约简,拟合度较高,达到0.938。可见采用的冗余数据识别和约简方法能够有效地解决单数据源数据冗余问题。     

交通流异常数据检测研究及实证

针对道路交通系统实时交通流数据普遍存在的异常现象,提出一种基于最小二乘支持向量机的交通异常数据检测方法。运用相空间重构技术,将时间序列数据转换为矢量数据。运用训练数据构建基于最小二乘支持向量机的回归估计模型,通过计算实际值与预测值之间的残差来判别异常数据。以成都市一环路北二段进行实证研究,并与传统检测方法比较分析,结果证实该检测方法的有效性。     

基于城市交通大数据的车辆类别挖掘及应用分析

实时城市交通监控已成为现代城市管理的一个重要组成部分，视频监控采集的交通大数据在城市管理和交通控制方面得到了越来越多的应用；然而，全城范围内庞大的监控交通大数据还鲜少用于城市交通及城市计算研究。在一个省会城市全城范围内的监控交通大数据上展开了车辆类别挖掘及应用分析研究。首先，定义了周期性私家车、类出租车和公共通勤车三种对城市交通具有重要影响的车辆类别，将车辆类别定义与频繁序列模式挖掘算法相结合提出了相应的挖掘方法。在济南市一周1704个视频监测点，1.2亿次车辆记录数据上，验证了所提定义及挖掘方法的有效性；其次，以4个居民小区为例挖掘分析了居民出行的交通方式及与周围兴趣点（POI）分布关系，此外，还探索了城市交通大数据与POI相结合在城市规划、需求预测和偏好推荐方面的应用潜能。     

基于WSNs的语音通信机制设计与实现

针对目前无线传感器网络( WSNs)对语音传输的不足,设计和实现了一种基于WSNs的无线语音通信机制.采用了结合8051CPU和RF收发器的CC2530芯片作为微处理器,CMX649作为语音编解码芯片,并采用协处理器辅助CC2530对语音信息进行处理,在lEEE 802. 15. 4协议上实现对语音的传输.详细介绍了系统的硬件和软件设计,对通话距离与通话质量进行了测试.实验结果表明:设计的语音通信机制具有较远的传输距离,能达到较好的语音传输效果,满足实际对语音通信的要求.     

单源交通流数据质量控制算法设计与程序实现

针对基于数值计算理论的单源交通流数据质量控制算法作程序实现。首先进行程序流程设计,包括数据读取模块、数据识别模块、数据修正模块、数据存储模块和数据输出模块五部分,各模块间相互联系,按照数据流的方向结合在一起按顺序使用。其次,给出了详细的总体程序工作流程,并且针对其中的两个核心模块给出了具体的工作流程。最后,针对北京市实时交通流数据进行了检验,结果表明程序设计实现了算法设计的要求,并且极大改进了质量控制的自动化程度,提高了质量控制的工作效率。     

基于WiFi的语音矿灯设计

针对现有矿灯大多只具有照明、定位、环境感知等功能,没有语音对讲功能的问题,设计了一种基于WiFi的具有语音对讲功能的语音矿灯。该语音矿灯以工业以太环网和WiFi网络为传输平台,采用VoIP语音通信技术实现语音播放、音频采集、与调度台对讲功能;通过音频编解码芯片实现语音模拟信号与数字信号的转换,采用UDP协议将信号传输至调度台,完成语音数据的双向传输,实现语音对讲和矿灯照明一体化。详细介绍了语音对讲功能实现的关键技术:音频数据的编码格式和缓存管理、语音数据的可靠传输机制,用于确保语音播放的准确性;WiFi模块与微控制器STM32L151的低功耗休眠技术,用于降低语音矿灯平均电流,延长工作时间。测试结果表明:该语音矿灯能够满足调度台与井下工作人员之间的语音对讲需求,与WiFi基站通信距离可达400 m,与调度台之间的对讲传输时延小于1 s,语音矿灯之间的组播传输时延小于3 s;语音矿灯对讲时平均电流小于70 mA,空闲时平均电流小于5 mA。     

数据挖掘算法在交通状态量化及识别的应用

在智能交通系统（ITS）环境下,以交通检测器采集的海量交通流信息为对象,通过数据挖掘技术即数据获取、数据预处理、挖掘方法、结果分析与评价、模式应用等进行新的信息提取,提出了各阶段的要求和聚类分析及模式识别的算法,最后从海量数据中得到新的有用信息交通状态分类,同时使用实时采集交通流数据进行交通状态判别。实验结果表明识别状态能够准确反映实际交通状态。     

Congestion control technique for an integrated services local area network

A congestion control scheme for implementation in an integrated services local area network is proposed. An outline is given of simulation studies that have been carried out to demonstrate the feasibility and effectiveness of such a scheme. The way that these studies can help identify the volume of real-time traffic that can be accommodated with an acceptable level of blocking is explained.     

Performance analysis of customer access to an ISDN

A simple analytical model has been developed to quantify the performance encountered by voice and data traffic during customer access to an ISDN. This enables a direct comparison of traffic capabilities of various proposed implementation schemes currently under consideration by telecommunications administrations.     

Analysis of the effect of security on data and voice traffic in WLAN

This paper investigates the impact of security on the performance of WLAN. More specifically, it analyzes the impact of different encryption techniques used by two security protocols, namely Wired Equivalent Privacy (WEP) and Wi-Fi Protected Access (WPA) on the throughput and delay over WLAN IEEE 802.11g. Moreover, in this paper, we investigate the effect of encryption on the performance of wireless transmission during the handover process when a mobile device disassociates from one access point and re-associates with another one. In addition, this paper addresses the impact of the use of security in WLAN specifically on voice traffic namely the incurred transmission delay and jitter.     

Design of a gateway for VBR voice traffic over N-ISDN and B-ISDN networks

An Internetworking Function (IWF) carrying VBR voice traffic between N-ISDN and B-ISDN networks will be considered. Based on the way the partially filled cells problem is addressed, two techniques governing the operation of the IWF will be proposed. In the first technique, bit stuffing will be considered, while in the second, delaying these unfilled cells will be suggested. It is assumed that in both cases, speech silence detection is employed. As an evaluation measure, the bandwidth utilization, the cell loss, the percentage of stuffing, and the delay performance will be computed. Results will be then presented and analyzed, followed by conclusions and suggestions.     

Voice/data integration on Ethernet — backoff and priority considerations

Astudy of the CSMA/CD protocol as given in the Ethernet specifications, for real time (voice in particular) and mixed voice/data environments is described. The objectives of the study were to predict the Ethernet's behaviour under the current protocols, to point out potential improvements within the realm of the current specifications, and to improve understanding of the network behaviour in these environments. Results are given in terms of relative performance of the various system measures given as a function of system and traffic models. It is shown that the support provided to real time voice on Ethernet strongly depends on the assumptions made with regard to the coexisting data user's traffic and that with certain data traffic patterns the observed voice delay is much worse than previously reported. Finally, it is shown that without violating the Ethernet network specification, voice service can be improved by simple adjustments of the Ethernet retransmission ‘backoff’ algorithm.     

An improved method for voice pathology detection by means of a HMM-based feature space transformation

This paper presents new a feature transformation technique applied to improve the screening accuracy for the automatic detection of pathological voices. The statistical transformation is based on Hidden Markov Models, obtaining a transformation and classification stage simultaneously and adjusting the parameters of the model with a criterion that minimizes the classification error. The original feature vectors are built up using classic short-term noise parameters and mel-frequency cepstral coefficients. With respect to conventional approaches found in the literature of automatic detection of pathological voices, the proposed feature space transformation technique demonstrates a significant improvement of the performance with no addition of new features to the original input space. In view of the results, it is expected that this technique could provide good results in other areas such as speaker verification and/or identification.     

基于云模型的动态交通数据流软划分算法

提出了一种交通数据流软划分算法,该算法利用STREAM算法对交通数据流进行了聚类分析,得到了能够反映交通状况不同特征的聚类结果,然后对聚类结果进行了数据挖掘和交通数据流趋势预测。最后在数据流值预测结果的基础上,采用基于云模型划分的算法对交通的预测流值进行分析,得到了更加灵活的控制策略。     

流量测量的关键技术分析与研究*

流量测量在流量计费、网络资源优化和异常检测等方面有广泛的应用。针对传统的流量测量模型缺乏可扩展性的缺点,提出了一种适用于高速网络的可扩展的流量测量模型。该模型采用报文分批处理的思想引入了两级缓冲区结构,使得缓存报文和流量统计两个过程同时进行。基于此模型,抽象了流量测量的两大关键技术,即流量抽样测量技术和概要数据结构。运用此模型进行高速网络的流量测量,不仅会降低资源的需求、提高分析的速度,而且还不会失去准确性。     

基于城市交通监控大数据的工作位置推理方法

基于时空数据的用户位置推理在产品推荐、精确营销、交通调度及城市规划等实际应用中有着重要的作用,然而,基于城市交通监控数据的位置推理问题尚未被探索,因此,提出了一种面向稀疏摄像头交通监控数据的工作位置推理方法。首先,收集了路网、兴趣点（POI）等城市交通外围数据,并通过路网匹配的预处理方式获取到了一个含有摄像头、POI等丰富语义信息的真实路网;其次,通过聚类车辆轨迹中所提取的起点-终点（O-D）对来获得车辆重要的停留区域,即候选工作区域;之后,利用所提的in/out访问时间模式的约束,从多个候选区域中匹配出最大可能的工作区域;最后,利用所获取的路网信息和路网周中POI的分布信息提取出车辆的可达POI集合,从而进一步缩小车主的工作位置范围。在一个省会城市真实的交通监控数据集上的综合实验评估和案例分析验证了所提方法的有效性。