Capacity and Protocols for Integrated Voice/Data CDMA Systems

该项目的目的是开发一种分析框架，为话音/数据集成CDMA系统中话音和数据的交互提供新的理解。从这项研究中获得的见解将进一步用于开发新的算法。本文将分析这种多媒体无线系统中数据的可变性及其对语音业务容量和性能的影响。针对延迟敏感语音优先的数据业务，提出了新的易于实现的实时呼叫接纳和呼叫建立后控制策略。不同类别的服务可能需要的可变服务质量（QoS）的影响将包括在分析框架和方法中。此外，协议层之间的交互将包括在数据访问控制方案的研究和设计中。具体来说，我们将分析端到端协议（如TCP（传输控制协议））与无线电链路上的数据访问控制方案之间的相互作用。分析方法：分析这些问题遇到两个严重的困难。即使所有的信息（例如，相关随机变量的分布）是可用的，精确的分析似乎超出了目前的理论，导致计算易于处理的解决方案。其次，所需的发行版通常不可用。相反，我们开发了简单、合理准确的分析方法，在适当的时候只使用均值和方差，在必要的时候只使用分布假设。数据控制策略：认识到语音对延迟敏感，我们开发了数据访问控制策略。一种类型的实时控制基于与语音质量相关的易于测量的数量，并使用向数据用户广播的许可概率。该提案将调查与数据控制算法优化相关的各种问题。我们将为传统和多用户接收器开发和分析这种控制策略。协议层之间的交互：当协议层是独立设计的，但最终产生了意想不到的交互时，可能会出现严重的问题。我们在TCP下运行的数据访问控制策略的背景下解决这个问题。我们研究了由于这种相互作用而产生的问题，并根据这种相互作用进行了适当的修改。项目意义：该项目将对语音/数据集成的复杂问题提供更全面的理解。这将包括发展便于使用并与预期可获得的有限资料相一致的分析方法。将为传统和多用户接收机开发新的有效实时算法。此外，研究数据链路重传方案与更高层可靠性协议（如TCP）之间的相互作用，应该为系统设计提供有价值的见解，以实现不同流量的最佳共存。因此，这项工作将更好地理解具有不同流量特征和不同QoS要求的服务类型之间以及不同协议层之间的复杂交互。这将导致改进数据实时算法的设计。

The aim of this project is to develop an analytical framework that provides new understanding of the interaction of voice and data in an integrated voice/data CDMA system. The insights obtained from this study will be further used to develop new algorithms. The effect of the variability of data and its impact on the capacity and performance of the voice service in such a multimedia wireless system will be analyzed. New easily implemented real-time strategies for call admission, and control after call setup will be derived for the data service with delay sensitive voice being given priority. The effect of variable quality of service (QoS) the different classes of service may require will be included in the analysis framework and methodology. Further, the interaction between protocol layers will be included in the study and design of access control schemes for data. Specifically, we will analyze the interactions of end-to-end protocols like TCP (Transmission Control Protocol) with data access control schemes on the radio link. Analytical Methodology: Analysis of these problems encounters two serious difficulties. Even if all the information (e.g., distributions of relevant random variables) was available, exact analysis appears to be beyond current theory leading to computationally tractable solutions. Secondly, the needed distributions are often not available. We develop instead simple, reasonably accurate, analytic methodology which uses only means and variances, when appropriate, and distributional assumptions when necessary. Data Control Strategies: Recognizing that voice is delay sensitive, we develop data access control strategies. One type of real-time control is based on readily measurable quantities related to voice quality and uses broadcasted permission probabilities to the data users. The proposal will investigate the varied questions associated with the optimization of the data control algorithms. We will develop and analyze such control strategies for both conve ntional and multiuser receivers. Interaction between Protocol Layers : Serious problems can arise when protocol layers are designed independently but end up with unexpected interactions. We address this in the context of a data access control strategy operating under TCP. We investigate the problems due to such interactions and appropriate modifications in the light of the interactions. Significance of this Project: This project will provide a more complete understanding of the complex issues of voice/data integration. This will include development of analytical methodology which is convenient to use and consistent with the limited information expected to be available. New effective real-time algorithms will be developed for both conventional and multiuser receivers. Further, the study of the interactions between data link retransmission schemes and higher layer reliability protocols (like TCP) should provide valuable insights into system design for optimum coexistence of diverse traffic. The work will thus provide a better understanding of the complex interactions both between service types with different traffic characteristics and different QoS requirements, and between different protocol layers. This will lead to improved design of real-time algorithms for data.