自适应分组长度策略获得最低延迟

发布时间：2018-06-26 18:09

本文选题：K-PK策略 + 端到端延迟　；参考：《计算机应用研究》2017年06期

【摘要】：无线网络容易受到干扰,信道质量波动大。为了保证实时性要求较高应用的延迟要求,减少分组头部开销,通过对网络层以及物理层的联合优化提出了基于符号个数的自适应分组长度策略(K-PK策略)。该策略是将泊松分布到达的符号流中连续到达的k个符号聚合,添加固定长度的分组头部形成分组。仿真与分析表明,在一定信道误码率条件下,存在一个最优k值能获得最低延迟,平衡分组头部开销。K-PK策略相比基于固定时隙组包策略(T-PK策略)更能适应信道质量比较差以及波动大的情况,更符合现实的需要。
[Abstract]:Wireless network is vulnerable to interference and the channel quality fluctuates greatly. In order to ensure the delay requirement of high real-time application and reduce packet header overhead, an adaptive packet length strategy (K-PK strategy) based on symbol number is proposed through the joint optimization of network layer and physical layer. The strategy is to aggregate k consecutive arriving symbols in a Poisson distribution arriving symbol stream and add fixed length packet headers to form packets. Simulation and analysis show that there is an optimal k value to obtain the lowest delay under the condition of certain channel bit error rate (BER). Compared with the fixed-slot packet formation strategy (T-PK), the balanced packet head overhead .K-PK strategy is more suitable for poor channel quality and high volatility, and meets the needs of reality.
【作者单位】：暨南大学信息科学技术学院;
【基金】：国家自然科学基金资助项目(61373125) 广东省科技计划资助项目(2013B010401016)
【分类号】：TN92

【参考文献】