用重要性抽样方法估计分组码的未检概率

Estimation of Undetected Error Probability of Block Codes Using Importance Sampling

分组码的未检概率在ARQ体制中占有重要的地位。未检概率的直接计算需要了解码的重量分布,而后者对大部分码是未知的,因此,用计算机仿真方法估计未检概率显得特别重要。用常见的Monte-Carlo(MC)方法估计未检概率时遇到较大的困难,就是当未检概率较小时,MC方法所需的仿真时间是不能接受的。本文用重要性抽样方法估计纯检错系统的未检概率,不仅比MC方法的效率高得多,而且获得了满意的仿真精度。

The undetected error probability of block cedes is an important parameter in automatic repeat request (ARQ) systems. Its direct calculation requires a knowledge of weight distributions which remain unknown for most of block codes Therefore it is essential to estimate undetected error probabilities with computer simulation. However, there is such difficulty that, when common Monte-carlo (MC) techniques are adapt ed to estimate small undetected probabilities, they take too running time to be accepted In this paper, the undefected error probabilities of pure error detecting systems are estimated using an importance sampling technique which not only has more efficient than the MC technique, but also obtains satisfying pricision of simulation.