采用嵌入马尔可夫链理论和概率母函数的方法,对在离散时间状态下的非对称门限服务轮询系统进行了分析,根据终端站信息分组在循环期内的平均等待时间有较好的近似性,分析了信息分组的平均等待时间.计算机模拟的结果表明了仿真与理论的一致性.
In view of Gauss noise characteristic, this paper proposed a denoising method carries on the gradation compensation with the variable step based on Pulse - Coupled Neural Networks (PCNN). This method worked well, and could manifest the PCNN' s nature capture characteristic better compared with other filtering methods. The experiment results showed that it had manifested a better performance in the signal - to - noise ratio improvement factor.