期刊论文

中文

武汉轻工大学学报

2095-7386

2019

38

2

44-49

10.3969/j.issn.2095-7386.2019.02.009

中国食品公益专项研究基金（201513004-3）；

本校为第一机构

针对现有的脉冲神经膜系统只能进行无符号二进制整数的乘法运算,而实际应用中更多地需要处理有符号整数的乘法运算,故用脉冲神经膜系统实现有符号整数的乘法运算更能满足实际需要。首先,设计了并行乘法器中并行数据的数值位与符号位的分离。其次,设计了用来计算任意两个有符号整数乘法的并行乘法器的脉冲神经膜系统。接着,设计了并行乘法器中两个并行数据的数值位与符号位的分离。最后,设计了能够执行任意两个有符号整数乘法运算的脉冲神经膜系统。系统仿真表明,所设计的脉冲神经膜系统都能准确地执行有符号整数的乘法运算。当前问题的解决,将有助于脉冲神经膜系统的生物型CPU的设计。

For the existing spiking neural membrane system, it can only multiply unsigned binary integers, but in practice, it is more necessary to multiply signed integers. Therefore, it is more suitable to multiply signed integers by spiking neural membrane system. Firstly, the separation of numeric bit and symbol bit of parallel data in parallel multiplier is designed. Secondly, a spiking neural membrane system is designed to compute the parallel multiplier of any two signed integer multiplications. Then, the separation of numeric bit and symbol bit of two parallel data in parallel multiplier is desig...