介绍了一种利用现场可编程门阵列(FPGA)对生物神经系统模型进行高速并行硬件仿真的新方法。提出了流水线算子和流水线模型的概念,为单核心多模型的实现方案提供了严格的数学推导。完成了12级流水线FPGA流程的设计,再现了Morris-Lecar神经元模型的非线性动力学特性。对设计方案的效率、计算误差以及硬件资源的消耗情况进行了详细的统计分析。
This paper presented a new method of high-speed parallel hardware simulation for biological neural system model which used field programmable gate array(FPGA).It defined the concept of pipelining operator and pipelining model,which could supply a mathematical deduction for realizing the single-core multi-model.Accomplished a 12-stage pipelining design flow,which could perform nonlinear dynamics of Morris-Lecar model.And did statistical analysis for the efficiency,error and resource consumption.