自回归背景噪声生成方法及其VLSI实现

On Generation of Background Noise with Autoregressive Data Model Technique and its VLSI Implementation

讨论了一种用自回归（ＡＲ）数据模型拟合生成具有指定功率谱结构的背景噪声的方法及其ＶＬＳＩ实现，提出了一种基于ＶＬＳＩ支持的双自回归滤波器，完成了对这种滤波器的设计方法—逆设计方法—的研究和它的工程建造工作．一个实际应用系统运行的结果，证明了这种背景噪声模拟方法的正确性及技术实现的合理性．

China's defence needs motivate the authors to study the generation of background noise that satisfics stringent requirements. The spectral structure of simulated background noise should approach closely that of real background noise and the simulation should be real-time.There appears to be no ready made solution for the above problem in the open literature. For example, AR (autoregrcssive) filter is, in the authors' opinion, needed for solving the problem but its detailed design for the authors special purpose, is not available.In the design of the needed AR filter, the authors make use of ten equations: eqs(3)～(12), which arc all well-known. The authors call their inverse-design method 'the AR data model fitting technique', in which the essential feature is that frequency response of AR filter is made to approach as closcly as possible the spectral structure of background noise. Only an outline of the inverse-design method is given in the full paper. To make real-time operation possible, much careful attention is paid to making purchased TMC 2310 and TMS320C25 chips form core of AR filter so as to provide VLSI support.The above AR filter with its VLSI implementation has been functioning satisfactorily in several operational engineering systems.