智慧型天線接收機之DOA與Beamforming演算法即時實現

Abstract

本論文是針對上鏈(Uplink)智慧型天線接收機的實現，需要使用到達角度估測(DOA)理論與波束形成(Beamforming)演算法，因為所需計算量龐大，因此構建多數位訊號處理器的實驗平台，為了有效的分散運算量，研究在多顆數位訊號處理器之間、個人電腦與數位訊號處理器之間，溝通與資料快速傳遞之方法。論文中，在數位訊號處理器上應用矩陣運算即時實現MUSIC、ESPRIT及MVDR演算法，對採樣到的實際訊號作演算法的驗證，並對演算法作個別的效能評估且適當的分配演算法的運算量到多顆數位訊號處理器上，即時實現智慧型天線接收機的功能。

This thesis focuses on real-time implementation of uplink smart antenna receiver based on the DOA（Direction-of-Arrival）theory and the Beamforming theory. The number of operations involved in computing DOA and Beamforming algorithms is large, so a computing platform is constructed using multiple DSPs (Digital Signal Processor) architecture, in order to spread efficiently the computation load. Communication and fast data transfer methods between DSP and DSP or DSP and PC are investigated. In this thesis, MUSIC, ESPRIT, MVDR algorithms are implemented using matrix computation in real-time and validated in applications with real data. Computing load of individual algorithm is evaluated. Suitably spreading computing load on multiple digital signal processors can implement uplink real–time smart antenna receiver.