Based On Qr Decomposition The Adbf Algorithm And Dsp Implementation

Adaptive beamforming,widely applied in Radar, Sonar and Communication, isone of the important part of signal processing. For recent forty years, adaptivebeamforming theory has been established and many algorithms have been developed.Nowadays, the algorithm of data domain which is numerically stable and has inherenthigh efficient parallelism, is the research hotspot. Two modified algorithms on QRdecomposition type SMI are studied, the schemes of above mentioned algorithms onthe ADSP-TS101S EZ-KIT Lite emulational board are also given and implemented inthis dissertation,. Corresponded to them, the main work of this dissertation is flollows:1. The conventional QRD-SMI algorithm is discussed, a Systolic structure easilyimplemented on DSP is presented, and a scheme tested on the hardware emulationalplatform is presented as well. Experiments show that the algorithm and scheme areeffective.2. A modified QR decomposition SMI algorithm (MQRD-SMI) is studied, andcomputer simulation has been accomplished. A Systolic structure about the newalgorithm easily implemented on DSP is presented, and a parallel scheme tested onthe hardware emulational platform is presented. The experiments show that thealgorithm and scheme are real-time and effective.3. Another modified QR decomposition SMI algorithm (IQRD-SMI) is studiedbased on MQRD-SMI, and computer simulation has been accomplished. A Systolicstructure about the new algorithm easily implemented on DSP is presented, and aparallel scheme tested on the hardware emulational platform is presented as well. Theexperiments show that the algorithm and scheme are real-time and effective.4. We present the method of diagonal loading about three QR decompositionalgorithm on multiple DSPs.