Research Of The Adaptive Beamforming Technology Based On M9703

This paper mainly study the realization and verification of adaptive beamforming technology, and build up the hardware of the system by using System Vue software and high speed data acquisition and analysis system, which provide a new solution for solving the problem such as long design cycle and verification complexity existing in radar system design.Firstly, the difference between LMS algorithm and QR_RLS algorithm are given in this paper。Some conclusions are got through analyzing the structure and the complexity of two algorithms. For example, the LMS is easily affected by convergence factor, but it is easy to design in FPGA or other hardware due to its simple structure. And the QR_RLS has faster convergence rate than the LMS, with its parallel processing architecture and high data throughput. And then the QR_RLS is more suitable for the module design of digital array radar beamforming.Secondly, an adaptive beamforming system is built by using the modules in System Vue software, thus the algorithm authenticity is verified successfully. And then the MATLAB Interface is used to customize the beam forming module based on the LMS algorithm. The comparison between the two cases shows that: 1. System Vue software can be used for radar system simulation and verification, and it also can greatly reduce the development cycle; 2. Radar algorithm that has been formed can be ported to the software and be packaged into its own radar libraries.Besides, this paper researches the causes and the compensation of channel imbalance on the M9703. A channel calibration system is designed successfully in System Vue. The whole design shows, System Vue software can be used to design any complex radar signals. The signals also can be generated by M8190, and it can be collected and stored by M9703. This Semi-physical simulation system can speed up the design process of radar system, it also can generate real radar data and reduce the workload of the radar operator.Finally, the article uses two FPGAs in the M9703 to achieve a four-channel ADBF algorithm. We set up a semi-physical simulation system to generate a real signal, using the self-developed module. The results after processing through the PCIe backplane hardware back to software and get the beam pattern. Compared the signal and after beam forming, the simulation results are consistent with MATLAB. This design is creative make radar IF receiver and radar signal processor to achieve unity with the M9703 A full use of the capabilities of the instrument, but al so provides a new research direction for the design and simulation of radar system.