High Speed Implementation Techniques For Adaptive Sidelobe Cancellation Algorithms Of Phased Array Radar

The Phased Array Radar changes the phase amplitude function to achieve the purpose of completing regional beam scanning and choosing received signals.The individually controlling mechanism for each array antenna gives the Phased Array Radar excellent electronic beam steering capability,which facilitates the formation of targeted beams pointing to different airspace orientations.If the environment contains high-intensity interference signals with unknown locations,though the weak target signal will be received by the main beam,the residual intensity of the interference from the beam side lobes is still sufficient to cover the characteristics of the desired signal,influencing the radar extraction of parameters from the target.The adaptive sidelobe cancellation algorithm is a mature theory and reliable engineering method used to resist such interference.By using an auxiliary cancellation system,the interference echo from the side lobes in the main receiving channel can be reduced.With the deterioration of the interference environment and the diversification of radar functions in the future,the number of antenna arrays and the tasks of the processing system will inevitably rise.The great amount of data will cause the feedback speed of the results becoming slow,eventually leading to different levels of display delay,reducing the radar's real-time competitiveness.In order to match the development trend of phased array,it is important to explore a higher speed adaptive sidelobe cancellation algorithm for large radar information processing systems.In this thesis,the theoretical knowledge of adaptive sidelobe cancellation algorithm is introduced under the set signal type from the aspects of composition,optimal solution.The simulation environment uses a binary area array model which is close to the real-situation direction finding.On this basis,the effects of different parameters of the antenna and signal on the result and processing time are discussed.After fully understanding the underlying structural characteristics of the GPU and the rules of the CUDA compilation platform,the sidelobe cancellation algorithm is disassembled by the idea of parallel heterogeneous computing,and an high speed adaptive sidelobe cancellation based on the GPU multi-thread programming framework is realized.It focuses on accelerating the matrix multiplication process and complex matrix inversion process in the calculation.Through multiple partitioning,data and task double-sided parallelism and other means to optimize performance.The result on the CUDA platform is compared with the result of the CPU serial operation.When the destructive result is stably floating in a certain error range,the parallel calculation is clearly better accelerated,and the running time is significantly reduced.It can be obtained through experimental tests that the high-speed implementation technology is insensitive to the large increase in data compared to the traditional serial algorithm.On the contrary,as the data size increases,it can show better acceleration performance.