论文部分内容阅读
自适应抗干扰阵列系统通常采用数值稳定性良好的QR分解最小二乘法(QRDLS)进行自适应处理.在需要最优权值的应用中,采用“回代法”或“权冲洗法”从阵列中得到权值.这两种方法对阵列单元施加特殊控制,不能实时实现自适应处理,也难于进行VLSI实现.文中提出的最优权值并行提取算法可以克服上述缺点,它在数据矩阵被三角化的同时从阵列中得到最优权值.另外,由于算法利用Givens旋转完成,全无除法.采用本方法的自适应处理特别适宜通过CORDIC技术进行VLSI实时实现.
Adaptive anti-jamming array systems are usually adaptively processed using the QR decomposition least square method (QRDLS) with good numerical stability. In applications that require the optimal weight, weights are derived from the array using the “back-pass” or “flush-to-weight” method. These two methods impose special controls on the array elements, failing to implement adaptive processing in real time, and difficult to implement with VLSI. The proposed parallel algorithm of optimal weight can overcome the above shortcomings. It obtains the optimal weight from the array while the data matrix is triangulated. In addition, since the algorithm uses Givens rotation to complete, there is no division. The adaptive processing using this method is particularly suitable for VLSI real-time implementation by CORDIC technology.