This paper presents new pseudolinear estimation algorithms for angle-of-arrival (AOA) target motion analysis when the sensor locations are not precisely known. Sensor location errors can cause significant performance degradation for the existing pseudolinear estimator (PLE) and its variants. In particular, we analytically show that, in the presence of sensor location errors, the well-known weighted instrumental variable estimator (WIVE) is no longer asymptotically unbiased due to the non-vanishing correlation between the instrumental variable matrix and the pseudolinear noise vector. To ameliorate this bias problem, a novel bias compensation method is developed for the WIVE, which is proven to be approximately efficient for a large number of measurements under the small noise assumption. In addition, a selective-angle-measurement (SAM) strategy is integrated into the bias-compensated WIVE (BCWIVE) to strengthen the correlation between the instrumental variable matrix and the measurement matrix, alleviating the performance degradation in large AOA and sensor location noises, and unfavourable geometries. The performance advantages of the proposed BCWIVE and SAM-BCWIVE over the PLE, its WIVE variants and the maximum likelihood estimator are demonstrated by way of extensive simulation examples.

中文翻译：

---

存在传感器位置误差时的 AOA 伪线性目标运动分析


本文提出了新的伪线性估计算法，用于在传感器位置未知时进行到达角 (AOA) 目标运动分析。传感器位置错误可能会导致现有伪线性估计器 (PLE) 及其变体的性能显着下降。特别是，我们分析表明，在存在传感器位置误差的情况下，由于仪器变量矩阵和伪线性噪声向量之间的非零相关性，众所周知的加权仪器变量估计器（WIVE）不再是渐近无偏的。为了改善这个偏差问题，我们为 WIVE 开发了一种新颖的偏差补偿方法，该方法被证明对于小噪声假设下的大量测量来说是近似有效的。此外，偏置补偿WIVE（BCWIVE）中集成了选择性角度测量（SAM）策略，以加强仪器变量矩阵和测量矩阵之间的相关性，减轻大迎角和传感器位置噪声下的性能下降，和不利的几何形状。通过广泛的仿真示例证明了所提出的 BCWIVE 和 SAM-BCWIVE 相对于 PLE、其 WIVE 变体和最大似然估计器的性能优势。