Geolocation of multiple noncooperative transmitters using array processing

在这项研究工作的第一阶段（最初的 Engage 资助）中，研究问题被重新表述为同一频段中存在的多个非合作发射机的同时地理定位。这一功能的实现既包括设计适当且稳健的算法，也包括开发相关的硬件实现，以便根据软件定义无线电 (SDR) 方法在硬件测试台上进行验证。多个发射机要求采集站（三角测量位置需要其中至少2-3个**）配备天线阵列。为了在天线阵列尺寸适中（4-8 个单元）的多径传播环境中以实用且鲁棒的方式实现此功能，在第一阶段提出了一个两步过程，并在仿真中进行了验证，包括 1-** 应用盲源分离 (BSS) 算法来提取每个发射机的空间特征，2-** 从每个**空间特征估计主到达方向 (DOA)（大概对应于视线）。给定来自每个采集站的 DOA 估计，对位置进行三角测量变得微不足道。**在所提出的第二阶段中，两步方法及其变体的开发将在**模拟中继续，并且将为第一步中的 BSS 算法导出高效的硬件架构。**此外，将探索另一种新颖且更简单的方法，该方法使用与**第一步相同的 BSS 算法，但完全不需要 DOA 估计，因此不需要阵列校准或特定的阵列**拓扑。在第二阶段的范围内，第二种方法将完全在硬件中实现并**在测试台上进行验证。

In phase 1 of this research endeavor (original Engage grant) the research problem was reformulated as the**simultaneous geolocation of multiple noncooperative transmitters existing in the same frequency band. This**functionality is pursued both in terms of devising appropriate and robust algorithms, and developing associated**hardware implementations to be validated on a hardware testbed according to a software-defined radio (SDR)**methodology. The plurality of transmitters requires the acquisition stations (at least 2-3 of them being required**for triangulating positions) to be equipped with antenna arrays. In order to implement this functionality in a**manner that is practical and robust in a multipath-rich propagation environment with modest antenna array**sizes (4-8 elements), a two-step process was proposed in phase 1 and validated in simulation consisting of 1-**applying a blind source separation (BSS) algorithm to extract the spatial signature of every transmitter and 2-**estimating the main direction of arrival (DOA) - presumably corresponding to the line of sight - from each**spatial signature. Given a DOA estimate from each acquisition station, triangulating a position becomes trivial.**In this proposed second phase, development of the two-step approach and its variants will continue in**simulation, and an efficient hardware architecture will be derived for the BSS algorithm in the first step.**Furthermore, another novel and simpler approach will be explored which utilizes the same BSS algorithm as a**first step, but does away with DOA estimation altogether, thus not requiring array calibration or a specific array**topology. Within the scope of phase 2, this second approach will be entirely implemented in hardware and**validated on the test bed.