Threshold Vector Field Estimators in Non-Gaussian EMC Environments

The present paper extends the author's earlier work [1H4] on weak signal estimation in electromagnetic compatibility (EMC) environments to include simultaneous space-time signal processing of vector (electromagnetic) fields. Specifically, the emphasis here is on signal intensity and waveform estimation in telecommunications, where typically desired signals are narrowband, are received in the far-field, and have well-defined wavefronts, subject to directional transmission and reception. The principal interference is other similar ”intelligent" transmissions and can include both natural and man-made "unintelligent" interference, e.g., atmospheric noise, automobile ignition effects, etc. A principal aim of the present study is to establish limits on the reconstruction (“extraction") of desired signal waveforms in crowded spectral regions, where the Interference is usually highly "structured" and nongaussian, and to determine the extent to which taking account of the vector nature of both signal and interference can modify and potentially improve the process of signal recovery. A high level of generality is postulated, including general signals and geometries. The resulting processing algorithms are new, reducing to familiar (and earlier) forms when the interference is gaussian. Optimum maximum likelihood estimators are constructed in the threshold régimes, based on the associated optimum threshold detection algorithms. Again, three principal modes of reception are considered: (i) coherent; (ii) incoherent; and (ill) composite, which require different algorithmic structures. An important result is the potentially large (O(10-30 dB) or more) degradation in estimation accuracy encountered when conventional estimators based on gaussian noise are employed, paralleling a similar effect in similar detection situations [2], These vector estimation algorithms are expectedly more complex than the scalar field cases, but remain explicitly manageable. Examples are included, as well as a comparison of systems in typical, nongaussian interference environments.