Aperture synthesis from free-flying collectors without accurate metrology

Radio frequency aperture synthesis from multiple free-flying collectors is traditionally dependent on highly accurate intra-constellation metrology and shared clocks. We demonstrate that relaxed precision of relative positioning and clocking can be acceptable through computational means in post processing. If successful, this technology will allow the synthesis of sharp beams to increase directivity of RF reception and improve rejection of interferers. The work assumes a small cluster of RF collectors that can collect on an emitter at the same time, with position and time accuracy provided by commonly available aerospace GPS units. No cross links or inter-vehicle spacing is assumed. In lieu of accurate timing, correlation searchs are performed at the ground station. A key part of the idea is an alignment stage following collection. This is composed of a Cross Ambiguity Function (CAF) search across all possible FDOA/TDOA shifts, and a phase alignment search. These time, frequency and phase corrections are applied to the data, bringing them into alignment with each other. Maintaining this alignment over a longer time span is the challenge. This paper covers our work to date. This includes on-orbit results, a simulation model of the coherence-through-computation mechanism, consisting of models of the geolocation process, Doppler correction over time, and transforming input signals to a common time and frequency reference. A key success is a partial on-orbit test of the concept. We emulate a full constellation by using a single satellite, but collect the same VHF transmission at a variety of positions in a pass, providing the effect of varying timing and Doppler of independent satellites. We conducted 2200 full-pass collections. Using these real-world transmissions, we demonstrated that it is possible to post-process collects with significantly different Doppler and timing and coherently sum the “aligned” collects. A notable achievement is in comparing a single receiver's data with four collects aligned and summed. The initial collect (single collector view) was too noisy to be used - No useable information content was able to be extracted. However, after four independent collections were aligned using our method and coherently summed the data had sufficient SNR to be useful; four copies produced a SNR improvement of approximately 3dB. Our ongoing work includes transmitting with two emitters to demonstrate nulling or selective reception in an environment with co-channel interferers, and experiments at higher frequencies. Currently coherence is only maintained for a short period of time for a given collect, so smoothly transitioning between correlation search periods may provide the ability to maintain coherence across an entire satellite pass.