A Tall Ship and a Star to Steer Her By

The vision: Sailing ships once navigated by the fixed stars and a nautical chronometer. Untethered from the terrestrial reference frame, future spacecraft will require more than a star chart to accomplish the same feat; instead, they may exploit the peculiar time and frequency structure of light from those distant stars to autonomously or semiautonomously determine their position and velocity in the celestial reference frame. 1.1 Broad Objectives: Two visionary innovations were proposed to allow spacecraft to assume the burden of navigation. The first extends the International Celestial Reference Frame (ICRF) to include not only orientation references (quasars), but also time (pulsars) and frequency (maser) references that allow position and velocity to be determined relative to the terrestrial reference frame (TRF) or planetary reference frames (PRFs). The second provides spacecraft with a direct means to determine their own position relative to the ICRF. 1.2 Proposed approach: The inspiration for the Tall Ship proposal was the progress being made towards using time-of arrival (TOA) and Doppler shift X-ray signatures of millisecond pulsars in support of autonomous spacecraft navigation. This technique, sometimes dubbed xNAV, will be tested at a proof-of concept level as part of the upcoming NICER/SExTANT experiment on the International Space Station (Arzoumanian et al. 2014). While xNAV is promising, the X-ray signal from pulsars is intrinsically weak, the detection instrumentation is cumbersome and single-purpose, and the overall resource burden on small spacecraft is likely to be large even after projecting the technology forward in time. In that respect, analogous techniques implemented using radio methods in the microwave bands (what might be called RNAV) may be preferable. The signal-to-noise ratio of the sources is higher in the microwave, the technology better lends itself to miniaturization, and large radio antennas and receivers are more synergistic with other spacecraft subsystems in the sense that they can contribute towards communication, and can share collecting area with a solar power system. Electronic beamforming technology will potentially reduce most of the pointing burden on the spacecraft as it matures, and (as described here), bright and ubiquitous astronomical masers, which provide frequency references for Doppler measurements, may complement or even replace the rare and difficult-to-detect pulsars as catalogue sources. The Tall Ship team therefore proposed to observationally determine, first, whether well-known astronomical masers are sufficiently stable on short time scales to constitute an ICRF sub-catalog that can be used for spacecraft navigation. Second, we proposed to assess the scope of spacecraft resources that would be required to exploit such a catalog, including antenna area, deployment mechanisms, power, and computational capability. Finally, we proposed to explore deployment of even larger antennas from small spacecraft that would be able to utilize both pulsars and masers for navigation. 1.3 Findings By way of overall summary, the prospect of autonomous navigation with astronomical masers proved less inviting in light of the Phase I research findings, but the prospect of using radio pulsars was significantly improved. The difference in the degree of difficulty is nowhere near as great as initially expected, and the navigational value of radio pulsar signals is found to be significantly greater. Moreover, a straightforward and inexpensive path to flight validation was identified. 1.3.1 RNAV with masers Despite the need for an extra integral to extract position information from Doppler shifts, we originally focused on astronomical masers because of the relative simplicity of detection, the ubiquity of bright sources, and the opportunity to use relatively small antennas to detect bright sources in the 3-30 mm (10-100 GHz) range. Initial estimates of signal strength suggested that 1 m2 would be adequate antenna area, a size nominally compatible with a CubeSat. Observations: Initial observations using the Haystack Observatory 37 meter telescope suggested that from an overall system-to-noise (SNR) perspective, water masers at 22 GHz were a better target than SiO masers at 43 GHz for a similar size antenna. Subsequent observations focused on studying the stability or predictability of spectral features at various times scales. In general, it was founded that on relatively short time scales (hours to days), frequency shifts of peaks did not exceed 20-30 m/s. While encouraging, this finding still lacks an order of magnitude or more of the precision needed to achieve the desired system performance for RNAV. However, strong evidence suggested that most of the residual changes were the result of instrumental error, and work continues both on improving the instrumentation and identifying features and patterns in the spectra that will further reduce the uncertainty. Antenna design: Purely from the perspective of performance, phased array antennas are appealing for RNAV because steering can be done totally electronically. Initial assessments, however, suggested that the corresponding power penalty of simultaneously processing signals from many antenna elements was unfavorable, and reflectarray technology is a more practical approach from a system perspective. Reflectarrays use many passive elements focused on a single feed to achieve the advantage of deployment on a flat surface without the power penalty. Reflectarrays offer a large upside potential as electromechanical methods of beamforming mature, and eventually they may offer steerability comparable to a phased array without the power penalty. In the short term, they offer performance comparable to a parabolic dish, but can be integrated into flat solar arrays (either on the back or, in a transparent configuration, on the front) with a negligible mass and volume penalty on the spacecraft. Scaling of the actual Haystack maser observations to a small space-based antenna suggests that the originally postulated 1-m2 antenna was overly optimistic, and that an aperture of at least 4 m2 is required to achieve the SNR needed to maintain positional accuracy of 10-20 km after a few months, even with ideal stable sources. While a factor of 2 in linear dimension seems modest, anything larger than a meter on a side is challenging for CubeSat implementation. Compounding the difficulty is the surface figure that must be maintained for quality measurements, typically considered to be 1/15th of a wavelength, or ~1mm at 22 GHz. This degree of surface control implies a level of stiffness, and hence mass, that isn't found in typical CubeSat solar panels. In theory, flaws in the figure can be calibrated in situ and compensated by tuning the antenna elements, but this capability has not yet been realized in lightweight space-based arrays. 1.3.2 RNAV with pulsars While astronomical masers only allow determination of the instantaneous velocity vector of a spacecraft, detection of time-of-arrival and Doppler shift of radio pulsars, particularly msec-scale pulsars, provides a direct determination of instantaneous position. At the time of submission of the Phase I proposal, the literature suggested that a capable pulsar detection system required of order 100 m2 antenna aperture, and imposed mass and power requirements that exceeded the capability of small satellites Becker et al (2013). More recent studies out of the same group (Jessner 2016), now suggest that the power and computational issues are modest, and an antenna size of 30-40 m2 (i.e. about 3x the diameter of the maser antenna) will be sufficient for detection. Moreover, at the recommended frequency of ~500 MHz, the antenna may consist of a lightweight mesh, and surface errors of 5-10 cm can be tolerated. As a result, we can consider relatively flimsy structures for pulsar detection that weigh no more than antennas for maser detection, and are substantially more tolerant to distortion. From a system perspective, this may prove simpler and less resource intensive to implement, with vastly superior navigational results. 1.4 Next steps The overall conclusion of this work is that radio navigation (RNAV) using pulsar signals in the UHF band is extremely promising in the near-term; the original maser focus of this work offers a potentially useful augmentation, though it faces near-term challenges in accommodating the precision antenna. The RNAV concept is most readily validated in the near term by leveraging existing assets, using data from Radio Astron, a 10-meter space-based telescope currently in a high orbit around Earth (Kardashev et al. 2012). Designed primarily for Very Long Baseline Interferometry (VLBI), Radio Astron is equipped with high precision timing and its absolute position is known to great accuracy. Should a Phase II effort be awarded, we proposed to independently demonstrate determination of Radio Astron position using existing or new pulsar timing data. Using new and existing technology, we also describe in this report a dedicated RNAV demonstration featuring a large-area deployable antenna developed by our partner Stellar Exploration, Inc. on a small satellite that would be deployed on a cis-lunar or deep space trajectory