A 1nrad Delta-DOR system

Earth-based radiometric tracking is the primary source of navigation data during interplanetary cruise. Three types of measurements are currently employed: Doppler, ranging and ∆DOR (delta-Differential One-way Ranging). ∆DOR measurements are based on Very Long Baseline Interferometry (VLBI). This radioastronomical technique allows the determination of the angular position of radio sources by measuring the time delay between received radio signals at two widely separated stations. The observed time delay is a function of the known baseline vector joining the two radio antennas and the direction to the radio source. Current ∆DOR measurements are based on the following concept: a spacecraft emits several tones or other signal components spanning at least a few MHz. The characteristics of the tones are selected on the basis of requirements for phase ambiguity resolution, measurement accuracy, efficient use of spacecraft signal power, efficient use of ground tracking resources, and frequency allocation for space research. The ∆DOR technique requires that a quasar and the spacecraft are tracked nearly simultaneously from both participating radio antennas. Typically, a ∆DOR pass consists of three or more scans of data recording, each of a few minutes duration. A scan consists of pointing the antennas to one radio source (e.g. the spacecraft) and recording the incoming signal. The antennas must then slew to the other radio source (a nearby quasar) for the next scan, and so on. The observing sequence is spacecraft-quasar-spacecraft or quasar-spacecraft-quasar. A minimum of three scans is required to eliminate clock-epoch and clock-rate offsets and then measure spacecraft angular position. Typically, a three-scan sequence is repeated several times. For a spacecraft, the one-way range is determined for a single station by extracting the phases of two or more received signal components. DOR observables are formed by subtracting the one-way range measurements generated at the two stations. The station differencing eliminates the effect of the spacecraft clock offset, but DOR measurements are biased by the desynchronization of ground station clocks (clock offset), propagation media effects and instrumental delays. For measuring the quasar delay, each station is configured to acquire data in frequency channels centred on the spacecraft tone frequencies. This configuration of the receiver ensures that the spacecraft-quasar differencing eliminates the effects of ground station clock offsets and, to a large extent, instrumental delays. By selecting a quasar that is angularly close to the spacecraft, and by observing the quasar at nearly the same time as the spacecraft, the effects of errors in the station locations, Earth orientation, and transmission media delays are strongly reduced. In 2005 ESA has successfully developed and later made operational use of the ∆DOR system. The technique has been since then used to successfully navigate all ESA interplanetary missions (i.e. Venus Express (VEX), Mars Express (MEX) and Rosetta). Several future missions (e.g. BepiColombo, ExoMars, Solar Orbiter) will rely on this technique as well to reach their final targets. Current ESA ∆DOR performance level has been evaluated by using Venus Express data collected in the time-frame 2007-2011 and from Rosetta data collected in 2011-2014 during the cometary approach. Such performance level can be estimated to be presently in the order of 10-15 nrad. A preliminary evaluation of the ∆DOR error budget and related apportionment into separate contributions was undertaken in the frame of the activity “ASTRA: interdisciplinary study on enhancement of end-to-end Accuracy for Spacecraft TRAcking techniques”. Results, although not fully conclusive, showed a good agreement of the error model with the measurements. The study concluded that the major error sources for present ESA ∆DOR system were: • Phase non-linearities in the recording bandwidth; • Thermal noise in the recorded S/C and quasar signal; • Uncalibrated media (both systematic and random components). Dispersive phase errors are caused by non-linearities of the phase response of the receiver across the spanned bandwidth. Due to the different nature of the signals being recorded (a narrowband monochromatic signal for the spacecraft and a broadband noise for the quasar), the presence of non-linear phase response (e.g. phase ripple) over the recording bandwidth can induce a systematic phase error, which in turn translates into a systematic delay error. A first estimation of the phase ripple introduced by the current ESA system was of about 0.5 degrees in each channel, making this error one of the major contributors to the overall ∆DOR error budget. Thermal noise contribution in the S/C DOR computation is driven by the available P/N0 (DOR tone power-to-noise-density ratio) and the spanned bandwidth (the widest frequency separation between downlink signal components). Current ESA deep space missions like MEX, VEX and Rosetta do not have the capability of transmitting dedicated DOR tones. Instead, telemetry harmonics up to order 20 are used, although their amplitude is so low that the total spanned bandwidth has to be limited to approximately 10 MHz. In general a wider bandwidth is desirable since it greatly improves the performances due to the fact that thermal noise contributions for both S/C and quasar DOR measurements are inversely proportional to the spanned bandwidth. Also, uncalibrated media, especially rapidly varying wet tropospheric delays, represents presently a major source of error in current ∆DOR systems. The current level of accuracy (i.e. 10-15nrad) would clearly be insufficient for missions with more demanding navigation requirements. Such missions are normally the ones foreseeing a direct landing on a planet (instead of an orbit insertion and a successive landing) or a very stringent navigation requirement for orbit insertion around a planet. In this case, angular accuracies of down to 1nrad may be required.