1. Introduction

The recent Optical PAyload for Lasercomm Science (OPALS) experiment successfully demonstrated an optical downlink from the International Space Station (ISS) to the Optical Communications Telescope Laboratory (OCTL) telescope located at Table Mountain near Wrightwood, CA [1]. The 50 Mbps Reed-Solomon, encoded, on-off keyed (OOK) modulated 1550 nm optical downlink was collected by the 1-m OCTL telescope and the atmospherically blurred focused spot was received with a large area (200 µm diameter) free space coupled avalanche photodiode detector (APD). Future space to ground optical communication systems can benefit immensely from much higher date rates in the 10’s to 100-Gbps range that these links are capable of supporting. To realize this benefit, much smaller active area detectors, typically single mode fiber (SMF) coupled receivers with higher bandwidth, will be required at the ground station telescope focus. Small telescope apertures can efficiently focus light onto these detectors but that requires the receive aperture to be smaller than the atmospheric cell size [2] which in turn will compromise the collection area or receiver gain and fail to support an operational link capability.

A remedy is to use adaptive optics (AO) correction of the atmospheric turbulence perturbed phase of the downlink beam wave-front so that near-diffraction limited focus spot sizes are achievable. AO systems have been routinely demonstrated to improve image quality in static astronomical observations as well as in imaging LEO objects for space surveillance [3,4] but have not been demonstrated in an operational space-to-ground optical link in the open literature.

We report, for the first time to our knowledge, the use of AO correction of an optical communication signal from a LEO platform to a satellite tracking ground station. When tracking LEO satellites from the ground, the received optical communication signal is disrupted by a combination of random atmospheric refractive index fluctuations and the relatively fast slew rate of the LEO satellite that effectively increases the transverse wind speed or rate at which the atmospheric turbulence moves past the ground receiving aperture. We demonstrate that fast closed-loop AO systems can be effective in closing high-bandwidth, LEO-to-ground optical communications links, even in daytime conditions.

The remaining paper is organized as follows. A notional link budget is presented to illustrate the critical role AO will play in enabling ultra-high rate space-to-ground communications from a future LEO spacecraft. The experimental setup is then described with an overview of the OPALS ground station and integration of the AO system provided by The Boeing Company. The results will then be discussed from an example overflight pass on Mar 4, 2015 followed by a second downlink opportunity on May 19, 2015 that culminated in the validation of the optical link as in the original OPALS demonstration.

2. Notional power budget

A 5 cm diameter unobscured transceiver transmitting 500 mW of 1550 nm laser power over a range of 1200 km is presumed for a power budget resulting in an incident irradiance of approximately 58 μW/m² at the ground aperture. A 400 km altitude LEO spacecraft at 75° zenith angle relative to the ground station was used. The large zenith angle is chosen in order to increase the link contact time. Reasonable optical transmission and wave-front error losses are assumed at either end of the link. Table 1 shows the net received power as a function of the ground aperture diameter in the first two columns.

Table 1. Aperture dependent received power

View Table

The split loss required in order to allocate 220 nW (−36.6 dBm) of downlink power for AO correction is shown in the third column and was chosen to be consistent with the power required by the AO system described below. Ground apertures less than 10 cm do not collect sufficient power. Assuming 10 photons/bit for differential phase shift keying (DPSK) modulation reported in the literature [5], −42.9 dBm optical power will be required for downlinking 40 Gb/s. The link margin corresponding to this requirement is shown in the fourth column of Table 1B. A portion of margin will be consumed in coupling to single mode fiber, fading and implementation losses. Conservatively allocating 10 dB for such losses would still leave healthy margin for 60-100 cm receiving aperture diameters that can be used either to increase data rate or range.

The Fried parameter r0, as low as 4.5 cm for daytime at 1550 nm and 75° zenith angle occurs (based on data gathered at Table Mountain, CA), will result in D/r0 values of 7-22 for 30-100 cm ground aperture diameters [6]. This will result in (D/r0)² (10’s to 100’s) spatial modes spread over ~D/r0 times the diffraction limited spot diameter of ~0.98*(λ/r0) with λ = 1550 nm. Without AO for correcting the atmosphere induced phase errors, single mode fiber coupling of the required optical power is not viable. Since future LEO-to-ground services would need to operate night and day for the 90th percentile clear sky atmospheric conditions, the notional power budget underscores the critical role that AO can play in achieving ultra-high rate data services.

3. Experiment

3.1 Overview of OPALS experiment

Details of the OPALS experiment have been published elsewhere [7]. The flight system was built by JPL as a low cost technology demonstration and included a two axis azimuth/elevation gimbal, a telecommunications derived fiber based master oscillator (DFB laser with external modulation) power amplifier laser transmitter, and a silicon charge coupled device (CCD) sensor for tracking. The laser collimator and camera were configured as a bi-static optical head. The avionics, including the 2.5 W laser transmitter were installed in a convectively cooled sealed air container to allow the use of low cost commercial off-the-shelf (COTS) avionics in space. The laser output fiber cable was fed through the sealed container and routed up the gimbal to the optical head. The OPALS flight system was delivered via a SpaceX launch and installed on an Express Logistics Carrier (ELC) pallet on the ISS in April, 2014. The output beam has a conservatively large full angle divergence of approximately 1.0 mrad to ensure sufficient received power at the ground station. The OCTL ground station implemented a multi-beam beacon comprised of four 976 nm laser diodes with 2.5 W per beam. The downlink signal was received by the APD and stored following clock and data recovery. The stored files were processed with a software decoder in order to recover the transmitted information. File transmissions on the optical link included HD video files, text files and PN data sequences [8]. Several successful links were performed previously during day and night conditions over the typical 2 min pass. The pass duration was constrained by the geometry of the flight system placement on the ISS and laser safety considerations prohibiting illumination of any ISS structure [1].

3.2 AO receiver system

The OPALS ground system at the OCTL was modified with the addition of a compact AO system test bed provided by The Boeing Company, El Segundo, CA. This AO system, developed by Jeff Barchers and purchased from SAIC in 2007, incorporates a specially designed self-referenced interferometer (SRI) wave-front sensor (WFS) and a Boston Micro Machines 1000-actuator MEMs deformable mirror (DM) with a direct-drive DM controller. The WFS includes a 19 kHz frame-rate FLIR InGaAs array camera which enables high-bandwidth AO operation. A polarization controller is incorporated into the AO system which adjusts the downlink signal polarization prior to the SRI in order to maintain the desired power splitting in the SRI during the downlink pass. The advantage of the SRI WFS over other AO systems such as the Shack-Hartmann based wave-front sensor, is the ability to correct high levels of turbulence at fast frame rates since the atmospherically-distorted phase is directly measured and applied to the MEMS-DM without need for wave-front reconstruction [9]. The optical path to the APD detector was replaced with relay optics coupling the OPALS downlink to the AO system as shown in Fig. 1. The telescope pupil was imaged on the AO system entrance in a collimated beam. The output is an atmospherically-corrected beam that was split between a tracking camera that measured the corrected Strehl ratio and SMF-28 fiber with 10.4 µm mode field diameter for recording power and communications. For the communication link, a commercial low noise amplifier (LNA) was spliced onto the output fiber along with a two stage temperature tuned narrow band fiber Bragg grating filter followed by a SMF OC1 receiver that consisted of a PIN detector integrated with a transimpedance amplifier (TIA). The 3 GHz Bragg grating filter was inserted to improve detection SNR of the optical communications signal and was required to track the center frequency of the Doppler-shifted optical signal from the ISS. The filter was manually temperature tuned during the pass to optimize performance and match the flight laser wavelength that could potentially drift as there was not enough signal on the filter monitor port for autonomous control.

 

Fig. 1 Schematic of OPALS ground system showing addition of Boeing provided AO system. FSM – fast steering mirror, OAP – off-axis parabola, MM – multimode, BS – beam-splitter, DM – deformable mirror, SRI –self-referencing interferometer, LNA – low noise amplifier, FBG – fiber Bragg grating, PD - photodiode

Download Full Size | PDF

Initially the AO correction was demonstrated in static ground-to-ground links over a 1.6 km atmospheric path. A remotely placed laser transmitter allowed beam alignment and tuning of the AO system parameters. Only power in fiber (PIF) measurements with the SMF output for recording power were performed at this stage in order to verify that the minimum OPALS downlink power would be adequate to operate the AO system. Approximately 220 nW of incident power was coupled into the AO system, 90% of which was utilized by the AO system. At this power level, full lock of the AO system was achieved. Several experiments were then conducted during ISS overflights with the OPALS flight system in order to verify AO system performance on the space-to-ground link with the objective of quantifying the PIF. This is in turn would determine that sufficient power, on the order of a few nano-Watts, could be achieved to close a communication link. After this was successful, the communication receiver was added and tested by downlinking a video file.

4. Results and discussion

During operation, the AO system senses and corrects the wave-front distortions present in the laser downlink signal. The AO system was manually enabled by operator when sufficient downlink power was received on the wave-front sensor. The AO system maintained good correction as long as there was sufficient signal strength on the WFS camera. Data taken from a single pass verified that with the OPALS downlink signal, the AO system was locked over an entire pass duration of 134 seconds. The ISS was tracked from 20° elevation through a maximum elevation angle of 60° in this case on Mar 4, 2015 at a local time of 2:54 pm with clear skies. Typical results of the output far field spots on the tracking camera are shown in Fig. 2, without and with the AO loop closed. The AO system tilt tracking was switched off during open loop measurements although telescope tracking was still operating. The open loop far field spot image shows the effect of the higher order wave-front distortions on the down link signal that are corrected by the AO system. The PIF was sampled at 10 Hz for 1 second at 6 second intervals and is shown in Fig. 3 along with measured Strehl ratios and pass elevation angle. The plotted Strehl ratios are computed by the AO system using the residual error between the instantaneous residual wave-front measured by the SRI WFS and a flat-wave-front calibration phase value. As indicated in the figure, observed Strehl ratios during closed loop operation averaged ~0.6 over the pass. A Strehl ratio of 1 represents ideal beam quality and the value of 0.6 demonstrates good wave-front correction. Without AO correction the Strehl ratio was < 0.02. The AO open loop and AO closed loop PIF probability distribution for the pass is shown in Fig. 4. The improvement in fiber coupled power is evident in the figure, with a 16dB improvement in the mean and a narrow distribution of 8dB during closed loop AO operation. A full aperture power measurement system has been added for future downlinks that will be used to normalize the power at the AO system and to determine the fiber coupling efficiency both open and closed loop. It was not operational for the current data sets. One of the constraints in our experiment was the low mean PIF of 2-3 nW since most of the received downlink power is used to achieve lock of the AO system. It should be noted that the OPALS experiment was not designed to be operated with an AO system but was used to demonstrate the feasibility of using AO on an optical downlink from a fast slewing LEO satellite. In an operational application, the link would be designed such that the higher required powers can be coupled into the fiber receiver, by narrowing the beam divergence on the transmit beam for instance.

 

Fig. 2 Tracking camera output of AO system showing open and closed loop case.

Download Full Size | PDF

 

Fig. 3 AO system performance showing Strehl and PIF during a single OPALS pass along with elevation angle. The AO loop closed after 12 seconds.

Download Full Size | PDF

 

Fig. 4 Open and closed loop power in fiber probability distribution of the pass shown in Fig. 3.

Download Full Size | PDF

The final communication link test was performed by capturing the downlink signal through the LNA and narrow band filter on May 19, 2015 at 9:45 am local time under clear sky conditions. The received signal for the pass in which the communication link was tested was somewhat weaker than the pass shown in Figs. 3 and 4. The exact reason for the lower power incident into the AO system was not identified though different atmospheric conditions or poorer pointing performance by the flight system could easily account for the variation. As a result, the data link was more intermittent. The full HD video was reconstructed in post processing and the BER, smoothed with a 1 second moving average, was measured. Because a software Reed-Solomon decoder can count the number of bit errors it corrects, it can be used to monitor the Bit Error Rate (BER) of the underlying uncoded link. For the two-minute OPALS pass, the BER of the uncoded optical channel is shown in Fig. 5. Gaps are evidence of the intermittent nature of the link. Points are also given in the graph for the times when there was not enough signal on the WFS for the AO system to correct the wave-front. During the gaps, link performance was not good enough to establish code-word synchronization, or the code-words contained too many errors. During this track, about 27.5% of the recorded data was decodable. The intermittency of the communication channel shown by the gaps in the BER is not only due to the AO system losing lock due to low received signal power but also due to the inadequate temporal tuning of the fiber Bragg grating (FBG) filter required to match the narrow filter band-pass with the Doppler-shifted downlink wavelength and the delayed clock and data recovery each time the link has to be re-established. These times correspond to lack of detected BER even when there was sufficient signal on the WFS for closing the AO loop. Although these contributed to the limited link performance, a full HD video transmission was successfully accomplished during the pass which was the success criteria for the original OPALS experiment and validates the feasibility of using AO on a LEO-to-ground optical downlink.

 

Fig. 5 BER as a function of time through the second pass showing intermittent link closure. Upper points reflect times when there was not enough signal on the WFS for the AO system to correct the wave-front.

Download Full Size | PDF

5. Summary

AO correction was performed on the OPALS downlink signal directly via a SRI based AO testbed with the corrected signal input to a SMF coupled communication receiver consisting of a LNA and temperature tuned narrow bandwidth filter and PIN photodiode. The performance of the system was characterized by the AO loop providing consistently high Strehl and reduced PIF signal fluctuations during LEO satellite downlink tests. Reconstruction of an HD video signal as in the original OPALS demonstration was also possible and demonstrates the ability to perform a communication link through the AO system. This represents the first published demonstration of AO correction on a LEO-to-ground optical communication signal and paves the way for high rate, multi-Gbps optical downlinks with LEO satellites and seamless connection to terrestrial fiber networks.