The greatest advantage of laser communications (lasercom) over radio frequency (RF) for space-to-ground communications-narrow beamwidth-also poses the greatest technical challenge. The considerably smaller transmit beamwidth in optical communications, typically on the order of tens of microradians, imposes stringent demands on the pointing system, especially in the presence of spacecraft vibration. Inaccurate beam pointing can result in significant signal fades at the receiving site, producing a large number of burst errors. One of the important steps in realizing space-ground laser communications, therefore, is to track the target (or receiver) with residual pointing error that is small compared to the transmit beamwidth.
FIGURE 1. The gimbal-mounted telescope optics assembly consists of a 10-cm telescope, a CCD array detector (top-mounted box), and associated optomechanical components.
The Optical Communications Demonstrator (OCD) program at the NASA Jet Propulsion Laboratory (JPL; Pasadena, CA) was created to address this issue in a laboratory environment and will validate several key free-space lasercom technologies such as beacon acquisition, precision beam pointing, and high-bandwidth tracking. The OCD terminal uses a patented, low-complexity architecture with only one fast steering mirror (FSM) and one focal-plane array (FPA) for all acquisition, tracking, and point-ahead functions.1 The large field-of-view (FOV) charge-coupled-device (CCD) array detector is used in a "windowed" mode for achieving high frame rates that are required in the tracking mode. This is different from the typical approach of using quadrant photodetectors for spatial tracking. Because the OCD was designed mainly for dumping data back to Earth from deep space at modest data rates (less than 10 Mbit/s), and from Earth-orbit at very high rates (100 Mbit/s to several gigabits per second), there is no uplink communications detector on the terminal. This reduction in the number of components is expected to increase system reliability. The OCD targets the higher Earth-orbit data rates because the first flight demonstrations are expected to be from Earth orbit.
The OCD consists of several components: the telescope optics assembly (TOA), a gimbal for coarse pointing, a fiber-coupled laser transmitter, and electronics to control and operate the terminal. The TOA is mounted on the gimbal and consists of a 10-cm telescope, a CCD array detector, a two-axis fine steering mirror, and other optomechanical components (see Fig. 1). The transmit laser is located away from the telescope and is coupled to the TOA using a single-mode optical fiber. Fiber-coupling of the laser transmitter eases thermal management issues and makes it easy to put in a different laser. Though some of the electronics for the CCD are mounted on the TOA itself, most of the electronics (including the processor) are placed away from the TOA and gimbal.2,3
Optomechanical design
The OCD optical assembly has three paths: a transmit channel to relay light from the laser to the output aperture through the FSM; a receive channel to relay the incoming beacon to the CCD; and a boresight channel to divert part of the transmit signal to the CCD (see Fig. 2).4,5 Because the acquisition path is not steered by the FSM, the beacon spot on the CCD moves around due to spacecraft jitter. The function of the electronics is to point the beam by controlling the FSM so that the transmit laser spot (from the boresight channel) on the CCD maintains a given vector distance from the beacon spot.
FIGURE 2. The inner workings of the telescope optics assembly include optical elements for directing the light beam along the transmit, receive, and boresight paths.
In the transmit channel, the 840-nm laser light from the fiber-coupled laser is expanded and collimated before being deflected by the FSM and transmitted out of the telescope. Nearly all the power in this beam passes through a dichroic beamsplitter and exits the OCD telescope aperture. The resulting far-field pattern is shown in Fig. 3. The observed size of the Airy disk agrees well with the predicted value of 21 µrad. The observed far-field profile, however, exhibits asymmetry as well as unusually large secondary rings. Better alignment between the primary and secondary is expected to improve the performance.
In the receive channel, the incoming beacon at 780 nm is collected by the 75-cm2 area of the OCD telescope. Nearly all of the collected energy is reflected off the dichroic beamsplitter and passes through a narrowband optical filter that rejects the out-of-band background noise before reaching the CCD camera. Almost 65% of the collected energy reaches the array detector.
The CCD-camera specifications for the noise-equivalent exposure and saturation-equivalent exposure are 15 pJ/cm2 and 45 nJ/cm2, respectively. For a 2-kHz frame rate (that is, 500-µs exposure time) and 16 ? 16-µm pixel size, we thus require about 1 nW of power at the CCD for near-saturation illumination. Note that the point-spread function of the optical system results in an Airy disk covering about 3 x 3 pixels. Typically, when the peak pixel value is near saturation, the neighboring nonzero pixels account for about four times the peak value. Given the size of the OCD aperture and the CCD sensitivity, incident intensities in the range of 20 pW/cm2 are needed to achieve reasonable signal-to-noise ratios for tracking purposes.
FIGURE 3. As seen by the divergence camera of the Lasercomm Test and Evaluation Station, the far-field pattern of the OCD is approximately 21 µrad. The station was built at the JPL as a general- purpose diagnostic tool to test any optical communication terminal with aperture up to 8 in. in diameter.
The third channel, or boresight channel, provides a real-time reference for the point-ahead angle by reflecting a fraction of the transmit signal off the dichroic beamsplitter and then off a "retromirror" that reflects the beam onto the CCD array. Attenuation of the 840-nm transmitted beam by the retromirror (which is uncoated on one side) and by the 780-nm narrowband filter result in a millionfold (60 dB) reduction in power by the time the beam reaches the CCD. This is acceptable if the average transmit power is around 1 mW. Further attenuation of around 20 dB is necessary to use a laser with 50-mW average power. Researchers at the JPL plan to achieve the higher attenuation in the boresight path by putting an antireflection coating on the retromirror.
Electronics hardware
The acquisition and tracking electronics for the OCD consist primarily of a Texas Instruments TMS420C40 (C40) based digital-signal-processing card and a tracking-processor-assembly (TPA) card set. The C40 digital-signal-processor card is a commercially available component and provides local data processing of camera images and control of the FSM and gimbal. The TPA is a custom-built card set used to interface the C40 card to the CCD, FSM, gimbal, and laser. The C40 card communicates with the TPA cards through a high-speed global bus (see Fig. 4).
FIGURE 4. In Optical Communications Demonstrator, the electronics communicate with the telescope optics assembly, gimbal, and host spacecraft.
A 486/586-processor-based single-board computer is the main instrument controller. It accepts commands from and sends status information to the bus controller through the MIL-STD-1553B card. The C44 (surface-mount version of the C40) digital-signal-processor card is slave to the x86 computer and performs all the acquisition- and tracking-related processing.
Acquisition and tracking
The acquisition and tracking function is achieved with a 128 x 128-pixel array CCD camera from Dalsa Inc. (Waterloo, Ont., Canada). The pixels are square and 16 ?m on a side. The CCD is mapped to a total FOV of about 1.2 mrad, or about 10 ?rad/pixel. Operating at 16 MHz, the CCD has 8 bits of resolution per pixel. The CD image data are read into the digital-signal processor through a Unix named pipe on the TPA card. The camera`s electronics were modified by 20/20 Designs Inc. (Los Altos, CA) to allow fast readout of small subframes by dumping unwanted rows and columns of data. When focused, the beacon and laser spot occupy fewer than 3 ? 3 pixels, sufficient to provide a centroid accuracy of about 0.1 pixel.
During acquisition of the beacon, the control software scans the 128 x 128 CCD camera image by using a 5 x 15 windowing scheme to locate the beacon position in the CCD field of view. The FSM is held fixed during this acquisition phase. The scan starts at the bottom right of the CCD field of view. If a beacon is not found, the next 5 ? 15 window directly above is scanned. This process is continued until the top of the CCD FOV is reached and is denoted as a stripe. If the beacon is still not found, the processing continues to the left to begin a new stripe. A valid beacon is defined as one or more pixels having a value greater than a predefined threshold. The transmit laser is turned off during the acquisition process so as not to confuse the acquisition algorithm. Once a valid beacon is found, the software enters the tracking mode and the transmit laser is turned on.
During tracking, a centroiding routine provides subpixel information on the location of the beacon and transmit-laser positions in the CCD field of view. This information is used to control the FSM and gimbal as well as adjust the active windows to be scanned on the next pass. The CCD dumps all pixel values during tracking except those from two 8 ? 8 windows around the laser and beacon spot to achieve a frame rate of 2 kHz. The function of the tracking routine is to maintain the transmit beam at a fixed position relative to the beacon using the FSM and to keep the transmit beam in the center of the CCD using the gimbal.
We have performed extensive analysis of the fine-pointing control loop. The 3-dB closed-loop bandwidth of the mirror response was measured to be more than 110 Hz in both axes. The 0-dB bandwidth of jitter suppression was measured to be near 60 Hz. Performance of the fine tracking loop was limited primarily by large system delays (the time between the exposure and application of control signal to the mirror) and dynamic performance of the mirror. The system delay was measured to be more than 1.25 ms in both axes and is in large part inherent to the CCD. The OCD FSM has a first resonance around 18 Hz, and its response falls off quickly beyond 30 Hz.
Significant performance increase is expected with active pixel sensors (APSs) and newer FSM mechanisms. The APS offers random access to pixels and therefore time need not be wasted in dumping rows and columns of pixels as in CCDs. We are currently evaluating new FSM mechanisms that have bandwidths of several hundred hertz, and we expect to obtain about 200 Hz, 0-dB jitter suppression bandwidth.
System-level demonstrations
Though the OCD was developed primarily for validating optical communication functions in a laboratory setting, it is now being used in the field for system-level demonstrations in the presence of an atmospheric channel. Specifically, we are using the OCD in ground-to-ground demonstrations between Strawberry Peak (in the San Bernardino Mountains) and the JPL Table Mountain Facility (TMF) near Wrightwood, CA. The distance between the transmitter (OCD at Strawberry Peak) and the receiver (TMF telescope) is 46.8 km, and the average height above sea level of the laser beam path is 2 km.
The 0.6-m aperture telescope receiver at TMF also sends a beacon laser beam to the OCD. Using the beacon as reference, the OCD transmits a modulated laser signal back to TMF simulating a bidirectional laser link conceived for space-to-ground optical communi cations.
Using this mountaintop-to-mountaintop configuration, we are evaluating OCD performance in the presence of atmospheric-turbulence-induced scintillation, beam spreading, and beam wander. Following completion of ground-ground demonstrations, environmental tests of the OCD in thermal, vacuum, and vibration chamber will be conducted.
OCD design, construction, and characterization, along with ground-ground demonstrations, have provided valuable experience in building and operating optical communication terminals. Work is already underway to construct an improved, flight-qualified, optical communication terminal for the International Space Station.8
ACKNOWLEDGMENTS
The research described in this report was carried out by the Jet Propulsion Laboratory, California Institute of Technology, under contract with NASA. The authors would like to acknowledge the work of S. Monacos, N. Page, A. Biswas, A. Portillo, C. Racho, B. Sanni, G. Ortiz, D. Erickson, K. Masters, J. De Pew, J. Brower, B. Kemp, and T.-Y. Yan.
REFERENCES
- J. R. Lesh, C.-C. Chen and H. Ansari, US Patent #5,517,016 (May 1996).
- C.-C. Chen and J. R. Lesh, SPIE Proc. Free-space Laser Communications Technologies VI 2123, p. 85 (1994).
- T.-Y. Yan, M. Jeganathan, and J. R. Lesh, SPIE Proc. 2990, 94 (1997).
- N. A. Page, SPIE Proc. Free-space Laser Communications Technologies VI 2123, p. 498 (1994).
- B. R. von Lossberg, SPIE Proc. Free-space Laser Communications Technologies VI 2123, p. 505 (1994).
- K. E. Wilson et al., SPIE Proc. 2990, 152 (1997).
- A. Biswas, K. E. Wilson, and N. Page, SPIE Proc. 3266, 22 (1998).
- J. Sandusky and J. Lesh, SPIE Proc. 3266, 128 (1998).