Coding and Modulation
Compression, Modulation, and Forward Error Correction
The telescope will begin taking 10 megapixel, 4-bit gray scale images one tenth of a light year away from Epsilon Eridani in order to ensure that some pictures are sent back with minimum corruption from the magnetic field caused by the star system.
The flow chart below illustrates the method of sending images back to Earth.
Each image is first compressed using a fractal compression scheme to compress the image to 1/170 of its original size. Besides the high compression ratio, this compression scheme is useful because it is resolution independent. This means that when the image is decompressed back on Earth, the resolution can be increased as needed to obtain a suitable image. Once the image is compressed, it is streamed at a bit rate of 1 bit/second. Raised cosine pulses with a roll-off factor of 0.5 are used for the link, which results in a bandwidth of 0.75Hz. A rate 1/2 turbo code is then used which increases the bit rate to 2 bit/s and the bandwidth to 1.5Hz. The modulation method used is QPSK, with the signal constellation shown below.
We desire errorless transmission at the receiving end (corresponding to a BER of10-6). Using the QPSK modulation alone, this requires an SNR of 10.52dB. However, the turbo coding reduces the minimum required SNR to about 1dB according to the graph below. To be certain of errorless transmission, our calculations allow for a minimum SNR of 2dB.
Source: Pratt, Bostian, Alnutt. Satellite Communications, Second Edition.
Copyright 2003.
It should also be noted that our assumed BER is quite conservative--the BER needed to obtain suitable images back on Earth is likely to be less than 10-6, especially given that our fractal compression scheme allows for resolution enhancement during the post-processing phase.
Our design parameters allow one image to be sent every 3 days.
Image Processing
Once the images are transmitted to Earth, they are decoded and decompressed. Finally, image processing algorithms are run on the digital images in order to remove noise and enhance the image quality. Namely, robust anisotropic diffusion will be implemented in order to remove obvious noise (such as dust) while preserving and sharpening the edges of the main features of the images, such as planets and rings that may be around the planets.
|
|
