On‐line data processing techniques for MST radars

We review the principles of pulse compression, frequency stepping, and coherent integration in some detail; coarse quantization and the calculation of spectral moments are treated very briefly. Pulse compression and/or frequency stepping permit full use of the average power capability of the radar transmitter, while maximizing the height resolution and minimizing problems of range ambiguity. Barker codes, complementary codes, cyclic codes, and quasi‐complementary codes have all been employed for pulse compression. Complementary codes are particularly useful in mesosphere‐stratosphere‐troposphere observations, since the correlation time of the medium is usually very long. Only binary codes have been used to date, but quaternary codes might also prove to be useful. Another topic worth exploring further is sample weighting, or windowing. Such “unmatched” decoding slightly reduces the signal‐to‐noise ratio, substantially reduces the range sidelobes associated with matched decoding, and significantly increases the computation required for decoding, since multiplications are needed. The sidelobe reduction might be worth both added costs in some cases. Coherent integration is a crude but effective and easily implemented form of digital frequency filtering which can reduce drastically the subsequent computations required. Coarse quantization, usually combined with special purpose hardware, is another way of reducing the computation load, with usually only a small loss in statistical accuracy. Computation is also reduced by calculating spectral moments directly from only a few lagged products, but these moments alone may not be sufficient if ground clutter or other contamination is important.