Optimum Signal Processing for an Adaptive Communications System.

This report contains the results of a study of the feasibility of developing an adaptive communication system based upon sampling in frequency of the propagation path transfer function. An optimum method of sampling and adaptive filter implementation is developed based upon the Sampling Theorem. The spacing of test lines CW frequency components is shown to be the reciprocal of the maximum spread in delays of the multipath components. This results in the number of test lines being approximately equal to the product of the spread in multipath delays and the information signal bandwidth. Typical propagation models ranging from HF ionospheric circuits to microwave troposcatter circuits were all found to have bandwidth delay products of the order of two to twenty. The bandwidth required for each test line was found to be of the order of the doppler spread of a CW signal transmitted over the path and is less than 10 Hz for most propagation modes. The resultant fractional reduction in bandwidth available to the signal is of the order of the product of the doppler spread and the delay spread and is typically less than ten percent. This reduction in effective bandwidth is considerably less than that which results when other techniques, such as diversity or error correcting coding, are used to compensate for propagation path distortion. It is concluded that the technique is feasible and a prototype development and testing program is recommended. Author