| Adaptive processing algorithms on antenna arrays have become a necessity on many systems which are deployed in dynamic environments. Conventional algorithms may utilize assumptions which may not be realizable in practical situations. Such algorithms which rely on larger durations of training data may have trouble with blinking jammers, non-homogeneous clutter, and other types of adversarial effects which are not constant for a relatively large duration of time. In this paper, the Direct Data Domain Least Squares(D3LS) approach is studied as an alternative to relax some of these assumptions. Also, one of the fundamental assumptions which some conventional algorithms are relying on is an ideal environment for which the antenna is deployed. The case where an antenna is deployed over a real practical environment, such as a lossy earth ground or sea, is studied. A transformation method is discussed which will transform this non-ideal deployment into an equivalent free space point radiator model. Once we obtain a free space model of our particular situation, we are free to implement any algorithm, not limited to just the ones described in this paper, which is optimized with a free space model. The D3LS approach described allows us to simultaneously null out different signals at different frequencies, and, at the same time, create peaks for multiple signals at multiple frequencies, all within the same antenna array. Examples are illustrated for each step of the process which are followed by a detailed numerical example. Finally, some conclusions and future work prospects are discussed in the last section. |
