Statistical optimization of antenna processing systems

This paper treats the antenna as an information processing device. By representing the antenna as a spatial (as well as temporal) frequency filter, information theory concepts can be applied so as to maximize the information content or rate at the system output or to minimize the information content in the system error. Rather than attempt to satisfy classical antenna design criteria such as maximization of gain, minimizing of side-lobe level, etc., the approach used here is to optimize the system by maximizing the useful information rate of the receiver subject to the physical constraints of the system. A convenient way of evaluating and comparing different signal processing systems is to determine their respective output information rates. Various non-linear as well as linear processing systems can be compared on this basis. For linear, one-way antennas, the optimum antenna in most cases is one which has uniform shading. For linear, two-way antennas, the optimum shading is non-uniform. The theory is illustrated with a nonreciprocal two-way array whose parameters are adjusted so as to yield the optimum antenna system according to the above criteria.