Cognitive Adaptive Array Processing for Radar

Usually adaptive array processing is presented with exotic matrix equations that are difficult to understand and do not give a physical feel to what is going on so one can improve on the jammer suppression. This tutorial is present the subject from physical point of view which gives one great insight into the best way to do jammer suppression. We show how to do Cognitive Adaptive Array Processing (CAAP). The immense power of CAAP is demonstrated. Digital beam forming (DBF) makes CAAP the processing of the future.

Barrage Jammer: It is shown how CAAP reduces by orders of magnitude the computation complexity, the number of training samples needed and the adapted antenna sidelobe degradation. For example assume an N = 10,000 element array and J = 1 barrage noise jammer. With the classical Sample Matrix Inversion (SMI) one needs to have K = 5N = 50,000 training samples to get a signal-to-interference (SIR) within 1 dB of optimum. Plus it requires the inversion of an NXN = 10,000X10,000 interference matrix which necessitates on the order of a billion operations (multiplies and divides). Also it degrades many of the adapted antenna sidelobes by 30 dB for a 40 dB un-adapted sidelobe level pattern. In contrast with CAAP it is shown that only K = 4 training samples are needed, only the two sidelobes next to the jammer are degraded and only about 2dB, and no matrix inversion is needed, only 13 operations for a reduction of 100 million operations. It is shown that CAAP lets us go from the complex SMI processor to a simple sidelobe canceller (SLC) having one aux antenna beam for the case of one jammer. Similar advantages are shown for the case of J jammers. With J jammers we go from an SMI canceller to a SLC having J high gain aux beams pointing at the jammers. To estimate the number and location of the jammers modern estimation and super-resolution methods could be used. It is shown that SMI actually does in the math what CAAP is doing, except that it does it very inefficiently or with poor performance or both. Like CAAP, SMI actually does SLC in the math but transparent to the user. It is pointing beams at the jammers using them to do SLC just like with the CAAP. However because the SMI does not have a perfect estimate of the NXN size interference covariance matrix ME, it also points beams at N-J-1 directions where there are no jammers and as a result degrades the adapted antenna sidelobe level where there are no jammers. For SMI these jammer beams are called eigenbeams. They result physically from the eigenvectors of ME. Amazingly Sidney Applebaum pointed this out in his seminal paper and report back in 1974. Covered will be the expressing of the adapted antenna pattern in terms of its eigenbeams, which is known as the principal component formulation. It is pointed out that CAAP represents a very fruitful area for future study. One area is coping with different jammer combinations in the sidelobes and main lobe. For example Gabriel in his original work showed that if you had two jammers of nearly equal strength close together in one sidelobe the eigenbeams are a sum beam and a difference beam [9, p. 185]. Probably having two squinted main beams for the aux would do just as well in general. The advantages that may be gained if the J jammers are separated from each other by more than a beamwidth should be explored. A lot can be learned for cognitive adaptivity by doing simulations for different cases. This is now easy to do using Mathworks MATLAB and Phased Array System Toobox.

Hot Clutter Jammer: We show how with hot clutter present we can use CAAP to cancel out the jammer even if it is coming into the main lobe where we are looking for a target. We show how with CAAP we cancel out main beam jammer without any loss of signal strength. Main lobe jammer cancellation without signal cancellation using CAAP. Amazing.

Repeater Jammer: Apply CAAP to repeater jammers using locations of repeaters to spoof them and lower the radar signals they can see.

Low Probability of Intercept: Apply CAAP to it.

MIMO vs Conventional Array Radars: Compare these two re the above jammers. Examine both Ubiquitous and Machine Gunning Conventional Radars. These conventional radars detailed. Results applied to thin/full and full/thin array systems.