Study Of Signal Reconstruction For Distributed Shipborne Surface Wave Over-the-Horizon Radar

Distributed multi-shipborne surface wave OTHR (Over-The-Horizon Radar) isa new radar system which is used for over the horizon sea detection. This kind ofradar not only has many advantages of general ground wave over-the-horizon radar,such as strong anti-jamming capability, anti-stealth and anti-low-altitude assault,but also has an unparalleled advantage in viability, mobility, and the detection range.Through the reasonable signal processing, it can overcome the traditional singleshipborne radar system’s low azimuth accuracy and poor matching ability inmulti-target type. However the great performance also makes the system structureand signal processing more complicated. Currently the distributed shipboard OTHRsystem research is in its initial stage, there are many key technologies needed tosolve. One of the problems is how to realize the cross-platform signal coherentprocessing with the signal and system construction.Distributed shipborne surface wave over-the-horizon radar (DS-OTHR) locatesradars on different ship platforms by reconstructing the signal-space to get thevirtual array, and then solve the above problem by signal processing on the array.This paper will study the construction of a virtual array in depth.In this paper, firstly, established a real-time positional relation between thereceive and transit platform on the basis of the dynamic geometric structure of thedistributed multi-shipborne OTH radar. Then analyzed the feature of FMPCWsignal and derive the beat frequency expression of the echo signal, and given thephysical significance of the beat frequency. Secondly, understood thetwo-dimensional FFT which is the typical HF ground wave radar signal processingmethod. Studied the Doppler broadening caused by the random dynamic structure.Then, assuming the platform is not moving got the constructed virtual array ofdistributed systems based on the rotational and translational transform. Howeverthis method is restricted in the scope of application. So we introduced arrayinterpolation technology. In order to reduce the interpolation errors, we chooseseveral non-overlapping narrow angular sectors which cover the pre-estimate ofDOA as the transform domain. Thus, we got the complete virtual array with greatcontrol in errors. At last, by simulating with computer, it proves that virtual arrayretain a better performance.