In multi-static radar (MSR), the transmit/ receive aperture is divided into a number of sub-apertures that can be placed in various locations relative to each other. These locations can be chosen to optimize the performance of the radar in terms of some specific task. Two multi-static approaches have been investigated:

  • Closely spaced apertures – distributed aperture radar (DAR)
  • Widely spaced apertures

Realizing the greater capability of MSRs requires unique waveform and signal processing approaches. A computer simulation has been developed that permits the analysis of MSR signal processing.

Figure 1. Multi-Static Imaging

Multi-static radars, in a distributed aperture mode, can potentially provide significantly improved target tracking because of the large baseline between the various apertures. The resulting angular resolution can be orders of magnitude better than the resolution of a monolithic system (single large radar). This capability comes with a cost because of the resulting grating lobes (multi-statics with evenly spaced apertures) or high side-lobes (multi-statics with randomly spaced apertures).

The same angular resolution can provide improved electronic protection (EP) capability. For a single-aperture radar, jammers located near targets of interest cannot be nulled without impacting the antenna main-beam and therefore, the target returns. But the multi-static system, with its very-long baseline receive aperture gain on the target can be maintained while a deep null is placed in the direction of the jammer.

Imaging and Discrimination

Two-dimensional images of moving targets can be obtained through inverse synthetic aperture radar (ISAR) processing. The range and cross-range dimensions of radars viewing the target from widely separated angles will achieve target- centered resolution in different dimensions. For example, two radars independently viewing an object in its plane of motion (linear, rotating) with 90º of separation will provide complementary information: the range resolution of one radar will be the cross-range resolution of the other, and vice versa.

Coherent fusion processing of the data from these two radars can provide improved resolution. Fusion of the data from the bi-static path can further improve the resolution. Also, two or more radars viewing an object from different angles not in its plane of motion can provide three-dimensional images. The overall 3D resolution of the object will be a function of the range and cross-range resolution of the individual radars and their angle separation as viewed from the target location.

Multi-Static Interference Rejection

Figure 2. Coherent Fusion Imaging Experiment. Coherent image for the test target consisted of two large scatterers on a vehicle that was driven throughthe radar’s field of view.

Interference can be rejected only if the target and interference are resolvable in the dimensions/domains in which the processing is being performed. Continuously radiating point sources (jamming) can be rejected in the spatial dimension if the target and EMI are separated in angle, and cannot be rejected when that separation is sufficiently small. In general, spatially continuous interference cannot be adequately suppressed by conventional non-adaptive means, i.e. by processing separately in either the spatial or Doppler domains.

These techniques fail because they do not handle the space-time coupling inherent in the clutter signal return. Consequently, leakage from one domain to the other limits the amount of suppression that can be achieved by operating in these domains separately. For conventional single-aperture radars, the cross-range resolution may be so large that the target effectively falls within the main beam antenna spatial response.

In this case, conventional space-time adaptive processing (STAP) will not be able to adequately reject the interference. However, assuming a distributed aperture radar with high range and cross-range resolution, improvements in target-clutter separation is achievable along with improvements in interference suppression. Such architectures generally lead to space-time grating lobes that can degrade performance. Using simultaneous orthogonal waveforms to form narrow spatial main beams, it is possible to develop space/time/waveform adaptive processing to suppress grating lobes, reject the clutter, and detect the target in jamming, clutter, and joint jamming/clutter environments.

The potential for orthogonal waveforms in distributed aperture radar architectures (DAR) in achieving improved resolution, interference suppression, and target detection and tracking performance while simultaneously controlling space-frequency grating lobes was demonstrated. System operation involves radiating orthogonal waveforms from multiple sub-apertures of the DAR and then receiving and processing these waveforms at each subaperture. The use of orthogonal waveforms provides an additional dimension (waveform) beyond the standard spacetime dimensions typically used in conventional STAP for adaptive suppression of the interference background.

Adaptive processing using frequency diversity was simulated and demonstrated. The interference was first modeled as a single-point EMI source. The signal theory was then generalized to handle the more difficult problem involving distributed volumetric clutter.

EMI Rejection Simulation and Analysis

Several key considerations for systems employing advanced adaptive processing techniques using waveform diversity were considered. A distributed aperture system with N sub-apertures was assumed. Each sub-aperture is assumed to transmit a different (orthogonal) waveform. Each sub-aperture then receives target returns from each of the transmitted waveforms, resulting in a total of N × N returned samples for each radar range gate. In this situation, the classical space-time data cube is replaced by a data hypercube where the additional dimension is “waveform”. In this effort, the orthogonal waveforms were chosen to be relatively narrowband signals offset in center frequency.

Traditionally, adaptive processing has focused on the spatial and temporal dimensions leading to space-time adaptive processing. The spatial steering vector is related to the spatial look direction while the temporal steering vector is determined by the temporal look or Doppler frequency. In this case, the space dimension is augmented by the addition of the waveform domain. The space/waveform steering vector is determined uniquely by the look angle with a different spatial steering vector for each transmit frequency.

Distribution of the sub-apertures and separation of the transmit frequencies introduces two new degrees of freedom to the radar designer; namely, the spacing between the antenna sub-apertures and spacing between frequencies. Generally subapertures and/or sub-bands are separated by multiple wavelengths. Consequently, equally spaced elements/sub-apertures with equally spaced frequencies can lead to grating lobes that can reduce the effectiveness of the adaptive process. Analysis has shown that appropriate distribution of the elements in the spatial and temporal (frequency) domains along with weighting can eliminate, or at least mitigate grating lobes in their respective dimensions and reduce the amount of adaptive processing required to suppress interference.

Another concept borrowed from bistatic radar applies a spatial unwarping to the data to remove the range dependency and allow a larger secondary data set size to be used. Based on these considerations a high-fidelity, multi-static radar simulation was developed and the performance of various geometries predicted.

Jammer Rejection Experiment

A rooftop experiment was accomplished at the Air Force Research Laboratory/Rome Research Site to verify the multi-static radar simulation. Five subapertures were located in roughly linear orientation. The total separation was about 200 feet. A moving target and a jammer were located about 6,000 feet away (Figure 2). The target was driven through the main beams of the five radars.

The suite of five transmitters collectively radiated five diverse frequencies that were recorded on each receiver channel. Activation of the jammers produced a 30- dB jammer/noise ratio and completely masked out the target vehicle return. After multi-channel, multi-waveform processing, jammer cancellation occurred along with target detection. Data was collected and adaptively processed as the target vehicle traveled along a course that brought it past the jammer. The SNR decreases significantly when the target and jammer are in the same direction. Had the sampling interval been smaller than 0.3 milliradians, the null may have been deeper.

Exo-Atmospheric Volume Clutter

The simulation was extended to model exoatmospheric clutter in addition to jamming. The simulation generates an individual scatterer model that keeps track of particle position and velocity as a function of time. These scatterers are range-gated and parsed out of the scatterer cloud. Ranges are then calculated for a set of spatially diverse subapertures, each transmitting an orthogonal waveform relative to each other. Each sub-aperture receives each transmitted waveform. Therefore, assuming Na sub-apertures, there are Na2 different mono-static and bi-static radar combinations (Na monostatic radars and Na–Na bi-static radars). The range and frequency information are then used to calculate phase difference values for each of the Na2 radars.

The resolution of a SAR or ISAR image is a function of how much of the Fourier space the measurements sample. The bandwidth of the ISAR measurement transforms to the radius in two-dimensional Fourier space. Bi-static measurements are more complex with the transform of the frequency being a function of that bandwidth and of the bi-static angle. Figure 1 shows the Fourier sampling for three sensors (two radars operating both mono-statically and bi-statically). The aqua and blue sectors are the Fourier sampling for the mono-static operation of these two radars. The green sector represents the bi-static sampling. For this geometry, the combined multi-static image would have a resolution 2.5 times better than a single mono-static image.

Rooftop Experiment

This experiment was also performed at the Air Force Research Laboratory, Rome Site in Rome, NY. Each radar was equipped with programmable waveform generators, frequency conversion equipment, timing and coherent local oscillators based on GPS receivers, as well as data recording servers with storage, processing, and display capability. A vehicle with two dominant scatterers was driven along the road in the target area. The objective of the experiment was to demonstrate an improvement in radar imaging capability by using data from both radars compared to mono-static data from a single radar. Imaging results are also presented in Figure 2.

Compared to conventional radars, multi-static radars have the potential to provide significantly improved interference rejection, tracking, and discrimination performance in severe EMI and clutter environments. They can potentially provide significantly improved target tracking accuracy because of the large baseline between the various apertures. The resulting angular resolution can be orders of magnitude better than the resolution of a monolithic system.

Two-dimensional images of moving targets can be obtained through inverse synthetic aperture radar processing. Coherent fusion processing of the data from multiple radars can provide improved resolution. Also, two or more radars viewing an object from different angles not in its plane of motion can provide 3D images.

This article was written by Russell Brown, Yuhong Zhang, and Richard Schneible of Stiefvater Consultants, Rome, NY; Michael Wicks of the US Air Force AFRL/RY; and Robert McMillan of the US Army, SMDCRDTC- TDT. For more information, Click Here