
MULTIRADAR TRACKING OF VHF RADAR PLOTS
"Scientific results cannot be used
efficiently by soldiers who have no understanding of them, and
scientists cannot produce results
useful for warfare without an understanding of the operations".
Dr. Theodore von Kármán
Abstract
Detection of flying objects with small radar cross section (Uninhabited Air Vehicles and Cruise Missiles  UAV&CM) and/or operating on low altitude is more effective by VHF radars than by radars operating in the "L" or "S" band. This is because STEALTH technology is ineffective in these bands. However central data processing (Multi radar Tracking) of the plots of these radars is difficult compared to that of radars operating at higher frequencies. The most important problems seem to emerge from the inaccurate measurements, high false plot rate, fragmentation of the target signals (plots) and poor jamming resistance.
This presentation offers a brief survey of the requirements of tracking and the peculiarities of VHF radar plots.
End of abstract
Thank you for this opportunity to describe our recent activities in the field of Multiradar tracking (MRT) of VHF radars.
To do so I shall address the following areas:
Under background, I intend to briefly review key events over the past decade, which collectively have helped to influence the development of new type of VHF radar network requirements.
I will then focus on specific requirements of modern airdefence system and the progress we are making in staffing changes to these requirements.
Lastly I will summarise and outline the way ahead and I will be ready to take your question.
But before commencing the briefing, I would next draw the conference's attention to this partial list of references, which essentially forms a baseline for much of the modern airdefence logic. We also intensively used them in the VHF radar requirements review progress.
Next, for better understanding why Hungarian military engineers are engaged in this progress, it is important to recognise that the catalyst for our activity is based on our history.
First, there were significant changes in the word radar astronomy when Zoltan Bay (Fig.1) measured the distance between the Moon and Earth. (Fig.2)
Figure 1.Zoltan Bay
Figure 2. Moon radar
Then the world famous mathematician Dr. Theodore von Kármán (Fig.3.), whom words should be one of the slogan of our conference. He achieved outstanding results in developing aviation and in buildingup modern missiles and who since 1951, had been the President of the AGARD of NATO.
Figure 3. Dr. Theodore von Kármán
Very quickly now, there have been a number of key threat which have affected the evolution of requirements of modern air defence. Well known where the airdefence activity starts. Firstly we should detect the target and build up the track as quickly as possible. Reference 3 has already analysed detection, tracking and Identification possibilities of Uninhabited Air Vehicles and Cruise Missiles.
I would like to drive your attention on this new type of threat. E.g. there is the AS17 UAV which has very small RCS, capable to fly only 200 m above the surface from the distance 200 km with high speed. (Fig.4.) This UAV is capable to manoeuvre with 10 G for a couple of minutes.
The Gulf War has shown how dangerous TBM could be and how new types of EW techniques are important.
Which kind of Sensors can detect UAV & TBM and which kind of AD C2 system MRT could track them?
In the context of these huge initial changes to the RCS situation (Fig. 5.) we should significantly change our air defence structure, first of all radars to make possible detect UAVs and TBMs and we must looking for the new MRT capabilities for AD C2 systems.
Figure 5. Example Radar Crosses Sections of "S" or Higher Band Radar
It is well known that the effectivity of surveillance radar network coverage depends on the degree of overlap of our radar network, which is usually high. We have to pay much more attention on VHF surveillance radars, than we did it before, because this type of radar is unique in its detection performance against new threats as well. (Fig. 6.)
Figure. 6. Overlapping of radar coverage
In the Fig. 7 we can see a wartime search radar scenario. In the radar coverage of the radar there are simultaneously operating UAVs and CMs, different type of aircraft, sort and medium range TBMs and the environment isn't friendly. I will point out that a lot of military leaders disagreed such dramatic change over 10 years ago, well before subsequent events would reinforce the wisdom of their peaceful logic. Operations became much more robust from an air command control perspective and need new type of radars and MRT principles.
Figure 7. Wartime Radar Surveillance Scenario
Application of metric wavelength (VHF band) possesses several benefits on various areas of surveillance radar technique.
Measurements at resonance frequencies prove a significant reduction of efficiency of stealth techniques at low frequencies. VHF frequencies, due to diffraction of propagation have the ability to detect low flying targets sometimes beyond the horizon.
On the other hand, there are several problems emerging from the application of metric band radars.
In case of LPI radar stations, built up on the basis of multistatic theory the high bandwidth is also an important requirement.
At the same time, reducing vulnerability of radars to signal detection by an intercept receiver, is one of the major objectives of modern radar design. The basic requirement for an LPI radar (to see and not to be seen) leads to the task of detecting targets employing signal with minimum radiated Spectral Power Density and using waveform modulation formats that make difficult to the intercept receiver to identify signal. This task has traditionally been performed using spreadspectrum signals. From the near future, the development of radars using FrequencyHopping (FH) method is expectable. These techniques also need high frequency agility and bandwidth.
In summary, in the recent years an increased demand for broadband VHF antennas has appeared.
On Fig. 9. you can see 3D Super long range earlywarning radar concept based on Oborona radar antenna system. There are some auxiliary antennas for SLC and SLB channels and a large primary antenna reflector here.
P14, (Oborona) modernisation technical performances are the following:
600 km, 1200 km for space and ballistic targets.
Figure. 9. P14 (Oborona)
There are feeding dipoles of low and high antenna beam formers controlled mechanically by operators. In our concept these feeding dipoles are used simultaneously for forming elevation monopulse beams.
On Fig.10. you can see how we would like to use two additional feeding dipoles for forming azimuth multibeams for azimut monopulse.
Figure 10. Monopulse beams for elevation and azimuth.
MultiRadar Tracker is the HEART of the AirDefense system
Airdefence experts are well aware that of MultiRadar Tracker (MRT) of AirDefense Command Control (AD C2) system after detection of targets on the basic level determines operational possibilities of airforce. If the radar measures positions of the target in the welldetermined discrete time frame, we can predict next coordinates of the target using MRT techniques. All kind of measurement has error, to compensate or reduce it we use MRT smoothing techniques. See Fig.11. (See Ref. 1.; Ref. 4.)
Figure 11. Target track smoothing and prediction
All kind of MRT practically based on these equations (See Ref. 2.):
Where: Y_{n}= Range measured at time t_{n};
= Predicted value of range and range rate at t_{n+1}.
Based on measurements made up to time t_{n}.
h_{n};g_{n} = Prediction equation smoothing constant;
t_{a} = time between measurements (data update time )
Almost all recently operational MRT filters are the following:
How do these filters differ? They differ in the selection of the weighting coefficients g and h. For some of these filters g and h depend on n. This is the case for the Kalman filter.
The most popular among them is the Kalman filter. Why is the Kalman filter so popular?
Target dynamics can be used directly to optimise filter parameters.
Makes use of a priori estimates if available.
Permits optimum handling of measurements
none quell time between measurements.
The next Kalman equations show the complexity of these equations.
Predictor equation :
Where:
Filtering equation :
Weight equation :
Where: Y_{n}= measurement matrix; R_{n}= observation noise covariance;
Predictor covariance matrix equation:
Covariance of random system dynamics model noise vector:
Covariance of measurement vector if
Corrector equation (covariance of smoothed estimate) :
Where: I = Identity matrix
Signal and plot specialities of the VHF radars:
Figure 12. VHF radar Plan Positioned Indicator
Does the Kalman filter give us optimal prediction?
What does optimal mean?
Does the Kalman filter apply to this problem?
Case where KALMAN filter is not optimum:
Non linear measurement.
Non linear target dynamics.
Case where Kalman filter is not optimum in the VHF radar case:
How can we do Kalman filter better?
Let's see a simple estimation problem. If x – unknown constant scalar and we want to estimate x given as a sequence of measurements. Measurements are corrupted by additive noise, which is statistically independent on sample to sample.
Some experts use so called extended Kalman filter solution, but it needs linearisation.We can use optimal non linear filter estimation as well, but it is much more complex, than Kalman filter equations. In this case for the simple estimation:
There are some other solutions for solving MTR problem. E.g. Multiple Hypothesis Tracker (MHT) where splits existing track when association ambiguity exists and follows each branch with probability calculation. (MHT very good at crossing targets tracking.) Or Probabilistic Data Association Filter (PDAF), but we can do only preliminary conclusions:
The key ingredient is the SHAPE of the conditional density.
From the portal conclusions we have already seen that VHF radar has more than twice better detection range than modern "S" band radars with many advance detection performances. (See Fig.13.) We have a low risk solution for solving emerging problems of Multiradar tracking of VHF radars.
Figure 13. Detection range of "S" and "VHF" radars
Final conclusion:
We have to analyse thoroughly the possibilities of the VHF radar system in the field of centralised MRT, against new type of ABM & TBM targets as well.
We have to optimise and harmonise VHF radar MRT opportunities with other band radar possibilities.
We must work out our concepts (modelling, threat database, and validation procedure) in strong coordination with the users.
References: