Boost Your Skills with On-Site Courses Tailored to Your Needs

Transcription

1

2 Boost Your Skills with On-Site Courses Tailored to Your Needs The Applied Technology Institute specializes in training programs for technical professionals. Our courses keep you current in the state-of-the-art technology that is essential to keep your company on the cutting edge in today s highly competitive marketplace. Since 1984, ATI has earned the trust of training departments nationwide, and has presented on-site training at the major Navy, Air Force and NASA centers, and for a large number of contractors. Our training increases effectiveness and productivity. Learn from the proven best. For a Free On-Site Quote Visit Us At: For Our Current Public Course Schedule Go To:

3 AESA Airborne Radar Theory and Operations Course Sampler Robert A Phillips AnnapolisStar@gmail.com Introduction Page 1

4 Objective Number 1 1) Learn how to interleave modes, intercept targets using advanced LPI techniques, and develop requirements for an AESA radar from the pilots point of view. AESA Radar engaging and launching missiles on three targets Introduction: Page 2

5 Objective Number 2 2) Present the theory of an AESA Radar and learn how to design the air-air and air-ground modes from the requirements up Antenna Receive Pattern using a diamond layout with true symmetric Dolph Chebyschev sidelobe weighting (from supplied Radar Theory ebook) Clutter Template Pulse Compress FFT Square law Detector CFAR Block Diagram for Search mode M of N Correlate Introduction: Page 3

6 Objective Number 3 3) Provide the simulations, tools, and references for putting the theory into practice 500 page interactive electronic book on class material including antennas, Space Time Adaptive Processing, Kalman filters, and automatic target recognition, with simulations & examples Win 7 Professional Radar mode design spread sheet with software AESA Radar Theory ebook "You cannot mandate productivity, you must provide the tools to let people become their best. Steve Jobs Introduction: Page 4

7 Some of the Questions to be Answered 1) How do you design and compute the performance for the AESA search modes? 2) How do you design an AESA mode to track 50 targets? 3) What is Space Time Adaptive Processing (STAP) and how do you design for it? 4) How can you use an AESA antenna to detect slow moving ground targets which are much smaller than the background clutter. 5) How do you design an automatic target detection and recognition mode? This sampler presents the top level charts from the course on how to answer these tough questions Introduction: Page 5

8 Sampler 1) Design of an AESA Medium PRF Search Mode AESA Radar in Medium PRF search Introduction: Page 6

9 MED PRF Search Block Diagram The Block Diagram for a MED PRF Search radar [Skolnick,fig 17.6] Sum Channel Altitude Speed Clutter Template [12] Smallest allowable Target size m 2 Skolnick Fig Compress FFT Square law detect CFAR Unfold Detects M of N Range Correlator We will use an AESA antenna and receiver with parameters like size, noise figure, power and cooling appropriate for a fighter type aircraft (from Stimson) to design the modes and compute the performance M of N Doppler Correlator Target Reports Range, Doppler, Cross Section Introduction: Page 7

10 Clutter Template from Supplied Simulation Tail aspect Head In this region the template tells us we should use a backend STC or guard channel In this region we are competing with altitude line Use special processing to blank returns In this region the template tells us we are competing with noise only and we can use the noise PFA threshold. In this region we are competing with Main Beam Clutter. Due to its magnitude we will use a notch filter The template [12] guides us in choosing a CFAR design Introduction: Page 8

11 Baseline MED PRF Search Parameters Derived parameters Parameter Value Comments FFT Size 512 Controls S/N and scan rate PRF 70KHz For good tail aspect visibility CHIP 0.5mics For reduced clutter PCR 4 Higher average power M of N 3 of 7 Range correlation TFA 30sec Specification time between FA s Freq Agile Look-Look Good LPI design Xmit Pulse 2mics Duty 14% Avg Power 471watts Pfa 5.8E-6 CFAR probability of false alarm The course will show the student how to select the parameters and enter them into the Mode Design spreadsheet Introduction: Page 9

12 MED PRF Single Scan Performance Single Scan PD - Low PRF.VS. Medium PRF Cross Section 5m 2 Chart from the mode design spreadsheet using VBA software from the ebook on Detection Theory (supplied with course) The Mode Design Spreadsheet 1) Guides the student in the designing a mode, 2) Captures the designs and 3) Compares the performance for different configurations Introduction: Page 10

13 Sampler 2) How to Track 50 Targets with an AESA Radar AESA Radar engaging and launching missiles on six targets Introduction Page 11

14 Vector Tracking Loop Error k ANT Steering vector k(α,β) k T (θ,φ) target Compute Monopulse Error k ANT k(α,β) antenna steering For monopulse vector processing see [9] Haupt and ebook on antennas Transform To NAV Coords Transform to ANT Coords Kalman Filter in NAV Target Relative Position General radar tracking loop Ownship position vector in NAV reference The Σ and channels are used to compute the error vector k in ANT coordinates Introduction: Page 12

15 Three Channel (Az,El,Range) Kalman [1] Extrapolate XΦX = For n in 1..3 P = ΦPP Q+ n n n n Gain Computation For n in 1..3 T ( ) 1 K = PH HPH + R T n n n n Where n is one of the 3 orthogonal channels Rng, Az, El P Update For n in 1..3 P ( 1 ) = KHP n n n State Update For n in 1..3 X= X+ KE n Nav n The lectures will define each matrix in the design Introduction: Page 13

16 Typical Track Performance RMS Velocity Error Angle Error Tracking a Steady 3G S Turn at 20nm. RMS velocity errors typically approach 200+ft/sec and are entirely adequate to guide missiles to intercept Introduction: Page 14

17 AESA Time Line 15 Target track interleaved with search while displaying a SAR image. Room for lots more!! Introduction: Page 15

18 Sampler 3) Space Time Adaptive Cancellers Introduction: Page 16

19 Space Time Adaptive Filters (Stimson, Haupt) The STAP canceller can remove multiple sidelobe jammer(s) without prior knowledge of the jammer(s) location or antenna gains. STAP uses an Interferometric (space based) canceller. For each expected jammer we need one receiver and Auxilliary antenna with a gain larger than the sidelobes of the main antenna. Adaptive Cancellers Stimson [3,Ch 40], Skolnick[4,Ch 9] Gain of AUX Target Standoff sidelobe jammer STAP computes jammer phase angles and antenna gains and applies a spaced based adaptive notch filter. By combining this with an FFT to separate moving targets we have a two dimensional Space Time adaptive filter Introduction: Page 17

20 The Adaptive Canceller [7] Elbert V V V See also Stimson [3,Pg509] Main AUX 2 AUX n Store samples from each channel in the rows of the H matrix H=[m a 2 a 3 a n ] X 1 X 2 X n The optimal weights X are the 1 st column of the inverse of the covariance matrix (H T H) -1 Sum the weighted outputs of the multiple antennas to cancel the jammer. Note the order of the matrix inverse is equal to the number of channels i.e. two channels means we have to invert a 2x2 matrix The space filter is a direct application of linear estimation theory [7] Introduction: Page 18

21 Example of STAP With Multiple Jammers Example of STAP with 4 Jammers. 4 Aux horns Target 10deg 20deg 30deg 40deg See ebook on Antennas for detailed simulation of multiple jammers Weighted Sum The optimal weights are: x=1st Column of CovarianceMatrix 1 The cancelled jammer output equation is: Output=Main+x Aux +x Aux +x Aux +x Aux One 5 th order Matrix Inversion and 25 dot products of length 10 Introduction: Page 19

22 FFT Before and After Cancellation The target cannot be seen in the FFT with 4 Sidelobe jammers. Notice the magnitude of the noise at 100 Q or more! After cancellation the target is easily seen in the FFT and the noise is down to 5 quanta Uncancelled Jammer + Target Cancelled Jammer + Target Example from ebook on Antennas Introduction: Page 20

23 Sampler 4) Slow Ground moving target indicator Main Beam Clutter Canceller Introduction: Page 21

24 Slow Moving Target Detection Combining the Interferometer technique (used in STAP) with multiple antenna beams we can implement a high performance mode to cancel main beam clutter and detect small slow moving targets in a situation which otherwise would be completely hopeless SAR display with outputs from the slow moving target detector One of the most impressive applications of an AESA canceller.. Introduction: Page 22

25 Spatial vs Frequency Filtering Tail aspect Head Frequency Filtering: With an FFT we can separate targets with different Doppler frequencies. This fast moving target is separated by frequency from main beam clutter and is easily detected with an FFT FFT range/doppler map Spatial Filtering This slow moving target, overwhelmed in an FFT by main beam clutter at the same frequency, can only be detected by spatial filtering with an interferometer The course will describe this essential diagram in detail Introduction: Page 23

26 Slow Moving Targets and Clutter Stationary target at angle θ t Large MBC Clutter at angle θ c In a space diagram the target and clutter are separable θ t θ c Angle Space Map Slow moving target at angle θ t Whereas in a normal FFT frequency diagram the target and clutter overlay each other and the smaller target cannot be detected Doppler Frequency Space Map A Spatial Notch with multiple antennas can remove the clutter Introduction: Page 24

27 Slow Mover - Canceller [Stimson Pg321] Get α,β for each FFT Cell Get Gain for each FFT Cell k ( θφ, ) k (, ) T The target at the same frequency as clutter -d/2 d/2 Left Rg x Filter matrix Cancel Clutter Right Rg x Filter matrix MBC αβ Recompute FFT The phase for clutter at angle αβ, : π d G ϕc = R k= sin( α)cos( β), G rel = λ G Using the canceller equation: GM Output = Main - Aux exp( j2 ϕ) G The cancelled clutter for each filter is: Cancelled = Left Right A little complicated but very powerful MBC comes from a known angle α,β n n CFAR A n Left Right exp( j2 ϕ c ) Slow moving ground targets Introduction: Page 25

28 Sampler S ATR Finds 3 S-300 Surface Air Missile Launchers with Pd>0.95 in 2 sec S 5) Automatic Target Recognition Target Detection S Bushehr nuclear power plant from Google Maps Introduction: Page 26

29 Automatic Target Detection Outline [13] SAR Targets + Clutter Data from MSTARS public website, algorithms from Lincoln labs and Mathcad image processing library CFAR Detector Binarize Image Open/Close Shapes Clumped Detects Get Enhanced Tgt Chips Detected targets sans clutter Edit Clutter False Tgts Compute Moments Statistics Library Clutter Shadow Removal Target Recognition Target List Detector uses general target signatures to find military like targets Introduction: Page 27

30 Theory of Moments from [11] HU Characterization of an image by statistical moments like variance, and kurtosis, and invariant moments like the eigenvalues is a common approach in ATR. The Uniqueness theorem states that you can completely reconstruct an image with knowledge of the moments of the image. If you use amplitude, translation, scale and rotation invariant moments you increase the power of this approach E All three E s in this example are uniquely identified by the same simple moments which are independent of where they are on the paper, their amplitude, scale or rotation We can also characterize tanks, trucks and guns by moments Introduction: Page 28

31 Example Automatic Target Recognition[13] Enhanced M113 Chip from ATD with feature vector consisting of moments, stats and pose Library Chips with same pose as detected target Pose=-30deg 1) Use the pose to index the library 2) Compute Score for each target in the library using feature vectors 3) The highest score is the ID Feature Vec BTR60 M113 BMP2 BTR70 T72 M109 M2 HMMW M1 Correlation Good Match Eigenvalues Area Combined Comparison of feature vectors for each target in library Introduction: Page 29

32 References 1) Decoupled Kalman filters for phased array radar tracking: Automatic Control, IEEE transactions on: Date of Publication: Mar 1983 Author(s):Daum F. Raytheon Company, Wayland, MA, USA 2) Blinchikoff and Zverev, Filtering in the Time and Frequency Domain ) Rabiner and Gold, Theory and Application of Digital Signal Processing ) Stimson, Introduction to Airborne radar ) Skolnick Introduction to Radar ) William Skillman Radar Calculations Artech House,1983 7) Estimation and Control of Systems Elbert 1984 Contains all aspects of linear estimation from least squares to the Kalman filter 9) Antenna Arrays - Randy Haupt IEEE Press 10) SDMS MSTARS Public Data Website Contains 1ft SAR images of military targets 11) M.-K. Hu, Visual pattern recognition by moment invariants, IRE Trans. Information Theory, vol. 8, no. 2, pp , ) Radar CFAR Thresholding in Clutter and MultipleTarget Situations Hermann Rohling AEG-Telefunken, IEEE Transactions On Aerospace and Electronic Systems VOL. AES-19, NO. 4 JULY 1983 Discusses clutter maps for describing clutter regions of differing clutter type. Excellent analysis of CA, GO CFAR and ordered statistic CFAR Introduction: Page 30

33 References 13) MIT Lincoln Lab Journal Archives Vol 10, Number Vol 8, Number Vol 6, Number Provides overview of the Automatic Target Recognition and Detection including Super resolution SAR, CFAR s and effects of polarization and resolution on recognition Introduction: Page 31