2. OBJECTIVE OF THIS WORK
• Optimal rules for compressing data at
each local sensor such that the fused
estimate is optimal.
• Data fusion with Estimated Weights.
• Implementation of these methods in LAN
in LINUX platform.
3. Data Fusion Algorithms
• Linear Estimation Fusion (LEF):
X = Pj (Pi+Pj)-1 Xi+ Pi (Pi+Pj) -1 Xj
• Weighted Least Square (WLS):
X=[HTP-1H] -1HTP-1y
5. Data Compression
• Lossless
• Lossy
Advantages in Data Compression
• Curse of Dimensionality
• Reducing communication overhead
6. Reduced SVD
Pc = U* S* VT*
• By keeping only the first k singular values of S
• And first k rows of VT and first k columns of U
• Pc is referred to as the Rank k Approximation of
P or the "Reduced SVD" of P
8. Assumptions
• Two non-maneuvering targets in a clutter
free environment.
• Zero mean, white Gaussian process noise
• Two radar system with same coverage area
9. RADAR 1:
• Scan rate: 0.5 sec
• Radar Freq: 2 Hz
• Scans : 150
• Location in LLA
– Latitude –10.502
– Longitude –80.1
– Altitude –10
• Standard deviation of errors in
– Range – 50 m
– Azimuth – 0.005759 radians
– Elevation – 0.0171radians
10. RADAR 2:
• Scan rate: 1 sec
• Radar freq: 1 Hz
• Scans : 75
• Location in LLA
– Latitude – 10.50
– Longitude – 80.1
– Altitude – 10
• Standard deviation of errors in
– Range – 100 m
– Azimuth – 0.005759 radians
– Elevation – 0.0171radians
15. Data Fusion with
Estimated Weights
• RP =TM.Rc.TMT
•
where,
CR
L =
TM 0 0
0 TM 0
0 0 TM
cos cos sin cos sin
sin cos 0
cos sin sin sin cos
m m m m m
m m
m m m m m
b e b e e
b b
b e b e e
TM =
16. Local Sensors
• XL(k) = CR
L XR(k)
• XL
^(k-1/k+1)= A XL
^(k/k) + KL(k+1)
{ZL(k+1) – H A XL
^(k/k)}
• XR
^(k-1/k+1) =A XR
^(k/k) +
CR
L KL (k+1) TM
{Z(k+1) – HAXR
^ (k/k)}
• KR(k+1) = CR
L KL(k+1)TM
17. Fusion centre (CMM)
• X^f(k + l/k + 1) = μiX^i (k + l / k + 1)
• μ l = PR
i (k+1/k+1) - PR
l(k+1/k+1)
(N-1) X PR
i (k+1/k+1
1
N
i
1
N
i
1
N
i
18.
19.
20. Implementation in Lan
Distributed Fusion
1. Without Compression Technique
2. With Compression Technique
3. Data Fusion with Estimated Weights
22. Method of Implementation
• Socket Programming in Linux Environment
1.Server
2.Client
• Server Socket
• Client Socket
• Now the processes can communicate
27. Techniques Available
• Based on retina model
• Influence factor modification and the
anova methods
• Multispectral multisensor image fusion
using wavelet transforms
• An estimation theory perspective
28. Techniques Available
• A pixel-level multi sensor image fusion
algorithm based on fuzzy logic
• A region-based image fusion method using
the expectation-maximization Algorithm
• Target tracking in infrared imagery using
weighted composite reference
• Function-based decision fusion
29. Military applications
• Automated target recognition
• Battlefield surveillance
• Concealed weapon detection
• Guidance and control of autonomous
vehicles
• Military command and control
30. Applications (Non military)
– Robotics
– Remote sensing
– Air traffic control
– Medical diagnostics
– Pattern recognition and environmental
monitoring
– Monitoring of complex machinery
– Artificial intelligence
31. REFERENCES
• Lei Wei Fong 2006, ‘Multi sensor data fusion with estimated
weights’ IEEE Transactions on Information Theory.
• Rong Li.X 2003, ‘Optimal Linear Estimation Fusion – Part I
,Unified Fusion Rules’,IEEE Transactions on Information Theory,
Vol 49,pp 2192-2207.
• Rong Li X. 2003, Jie Wang, ‘Unified Optimal Linear Estimation
Fusion – Part II Discussions and Examples’, IEEE Transactions
on Information Theory.
• Rong Li.X 2003, Keshu zhang, Peng Zhang, Haifeng Li ,‘Unified
Optimal Linear Estimation Fusion – Part VI ‘Sensor data
compression’, IEEE Transactions on Information Theory.
• Neil Matthew 2001 and Richard Stones, ‘Beginning Linux
programming’ 2nd edition.
32. REFERENCES
• Samuel S.Blackman1986 ‘Multiple Target Tracking with Radar
Applications’, Artech House Inc.
• Bar Shalom.Y 1995. and Xiao-Rong Li ‘Multisensor – Multitarget
Tracking.
• Bar-Shalom.Y1995., ’Multi target-Multisensor Tracking: Applications
and Advances volume II’ 1941- 621.3848
• Farina.A and Studer.F.A. ‘Radar Data Processing’, Research
Studies Press Ltd.1986.
• David L Hall2001, James Llinas , ’Handbook of Multi sensor Data
Fusion’ -- (Electrical engineering and applied signal processing) II.
Title. III. Series