; methods
Documents
Resources
Learning Center
Upload
Plans & pricing Sign in
Sign Out
Your Federal Quarterly Tax Payments are due April 15th Get Help Now >>

methods

VIEWS: 31 PAGES: 32

  • pg 1
									          Institute for Parallel Processing - Bulgarian Academy of Science




Methods for Data and Information Fusion

                Kiril Alexiev, Iva Nikolova

                  alexiev@bas.bg
            Tel: 9796620; 0898 898 616
       25A, Acad.G.Bonchev Str., Sofia 1113,
                      Bulgaria
Seminar 22.11.2006                                                           1
          Institute for Parallel Processing - Bulgarian Academy of Sciences




      Correct decision making (taking) in the
      security sector mainly depends on
      information, received from multiple
      sources. Often, the information is
      insufficient, unreliable and
      contradictive.



Methods for Data and Information Fusion                                       2
           Institute for Parallel Processing - Bulgarian Academy of Sciences



        Architecture of sensor network
                                                                      sensor
                                                                       node
                                                            routing
                                                              data
                              Communication                  sensor
              query                                           data

               sensor
                data                                                  sensor
                                                         routing       node
    user
                                                           data
                                                          sensor
                                                           data

Methods for Data and Information Fusion                                        3
          Institute for Parallel Processing - Bulgarian Academy of Sciences


              Definition of Data and
               Information Fusion
Wikipedia:
Sensor fusion is the combining of sensory data
  such that the resulting information is in some
  sense better than would be possible when these
  sources were used individually.
Better = more accurate, more complete, or more
  dependable

Methods for Data and Information Fusion                                       4
          Institute for Parallel Processing - Bulgarian Academy of Sciences


                    Definition of Data and
                     Information Fusion
     Authors remark:
     In definition:
     “combination of data” is not very suitable
     phrase. We have to find better one, for
     example “simultaneously processed data”



Methods for Data and Information Fusion                                       5
          Institute for Parallel Processing - Bulgarian Academy of Sciences



          Benefits from Fusion Process
    The first and the most important remark is that
     fusion process is necessary most of all to reduce
     (to filter) input information through its
     integration (merging) and generalization.
    Fusion process is necessary to improve accuracy.
    Fusion process is necessary to reduce
     uncertainty.



Methods for Data and Information Fusion                                       6
          Institute for Parallel Processing - Bulgarian Academy of Sciences


Structure of Data and Information Fusion (JDL)
    Level 0: Preliminary data processing – pixel or signal
      level data association and characterization.
    Level 1: Data alignment, association, tracking and
      identification.
    Level 2: Situation assessment.
    Level 3: Threat assessment.
    Level 4: Process Refinement includes adaptive
      processing through performance evaluation and
      decision or resource and mission management.

Methods for Data and Information Fusion                                       7
          Institute for Parallel Processing - Bulgarian Academy of Sciences



                         Paper's classification by fusion level




                         5%   3% 2%                               Preprocessing
                   11%
                                                                  Level 1
                                                                  Level 2
             15%                                                  Level 3
                                                  64%             Level 4
                                                                  Others




Methods for Data and Information Fusion                                           8
          Institute for Parallel Processing - Bulgarian Academy of Sciences



                        Paper's classification by sensor type




                 26%                                            Radar
                                                       40%      Visual
                                                                Infrared
                                                                Acoustic
               15%
                                                                Others
                       4%            15%




Methods for Data and Information Fusion                                       9
          Institute for Parallel Processing - Bulgarian Academy of Sciences



                          Paper classification by target type



                                                           Air targets

                    12%                                    Mobile Robots
             14%                             36%
                                                           Ground and/or mobile
                                                           targets
                                                           Submarine
                   31%                  7%

                                                           Others




Methods for Data and Information Fusion                                           10
          Institute for Parallel Processing - Bulgarian Academy of Sciences




                                 Level 1




Temporal data fusion                              Sensor data fusion




Methods for Data and Information Fusion                                       11
          Institute for Parallel Processing - Bulgarian Academy of Sciences


             Data association methods
 The Nearest Neighbor method associates the nearest
 measurement to the track prediction. The more
 complicated Global Nearest Neighbor minimizes
 cluster cost function in measurement distribution.
The probabilistic data association filter (PDAF) and
 its extension to multiple targets – joint PDAF (JPDAF),
 solve the same task of measurement identification in a
 simpler way. In the JPDAF hypotheses are built for the
 measurements and targets only for the current scan. In
 this way the number of hypotheses is additionally
 reduced but the chance of combinatorial explosion in
 dense target and clutter scenarios still remains.

Methods for Data and Information Fusion                                       12
           Institute for Parallel Processing - Bulgarian Academy of Sciences


              Data association methods
 In Multiple Hypothesis Tracking approach all measurements
  received at a scan are assigned to initialized targets, new targets or
  false alarms. A number of hypotheses are generated. Every one
  supposes a possible assignment scheme between measurements,
  received in all scans, and the targets - confirmed, new ones or false.
  Pruning and gating techniques are used to retain the most likely
  hypotheses and in this way to reduce their number
 Finite Set Statistics considers all measurements as measurements
  from a generalized sensor and all targets as a generalized target of
  interest. Fusion of information from one and the same sensor but from
  different moments of time


 Methods for Data and Information Fusion                                       13
          Institute for Parallel Processing - Bulgarian Academy of Sciences



                             Identification
    Two types of identification:
     Structural identification – more difficult
      Define structure (model), which in the best way
     corresponds to the observed system (process).
     Parametrical identification – a lot of algorithms
      Find (calculate) values of parameters, which
     characterize entirely considered system (process).


Methods for Data and Information Fusion                                       14
          Institute for Parallel Processing - Bulgarian Academy of Sciences


              Parameter identification
              Mathematical description

Linear dynamic system (Markovian presentation):


  x(k  1)  F (k ) x(k )  G(k )u(k )  v(k )
  z(k )  H (k ) x(k )  w(k )
Kalman filter gives optimal solution for Gaussian noises


Methods for Data and Information Fusion                                       15
          Institute for Parallel Processing - Bulgarian Academy of Sciences



                           Description
Markovian - semi Markovian         x(k  1)  F ( x(k ), x(k  1),..., x(0))

Linear - non-linear dynamic system                   x(k  1)  F ( x(k ))

Additive - non additive system noise            x(k  1)  F ( x(k ), v(k ))

Gaussian - non Gaussian system noise                   v  N (v, mv ,  v )

Additive - non additive measurement noise z(k )  H ( x(k ), w(k ))
Gaussian - non Gaussian measurement noise w  N (w, mw , w )



Methods for Data and Information Fusion                                       16
          Institute for Parallel Processing - Bulgarian Academy of Sciences


 The simplest tracking filter, considered in the paper, is alpha-
  beta filter. It is suitable for tracking of moving with constant
  velocity targets without steady-state error. The alpha-beta-
  gamma filter has ability to track even accelerating targets without
  steady-state error.
 Kalman filter is a classical optimal estimating algorithm for
  dynamical linear system with Gaussian measurement and system
  noise. The modification of Kalman filter - Extended Kalman filter
  is developed for non-linear systems. The EKF gives particularly
  poor performance on highly non-linear functions because only the
  mean is propagated through the non-linearity. The unscented
  Kalman filter (UKF) uses a deterministic sampling technique to
  pick a minimal set of sample points (called sigma points) around
  the mean.


Methods for Data and Information Fusion                                       17
          Institute for Parallel Processing - Bulgarian Academy of Sciences


 The theoretically most powerful approach for manoeuvring
  targets tracking is known to be Interacting Multiple Models
  estimator. Generalized Pseudo-Bayesian (GPB) estimators
  different orders, Fixed structure IMM, Variable Structure
  IMM, Probabilistic Data Association IMM are variants. The
  most important feature is that all these estimators use in parallel
  several models for modelling of the estimated system.
 Particle filters, also known as Sequential Monte Carlo methods
  (SMC), are sophisticated model estimation techniques based on
  simulation. Particle filters generate a set of samples that
  approximate the filtering distribution to some degree of accuracy.
  Sampling Importance Resampling (SIR) filters with transition
  prior as importance function are commonly known as bootstrap
  filter and condensation algorithm.


Methods for Data and Information Fusion                                       18
                Institute for Parallel Processing - Bulgarian Academy of Sciences

          Temporal data fusion Alan Steinberg








 NN        Nearest Neighbor                    PF      Particle Filter
 F         Alpha-Beta Filter                   PDAF    Probabilistic Data Association Filter
 KF        Kalman Filter                       JPDAF   Joint Probabilistic Data Association Filter
 EKF       Extended Kalman Filter                      FISST       Finite Set Statistics
 IMM       Interacting Multiple Model filter   Y/N     Good/Poor Choice;

     Methods for Data and Information Fusion                                             19
          Institute for Parallel Processing - Bulgarian Academy of Sciences

     Temporal data fusion Alan Steinberg




Methods for Data and Information Fusion                                       20
          Institute for Parallel Processing - Bulgarian Academy of Sciences


      Fully Centralized Measurement
            Fusion Architecture




Methods for Data and Information Fusion                                       21
          Institute for Parallel Processing - Bulgarian Academy of Sciences


   Fully Centralized Trajectory Fusion
              Architecture




Methods for Data and Information Fusion                                       22
          Institute for Parallel Processing - Bulgarian Academy of Sciences

          Distributed Decision Fusion
                 Architecture




Methods for Data and Information Fusion                                       23
          Institute for Parallel Processing - Bulgarian Academy of Sciences



                    Simple Example
   When both sources are reliable, there is a consensus and
   it is reasonable to find solution in the cross-section of
   and - sets of corresponding sources: x  D1  D2. If the
   two sources do not agree, we have x  D1  D2 . The
   hypothesis for reliability sources is no longer credible
   and three other hypotheses appear: 1) First source is
   correct, the second is incorrect; 2) First source is
   incorrect, but second is incorrect; 3) Both sources are
   incorrect. How to find the correct hypothesis? As a
   precaution, all available information is kept and we hold
   up .

Methods for Data and Information Fusion                                       24
          Institute for Parallel Processing - Bulgarian Academy of Sciences


                 Example – continue
   It is obvious that the first fusion method is the most
   informative because the information is refined to the
   intersection of sets given by each source. It is also the
   most “risky” approach because the real value of is
   assumed to be inside a smaller set than the two initial
   sets. The second fusion method is more reliable since all
   the information given by the two sources is preserved.
   The drawback of such an approach is a loss of accuracy
   since the set assumed to contain , is larger than each of
   the initial sets.

Methods for Data and Information Fusion                                       25
            Institute for Parallel Processing - Bulgarian Academy of Sciences


            Homogeneous sensor fusion
 AND Operator. This method transforms the output of the sensors in a binary
  yes/no consensus operating with logical AND. After that thresholds are applied
  to find the result. The procedure is very simple, intuitive and fast, if the values of
  thresholds are determined in advance. The method does not take into account the
  degree of confidence of each sensor.
 Weighted Average. This method takes a weighted average of available sensor
  data and uses it as the fused value. Usually the weights are proportional to
  accuracy of sensors or to credibility of sensor information.
 Voting. The voting schemes main advantage is computation efficiency. Voting
  involves the derivation of an output data object from a collection of n input data
  objects, as prescribed by the requirements and constraints of a voting algorithm.
  The voting algorithms can be quite complex in terms of content and structure of
  the input data objects and how they handle the votes (weights) at input and
  output.


Methods for Data and Information Fusion                                            26
          Institute for Parallel Processing - Bulgarian Academy of Sciences


                          Sensor fusion
 Bayesian Theory. The use of Bayesian inference theory is widely
   spread for the fusion of redundant information. The most known
   method is the Kalman Filter, that is optimal in a statistical sense (it
   presents the least square error). Bayesian theory is also used to
   establish the weights linking the sensors in a weighted average
   fusion architecture. Moreover, some reductions of superbayesian
   methods to probabilistic evidence combination formulas have been
   provided. Some problems arise in a Bayesian framework: I) it
   does not distinguish between “lack of evidence” and “disbelief”;
   ii) practical difficulties in setting the apriori probabilities:
   noninformative priors can cause a wrong bias of further reasoning;
   iii) it assumes that the knowledge sources are consistent.

Methods for Data and Information Fusion                                       27
          Institute for Parallel Processing - Bulgarian Academy of Sciences


                            Sensor Fusion
 Information Theory. Mutual information, in the
 form of the Kullback-Leiber divergence, has been
 used in [12] as a way of combining probabilistic
 masses (sensor outputs). This is yet another
 method of fusing two probabilities, this time with
 a non-bayesian law, adding some information on
 average image values (e.g. depending on lighting
 conditions). The local maximum of the mutual
 information is then taken as the fused value.
Methods for Data and Information Fusion                                       28
          Institute for Parallel Processing - Bulgarian Academy of Sciences


                            Sensor Fusion
     Belief Theory. Dempster-Shafer evidential
     reasoning is used to compute the belief of a given
     event from two or more assessments provided by
     different knowledge sources at a symbolic level.
     This theory is based on the premise that each
     source of information provides only a partial belief
     about a proposition. Problem – redistribution of
     conflicts.
     Dezert Smarandache Theory(DSmT). DSmT is
     analogous to Dempster-Shafer evidential reasoning
     theory but overcomes some drawbacks of this
     theory

Methods for Data and Information Fusion                                       29
          Institute for Parallel Processing - Bulgarian Academy of Sciences


                             Sensor fusion
Fuzzy Reasoning. Fuzzy sets and variables are used to
 deal with real-world models where the usual ideal
 mathematical assumptions are inappropriate. Under the
 fuzzy framework, the possibility theory has emerged to
 represent imprecision in terms of fuzzy sets and to
 quantify uncertainty through four proposed notions:
 possibility, necessity, plausibility, and credibility
 distributions .
 Geometric Methods, e.g. using uncertainty ellipsoids.
 Parametrical identification – if we know model, we can
 estimate parameters;

Methods for Data and Information Fusion                                       30
          Institute for Parallel Processing - Bulgarian Academy of Sciences



                      Level 2,3,4 fusion
        Belief Propagation Nets
        Markov Random Fields
        Factor Graphs
        Game theory




Methods for Data and Information Fusion                                       31
          Institute for Parallel Processing - Bulgarian Academy of Sciences


                 New research direction
 New tracking filters – may be FISST, may be
 new one
 Increased interest in image fusion methods -
 improvement of existing, search for new ones.
 Increased interest on higher level fusion – not
 only theoretical but engineering approach
 Decision level methods for fusion – like Dezert
 –Smarandache Theory or new ones.

Methods for Data and Information Fusion                                       32

								
To top