Skip to main content

Table 1 Pseudo-code for the particle filter-based ATR algorithm

From: Automated target tracking and recognition using coupled view and identity manifolds for shape representation

• Initialization: Draw X 0 j ~N ( X 0 , 1 ) , and α 0 j = α 0 j {1,..., N p }. Here X0 and α0 are the initial kinematic state and identity values, respectively.
• For t = 1,..., T (number of frames)
   1. For j = 1,..., N p (number of particles)
1.1 Draw samples X t j ~p ( X t j X t - 1 j ) and α t j ~p ( α t j α t - 1 j ) as in (10) and (11).
1.2 Compute weights w t j =p ( z t α t j , X t j ) using (12).
End
   2. Normalize the weights such that j = 1 N p w t j =1.
   3. Compute the mean estimates of the kinematics and identity X ^ t = j = 1 N p w t j X t j and α ^ t = j = 1 N p w t j α t j .
   4. Set [ α t j , X t j ] = resample ( α t j , X k j , w k j ) to increase the effective number of particles [39].
• End