Target tracking through a coordinated turn
Mark R. Morelande, Neil J. Gordon · 2006
Allowing for perturbations in speed and turn rate, the state of a target moving in a coordinated turn obeys a nonlinear stochastic differential equation which cannot be discretised exactly. Existing algorithms for coordinated turn tracking avoid this problem by ignoring perturbations in the continuous-time model and adding process noise only after discretisation. We retain this modelling by discretising using first and second order Taylor approximations to the continuous-time coordinated turn dynamic model. The discrete-time models are used as the basis for a particle filter for tracking a target moving in a coordinated turn. The performances of the discretisation techniques and the effect of different coordinate systems on tracking performance are examined.