A new approach to accelerate IMM algorithm of maneuvering target tracking based on CUDA

Abstract

According to the characteristics of Uniform motion in two-dimensional space, On the basis of the target motion model and observation model, adopted based on the Interacting Multiple Model algorithm (IMM) Kalman filter to the maneuvering target tracking. However, due to there are a lot of recursive and matrix operations in IMM algorithm, So the real time is reduced ,this also is the reason that it's works hard to achieve in project. This article through the CUDA to speed it up, Greatly improved the performance of real-time on the premise of do not affect arithmetic performance. Simulation results show that, even if when motion model is change the performance of real-time has also been improved so much, and the filtering error is also relatively small.