A FastSLAM Algorithm Based on the Auxiliary Particle Filter with Stirling Interpolation

The choice of the distribution model and the consistency of the result are very important for FastSLAM. This paper provides a method which combines the auxiliary variable model with FastSLAM and, uses Stirling Interpolation to approximate the nonlinear functions. It overcomes the drawbacks of the FastSLAM by using a model ignoring the measurement data and the approximation error for nonlinear functions. This approach improves the estimation accuracy and reduces the degradation speed of the particle. Simulation results demonstrate the excellence of the proposed algorithm.