Unscented Transform for SLAM Using Gaussian Mixture Model with Particle Filter

The aims of the simultaneous localization and mapping (SLAM) in a real-world environment is to obtain faster processing speed, more precise predictable results, and better system approximation and consistency. This paper proposes a combination of the Gaussian mixture model (GMM) with particle filter (PF) and unscented Kalman filter (UKF) for the robot SLAM. Also, the PF Markov chain Monte Carlo (MCMC) method is applied to get a better particle distribution. The application of the SLAM process will depend on the incoming measurement data; whether there is a landmark signal being detected or not. In the former condition, the new landmark position computing, the GMM updating and, the robot and the landmark position updating are needed and, in the latter case, robot and landmark position are predicted through a prediction equation. From the experimental results, it can be seen that the processing speed and the precision of the proposed method are better than that of the FAST SLAM and the UKF SLAM,especially in a dense map environment.