A stochastic map for uncertain spatial relationships
Proceedings of the 4th international symposium on Robotics Research
FastSLAM: a factored solution to the simultaneous localization and mapping problem
Eighteenth national conference on Artificial intelligence
Learning Occupancy Grid Maps with Forward Sensor Models
Autonomous Robots
Hi-index | 0.00 |
Rao-Blackwellized particle filters simultaneous localization and mapping can yield effective results but it has the tendency to become inconsistent. To ensure consistency, a methodology of an unscented Kalman filter and Markov Chain Monte Carlo resampling are incorporated. More accurate nonlinear mean and variance of the proposal distribution are obtained without the linearization procedure in extended Kalman filter. Furthermore, the particle impoverishment induced by resampling is averted after the resample move step. Thus particles are less susceptible to degeneracies. The algorithms are evaluated on accuracy and consistency using computer simulation. Experimental results illustrate the advantages of our methods over previous approaches.