The Study of Improving Kalman Filters Family for Nonlinear SLAM

Authors:
Wu Zhou;Chunxia Zhao;Jianhui Guo
Affiliations:
College of Computer Science and Technology, Nanjing University of Science and Technology, Nanjing, China;College of Computer Science and Technology, Nanjing University of Science and Technology, Nanjing, China;College of Computer Science and Technology, Nanjing University of Science and Technology, Nanjing, China
Venue:
Journal of Intelligent and Robotic Systems
Year:
2009

Citing 5
Cited 1

On the representation and estimation of spatial uncertainly

International Journal of Robotics Research
A Probabilistic Approach to Concurrent Mapping and Localization for Mobile Robots

Machine Learning - Special issue on learning in autonomous robots
Rao-Blackwellised Particle Filtering for Dynamic Bayesian Networks

UAI '00 Proceedings of the 16th Conference on Uncertainty in Artificial Intelligence
FastSLAM 2.0: an improved particle filtering algorithm for simultaneous localization and mapping that provably converges

IJCAI'03 Proceedings of the 18th international joint conference on Artificial intelligence
Toward multidimensional assignment data association in robot localization and mapping

IEEE Transactions on Robotics

Experimental Study of Iterated Kalman Filters for Simultaneous Localization and Mapping of Autonomous Mobile Robots

Journal of Intelligent and Robotic Systems

Quantified Score

Hi-index	0.00

Visualization

Abstract

When Extended Kalman Filter is used to solve the SLAM problem of a nonlinear system, the linearization error will lead to severe estimation error or even make the method to be divergent. After analyzing the linearization principle of Kalman filters family, two improved methods are suggested to decrease the linearization error. These two methods improve posterior estimation accuracy by revising the observation-update step. Simulation results indicate that the two methods are feasible. The method named `Mean Extended Kalman Filter' performs much better than EKF and UKF for nonlinear SLAM. And the iterated version of EKF and UKF even falls behind MEKF in estimation accuracy. In addition, MEKF is computationally efficient. With a view to both estimation accuracy and computational complexity, MEKF seems to be the best filter of the Kalman filters family for nonlinear SLAM. Experiments are carried out with `Car Park Dataset' and `Victoria Park Dataset' to evaluate the performance of MEKF based SLAM solutions. And the experimental results validate the effectiveness of MEKF in real SLAM applications.