Volume 16, Issue 12 (2-2017)                   Modares Mechanical Engineering 2017, 16(12): 490-500 | Back to browse issues page

XML Persian Abstract Print

1- Department of Electrical and Computer Engineering Isfahan University of Technology
2- Department of Electrical and Computer Engineering Isfahan University of Technology
Abstract:   (3524 Views)
In this paper, by introducing of development of two approaches based on the relative map filter (RMF); it has been tried to improve simultaneous localization and mapping (SLAM). The implementation of Extended Kalman Filter SLAM (EKF-SLAM) in large environments is not practical due to large volume of calculations. On the other hand, the observation and motion models of many robots are nonlinear and these cause the divergence of EKF-SLAM. The basis of RMF is relative distances between landmarks; therefore its equations are independent from the robot motion model. Also, the robot observation model can be linearly defined and its convergence is guaranteed. Despite these features, the relative filter proposed methods are faced with the problem of ambiguity in absolute positioning of robot and landmarks. In this article, ILPE (Improved Lowest Position Estimation) and IMVPE (Improved Minimum Variance Position Estimation) methods are introduced. In these methods, the ambiguity problem in localization and mapping of robot and landmarks are solved by sequential switching between absolute and relative spaces. The calculation volume of these methods does not depend on the number of landmarks and depends on the average number of landmarks observed in each scan of the robot. In this paper, the equations and the required algorithm to find the position of landmarks and robot are presented. Moreover by simulation, the performance and efficiency of the proposed methods are discussed in comparison with the previous methods including EKF-SLAM.
Full-Text [PDF 628 kb]   (4325 Downloads)    
Article Type: Research Article | Subject: robatic
Received: 2016/09/14 | Accepted: 2016/10/30 | Published: 2016/12/18

Rights and permissions
Creative Commons License This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.