Abstract
While FastSLAM algorithm is a popular solution to SLAM problem, it suffers from two major drawbacks: one is particle set degeneracy due to lack of observation information in proposal distribution; the other is errors accumulation caused by inaccuracy linearization of the robot motion model and the observation model. To overcome the problems, we propose a new Jacobian free CFastSLAM algorithm. The main contribution of this work lies in the utilization of Cubature Kalman Filter (CKF), which calculate Gaussian Weight Integral based on Cubature Rule, to design an optimal proposal distribution of the particle filter and to estimate the environment feature landmarks. On the basis of Rao-Blackwellized particle filter, proposed algorithm is comprised by two main parts: in the first part, a Cubature Particle Filter (CPF) is derived to localize the robot; in the second part, a set of CKFs is used to estimate the environment landmarks. The performance of the CFastSLAM is investigated and compared with that of FastSLAM2.0 and UFastSLAM in simulations and experiments. Results verify that the CFastSLAM improves SLAM performance. © 2012 IEEE.
| Original language | English |
|---|---|
| Title of host publication | 2012 IEEE International Conference on Robotics and Automation, ICRA 2012: Proceedings |
| Publisher | Institute of Electrical and Electronics Engineers Inc. |
| Pages | 3063-3068 |
| Number of pages | 6 |
| ISBN (Electronic) | 9781467314053 |
| ISBN (Print) | 9781467314039 |
| DOIs | |
| Publication status | Published - 2012 |
| Externally published | Yes |
| Event | 2012 IEEE International Conference on Robotics and Automation - RiverCentre, Saint Paul, United States Duration: 14 May 2012 → 18 Jun 2012 |
Conference
| Conference | 2012 IEEE International Conference on Robotics and Automation |
|---|---|
| Abbreviated title | ICRA 2012 |
| Country/Territory | United States |
| City | Saint Paul |
| Period | 14/05/12 → 18/06/12 |
Funding
This work was jointly supported by State Key Laboratory of Robotics and System of HIT (Nos SKLRS-2009-ZD-04, SKLRS-2010-ZD-15), National Nature Science Foundation of China (Nos 60909055, 61005070), China Postdoctoral Science Foundation Special Funded Project (Nos 201003144, 201104160), Fundamental Research Funds for the Central Universities of China (No 2009JBZ001-2) and Beijing Jiaotong University Research Program (No2010RC008).