학위논문 (박사)-- 서울대학교 대학원 : 전기·컴퓨터공학부, 2015. 8. 이범희. ; In unknown environments, multiple robots must have capabilities to sense and interpret their surroundings, and localize themselves before performing some missions such as exploring the mineral resources and rescuing people. It is usually called multi-robot simultaneous localization and mapping. To perform multi-robot SLAM more accurately, robots are required to build maps of their surroundings accurately. In addition, inter-robot measurements should be properly utilized in the SLAM process. In this dissertation, a novel Rao-Blackwellized particle filter based SLAM framework is presented using geometric information and inter-robot measurements for accurate multi-robot SLAM. For SLAM, a Rao-Blackwellized particle filter (RBPF) is basically one of representative methods. It takes advantage of linear time-complexity which is linearly proportional to the number of features by factoring the full SLAM posterior into the product of a robot path posterior and landmark posteriors. Additionally, it deals with multi-hypothesis data association using particles with their own data association. They makes it more robust than extended Kalman filter based SLAM. The proposed SLAM framework is divided into two major parts. First, the RBPF is improved using cooperation among particles in case of single robot SLAM, which is called Relational RBPF-SLAM. Here, the framework basically follows the process of the factored solution to SLAM using the unscented Kalman filter (UFastSLAM), which is an accurate instance of RBPF-SLAM. A concept of particle to particle cooperation is considered in the importance weight step and the resampling step to increase the SLAM accuracy and solve some inherent problems such as the particle depletion problem and the data association problem. The particle depletion problem is almost eliminated using the formation maintenance of particles which is controlled without any rejection or replication of particles during the resampling step. In addition, to overcome the data association problem, the posterior distribution is estimated more accurately by compensating improperly assigned weights of particles. Secondly, to reduce the accumulated robot pose errors and feature errors, inter-robot measurements are utilized in the proposed RBPF-SLAM framework. They can be measured when a rendezvous between robots occurs or robots share common features. To deal with the inter-robot measurements, a Kalman consensus filter scheme is involved in the proposed RBPF-SLAM framework, which is robust than the covariance intersection method. Several simulations and experiments show significant improvements of the proposed RBPF-SLAM framework in both the accuracy of robot poses and map quality by comparing the state of the art techniques, i.e. FastSLAM 2.0, particle swarm optimization (PSO) based FastSLAM, UFastSLAM, particle fission based UFastSLAM and PSO based UFastSLAM. ; Chapter 1. Introduction 1 1.1 Background and motivations 1 1.2 Related Work 4 1.2.1 Scan Matching based SLAM 5 1.2.2 Graph SLAM 7 1.2.3 Bayesian Filter based SLAM for Single Robot 9 1.2.4 Bayesian Filter based SLAM for Multiple Robots 16 1.3 Contributions 19 1.4 Organization 22 Chapter 2. Fundamental Techniques for Multi-Robot SLAM 23 2.1 Rao-Blackwellized Particle Filter based SLAM 24 2.1.1 Sampling Strategy 25 2.1.2 Feature State Estimation 28 2.1.3 Calculating Importance Weight and Resampling Strategy 29 2.2 Covariance Intersection (CI) and Kalman Consensus Information Filter (KCIF) in RBPF-SLAM 31 Chapter 3. Particle-to-Particle Cooperation in RBPF-SLAM 35 3.1 Weight Compensation using Particle Cooperation 36 3.2 Applicability of Particle Formation Maintenance 43 3.3 Particle Formation Maintenance 45 3.4 Overview of Relational RBPF-SLAM 50 3.5 Complexity of Relational RBPF-SLAM 51 Chapter 4. Robot to Robot Cooperation in RBPF-SLAM 53 4.1 Multi-Robot Initialization in the Unknown Initial Condition 53 4.2 Compensation in Rendezvous Events 54 4.3 Compensation in Feature Sharing Events 60 4.4 Overview of the Proposed RBPF-SLAM Framework 61 Chapter 5. Simulations 64 5.1 Verification for Needs of Weight Compensation and Particle Formation Maintenance 65 5.1.1 Simple Weight Compensation 65 5.1.2 Piecewise Average based Weight Compensation 67 5.1.3 Particle Formation Maintenance Test 69 5.2 Simulation with Unknown Data Association 72 5.3 Simulations for Robot Pose Consensus and Feature Consensus 75 5.3.1 Robot Pose Consensus Test I 76 5.3.2 Robot Pose Consensus Test II 79 5.3.3 Feature Consensus Test 82 5.4 Discussions 85 Chapter 6. Experiments 88 6.1 Line Feature Extraction 89 6.2 Tests using Car Park Dataset 91 6.3 Tests using Victoria Park Dataset 96 6.4 Indoor Experiments 102 6.5 Outdoor Experiments 106 6.5.1 Performance Comparison for Single Robot SLAM 107 6.5.2 Performance Comparison for Data Consensus 115 6.6 Discussions 116 Chapter 7. Conclusions 121 Bibliography 124 ; Doctor


    Access

    Download


    Export, share and cite



    Title :

    Cooperative Rao-Blackwellized Particle Filter based SLAM framework using Geometric Information and Inter-Robot Measurements ; 기하학적 정보와 로봇 간 측정값을 이용한 협조적 Rao-Blackwellized Particle Filter 기반 SLAM 프레임워크



    Publication date :

    2015-01-01


    Type of media :

    Theses


    Type of material :

    Electronic Resource


    Language :

    English



    Classification :

    DDC:    629



    Novel Rao-Blackwellized Particle Filter for Mobile Robot SLAM Using Monocular Vision

    Maohai Li / Bingrong Hong / Zesu Cai et al. | BASE | 2008

    Free access

    Geomagnetic Aided Navigation Using Rao Blackwellized Particle Filter

    Cuenca, Andrei / Moncayo, Hever | TIBKAT | 2023


    A Rao-Blackwellized particle filter for EigenTracking

    Zia Khan, / Balch, T. / Dellaert, F. | IEEE | 2004


    A Rao-Blackwellized Particle Filter for EigenTracking

    Khan, Z. / Balch, T. / Dellaert, F. et al. | British Library Conference Proceedings | 2004


    Geomagnetic Aided Navigation Using Rao Blackwellized Particle Filter

    Cuenca, Andrei / Moncayo, Hever | TIBKAT | 2023