Distributed Multi-Robot Cooperative Localization Using Bayesian Fusion on the Special Euclidean Group

multi-robot;distributed;cooperative localization;Lie group;exponential coordinate;Robotics

Li, Xiao

Johns Hopkins University

关键词: multi-robot;    distributed;    cooperative localization;    Lie group;    exponential coordinate;    Robotics;   

Others  :  https://jscholarship.library.jhu.edu/bitstream/handle/1774.2/38149/LI-THESIS-2014.pdf?sequence=1&isAllowed=y

瑞士|英语

来源: JOHNS HOPKINS DSpace Repository

【 摘 要 】

This thesis presents a new distributed cooperative localization technique using a secondorder sensor fusion method developed for the Special Euclidean group. Uncertaintiesin the robot pose, sensor measurements and landmark positions (neighboring robots inthis case) are modeled as Gaussian distributions in exponential coordinates. This provesto be a bettert for posterior distributions resulting from the motion of nonholonomickinematic systems with stochastic noise (compared to standard Gaussians in Cartesiancoordinates). We provide a recursive closed-form solution to the multi-sensor fusionproblem that can be used to incorporate a large number of sensor measurements intothe localization routine and can be implemented in real time. The technique can beused for nonlinear sensor models without the need for further simpli cations given thatthe required relative pose and orientation information can be provided, and it is scalablein that the computational complexity does not increase with the size of the robot teamand increases linearly with the number of measurements taken from nearby robots. Theproposed approach is validated with simulationrst conceptually in Matlab then morerealistically in the robotics simulator ROS/Gazebo. It is also compared with one of thecurrent state of the art methods (distributed EKF) and shows promising results.