A method for localizing a robot in a localization plane

• Main Authors: De La Fortelle Arnaud, Wirbel Emilie
• Format: Patent
• Language: English
• Published: 24.11.2017
• Subjects: CONTROLLING; GYROSCOPIC INSTRUMENTS; MEASURING; MEASURING DISTANCES, LEVELS OR BEARINGS; NAVIGATION; PHOTOGRAMMETRY OR VIDEOGRAMMETRY; PHYSICS; REGULATING; SURVEYING; SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES; TESTING

Navigation and localization is a crucial problem of robotics, particularly in a human populated environment. Although approaches exist to provide robots with a navigation system, these do not provide satisfactory solutions and improvements are desired. The invention concerns a method for localizing a robot in a localization plane associated with a bi-dimentional reference with two axis x and y comprising the following steps: determining (200) by odometry an estimation of the coordinates x1 and y1 of the robot as well as an estimation of its orientation Θ1; determining (202) an estimation Θ2 of the orientation of the robot by using a virtual compass; determining (204) an estimation Θ3 of the orientation of the robot by correlating parts of a reference panorama with parts of a query panorama; determining (206) an estimation x4, y4 of the robot position by using an Iterative Closest Points technique; determining the standard deviations σ_x1, σ_x2, σ_θ1 σ_θ2, σ_θ3, σ_x4, σ_y4 of the aforementioned estimations; determining (220) probability distributions G(x1), G(y1), G(Θ1), G(Θ2), G(Θ3), G(x4) and G(y4) of each available estimation using said standard deviations; determining (221) three global distributions GLOB(x), GLOB(y) and GLOB(Θ) and determining a global estimation xg, yg of the coordinates of the robot in the localization plane as well as an global estimation Θg of its orientation by applying maximum likelihood to the global distributions.