Geintra

{U}sing external cameras to achieve robot localization has been widely proposed in the area of {I}ntelligent {S}paces. {R}ecently, an online approach that simultaneously obtains robot’s pose and its 3{D} structure using a single external camera has been developed [8]. {S}uch proposal relies on a proper initialization of pose and structure information of the robot. {T}he present paper proposes a solution to initialization which consists of retrieving 3{D} structure and motion of a rigid object from a set of point matches measured by the camera. {A} batch {S}tructure from {M}otion ({SFM}) approach is proposed along a short path. {B}y incorporating odometry information available in the robot, the ambiguity generated by a single view in the solution is solved. {W}e propose to describe robot’s motion and image detection as statistical processes in which the uncertainty is properly modelled. {U}sing a {G}aussian equivalence of the processes involved, the {SFM} cost function is expressed as a {M}aximum {L}ikelihood optimization. {T}he paper shows the improvements of the approach in the presence of the usual odometry drift noise, compared with those using {E}uclidean distance as a likelihood. {T}he proposed method is assessed on synthetic and real data.